new5-AuthorGuide.tex

\section{Author Guide}
\label{aut}

This part of the manual documents the author interface of the \biblatex package. The author guide covers everything you need to know in order to write new citation and bibliography styles or localisation modules. You should read the user guide first before continuing with this part of the manual.

\subsection{Overview}
\label{aut:int}

Before we get to the commands and facilities provided by \biblatex, we will have a look at some of its fundamental concepts. The \biblatex package uses auxiliary files in a special way. Most notably, the \file{bbl} file is used differently and there is no concept of a style-dependent \file{bst} file. With \latex's standard bibliographic facilities, a document includes any number of citation commands in the document body plus \cmd{bibliographystyle} and \cmd{bibliography}, usually towards the end of the document. The location of the former is arbitrary, the latter marks the spot where the list of references is to be printed:

\begin{ltxexample}
\documentclass{...}
\begin{document}
\cite{...}
...
\bibliographystyle{...}
\bibliography{...}
\end{document}
\end{ltxexample}
%
Processing this files requires that a certain procedure be followed. This procedure is as follows:

\begin{enumerate}

\item Run \bin{latex}: On the first run, \cmd{bibstyle} and \cmd{bibdata} commands are written to the \file{aux} file, along with \cmd{citation} commands for all citations. At this point, the references are undefined because \latex is waiting for \bibtex to supply the required data. There is also no bibliography yet.

\item Run \bin{bibtex}: \bibtex writes a \env{thebibliography} environment to the \file{bbl} file, supplying all entries from the \file{bib} file which were requested by the \cmd{citation} commands in the \file{aux} file.

\item Run \bin{latex}: Starting with the second run, the \cmd{bibitem} commands in the \env{thebibliography} environment write one \cmd{bibcite} command for each bibliography entry to the \file{aux} file. These \cmd{bibcite} commands define the citation labels used by \cmd{cite}. However, the references are still undefined because the labels are not available until the end of this run.

\item Run \bin{latex}: Starting with the third run, the citation labels are defined as the \file{aux} file is read in at the end of the preamble. All citations can now be printed.

\end{enumerate}

Note that all bibliographic data is written to the \file{bbl} file in the final format. The \file{bbl} file is read in and processed like any printable section of the document. For example, consider the following entry in a \file{bib} file:

\begin{lstlisting}[style=bibtex]{}
@Book{companion,
  author    = {Michel Goossens and Frank Mittelbach and Alexander Samarin},
  title     = {The LaTeX Companion},
  publisher = {Addison-Wesley},
  address   = {Reading, Mass.},
  year      = {1994},
}
\end{lstlisting}
%
With the \path{plain.bst} style, \bibtex exports this entry to the \file{bbl} file as follows:

\begin{ltxexample}
\bibitem{companion}
Michel Goossens, Frank Mittelbach, and Alexander Samarin.
\newblock {\em The LaTeX Companion}.
\newblock Addison-Wesley, Reading, Mass., 1994.
\end{ltxexample}
%
By default, \latex generates numeric citation labels, hence \cmd{bibitem} writes lines such as the following to the \file{aux} file:

\begin{ltxexample}
\bibcite{companion}{1}
\end{ltxexample}
%
Implementing a different citation style implies that more data has to be transferred via the \file{aux} file. With the \sty{natbib} package, for example, the \file{aux} file contains lines like this one:

\begin{ltxexample}
\bibcite{companion}{{1}{1994}{{Goossens et~al.}}{{Goossens, Mittelbach,
and Samarin}}}
\end{ltxexample}
%
The \biblatex package supports citations in any arbitrary format, hence citation commands need access to all bibliographic data. What this would mean within the scope of the procedure outlined above becomes obvious when looking at the output of the \sty{jurabib} package which also makes all bibliographic data available in citations:

\begin{ltxexample}
\bibcite{companion}{{Goossens\jbbfsasep Mittelbach\jbbstasep Samarin}%
  {}{{0}{}{book}{1994}{}{}{}{}{Reading, Mass.\bpubaddr{}Addison-Wesley%
  \bibbdsep{} 1994}}{{The LaTeX Companion}{}{}{2}{}{}{}{}{}}{\bibnf
  {Goossens}{Michel}{M.}{}{}\Bibbfsasep\bibnf{Mittelbach}{Frank}{F.}%
  {}{}\Bibbstasep\bibnf{Samarin}{Alexander}{A.}{}{}}{\bibtfont{The
  LaTeX Companion}.\ \apyformat{Reading, Mass.\bpubaddr{}
  Addison-Wesley\bibbdsep{} 1994}}}
\end{ltxexample}
%
In this case, the contents of the entire \env{thebibliography} environment are effectively transferred via the \file{aux} file. The data is read from the \file{bbl} file, written to the \file{aux} file, read back from the \file{aux} file and then kept in memory. The bibliography itself is still generated as the \file{bbl} file is read in. The \biblatex package would also be forced to cycle all data through the \file{aux} file. This implies processing overhead and is also redundant because the data has to be kept in memory anyway.

The traditional procedure is based on the assumption that the full bibliographic data of an entry is only required in the bibliography and that all citations use short labels. This makes it very effective in terms of memory requirements, but it also implies that it does not scale well. That is why \biblatex takes a different approach. First of all, the document structure is slightly different. Instead of using \cmd{bibliography} in the document body, database files are specified in the preamble with \cmd{addbibresource}, \cmd{bibliographystyle} is omitted entirely (all features are controlled by package options), and the bibliography is printed using \cmd{printbibliography}:

\begin{ltxexample}
\documentclass{...}
\usepackage[...]{biblatex}
\addbibresource{...}
\begin{document}
\cite{...}
...
\printbibliography
\end{document}
\end{ltxexample}
%
In order to streamline the whole procedure, \biblatex essentially employs the \file{bbl} file like an \file{aux} file, rendering \cmd{bibcite} obsolete. We then get the following procedure:

\begin{enumerate}

\item Run \bin{latex}: The first step is similar to the traditional procedure described above: \cmd{bibstyle} and \cmd{bibdata} commands are written to th \file{bcf} file, along with \cmd{citation} commands for all citations. We then wait for the backend to supply the required data.

\item Run \bin{biber}: The backend supplies those entries from the \file{bib} file which were requested by the \cmd{citation} commands in the auxiliary file. However, it does not write a printable bibliography to the \file{bbl} file, but rather a structured representation of the bibliographic data. Just like an \file{aux} file, this \file{bbl} file does not print anything when read in. It merely puts data in memory.

\item Run \bin{latex}: Starting with the second run, the \file{bbl} file is processed right at the beginning of the document body, just like an \file{aux} file. From this point on, all bibliographic data is available in memory so that all citations can be printed right away.\footnote{If the \opt{defernumbers} package option is enabled \biblatex uses an algorithm similar to the traditional procedure to generate numeric labels. In this case, the numbers are assigned as the bibliography is printed and then cycled through the backend auxiliary file. It will take an additional \latex run for them to be picked up in citations.} The citation commands have access to the complete bibliographic data, not only to a predefined label. The bibliography is generated from memory using the same data and may be filtered or split as required.

\end{enumerate}

Let's consider the sample entry given above once more:

\begin{lstlisting}[style=bibtex]{}
@Book{companion,
  author    = {Michel Goossens and Frank Mittelbach and Alexander Samarin},
  title     = {The LaTeX Companion},
  publisher = {Addison-Wesley},
  address   = {Reading, Mass.},
  year      = {1994},
}
\end{lstlisting}
%
This entry is essentially exported in the following format:

\begin{ltxexample}
\entry{companion}{book}{}
  \labelname{author}{3}{}{%
    {{uniquename=0,hash=...}{Goossens}{G.}{Michel}{M.}{}{}{}{}}%
    {{uniquename=0,hash=...}{Mittelbach}{M.}{Frank}{F.}{}{}{}{}}%
    {{uniquename=0,hash=...}{Samarin}{S.}{Alexander}{A.}{}{}{}{}}%
  }
  \name{author}{3}{}{%
    {{uniquename=0,hash=...}{Goossens}{G.}{Michel}{M.}{}{}{}{}}%
    {{uniquename=0,hash=...}{Mittelbach}{M.}{Frank}{F.}{}{}{}{}}%
    {{uniquename=0,hash=...}{Samarin}{S.}{Alexander}{A.}{}{}{}{}}%
  }
  \list{publisher}{1}{%
    {Addison-Wesley}%
  }
  \list{location}{1}{%
    {Reading, Mass.}%
  }
  \field{title}{The LaTeX Companion}
  \field{year}{1994}
\endentry
\end{ltxexample}
%
As seen in this example, the data is presented in a structured format that resembles the structure of a \file{bib} file to some extent. At this point, no decision concerning the final format of the bibliography entry has been made. The formatting of the bibliography and all citations is controlled by \latex macros, which are defined in bibliography and citation style files.

\subsection{Bibliography Styles}
\label{aut:bbx}

A bibliography style is a set of macros which print the entries in the bibliography. Such styles are defined in files with the suffix \file{bbx}. The \biblatex package loads the selected bibliography style file at the end of the package. Note that a small repertory of frequently used macros shared by several of the standard bibliography styles is included in \path{biblatex.def}. This file is loaded at the end of the package as well, prior to the selected bibliography style.

\subsubsection{Bibliography Style Files}
\label{aut:bbx:bbx}

Before we go over the individual components of a bibliography style, consider this example of the overall structure of a typical \file{bbx} file:

\begin{ltxexample}
\ProvidesFile{example.bbx}[2006/03/15 v1.0 biblatex bibliography style]

\defbibenvironment{bibliography}
  {...}
  {...}
  {...}
\defbibenvironment{shorthand}
  {...}
  {...}
  {...}
\InitializeBibliographyStyle{...}
\DeclareBibliographyDriver{article}{...}
\DeclareBibliographyDriver{book}{...}
\DeclareBibliographyDriver{inbook}{...}
...
\DeclareBibliographyDriver{shorthand}{...}
\endinput
\end{ltxexample}
%
The main structure of a bibliography style file consists of the following commands:

\begin{ltxsyntax}

\cmditem{RequireBibliographyStyle}{style}

This command is optional and intended for specialized bibliography styles built on top of a more generic style. It loads the bibliography style \path{style.bbx}.

\cmditem{InitializeBibliographyStyle}{code}

Specifies arbitrary \prm{code} to be inserted at the beginning of the bibliography, but inside the group formed by the bibliography. This command is optional. It may be useful for definitions which are shared by several bibliography drivers but not used outside the bibliography. Keep in mind that there may be several bibliographies in a document. If the bibliography drivers make any global assignments, they should be reset at the beginning of the next bibliography.

\cmditem{DeclareBibliographyDriver}{entrytype}{code}

Defines a bibliography driver. A <driver> is a macro which handles a specific entry type (when printing bibliography lists) or a specific named bibliography list (when printing bibliography lists). The \prm{entrytype} corresponds to the entry type used in \file{bib} files, specified in lowercase letters (see \secref{bib:typ}). The \prm{entrytype} argument may also be an asterisk. In this case, the driver serves as a fallback which is used if no specific driver for the entry type has been defined. The \prm{code} is arbitrary code which typesets all bibliography entries of the respective \prm{entrytype}. This command is mandatory. Every bibliography style should provide a driver for each entry type.

\cmditem{DeclareBibliographyAlias}{alias}{entrytype}

If a bibliography driver covers more than one entry type, this command may be used to define an alias where \prm{entrytype} is the name of a defined driver. This command is optional. The \prm{alias} argument may also be an asterisk. In this case, the \prm{entrytype} driver serves as a fallback which is used if no specific driver for an entry has been defined.

Note that an alias declared with \cmd{DeclareBibliographyAlias} only <reroutes> the bibliography driver from \prm{alias} to \prm{entrytype}. Type-specific formatting directives still operate with the old \prm{alias} name. \cmd{DeclareBibliographyAlias} thus provides only a <soft> alias. If a complete alias is desired so that \prm{alias} and \prm{entrytype} are completely indistinguishable and use the same type-specific formatting, an approach with source mapping would be more appropriate (cf.~the mappings for \secref{bib:typ:als} in \secref{apx:maps:bibtex}, this would give a <hard> alias).

\cmditem{DeclareBibliographyOption}[datatype]{key}[value]{code}

This command defines additional preamble options in \keyval format. The \prm{key} is the option key. The \prm{code} is arbitrary \tex code to be executed whenever the option is used. The value passed to the option is passed on to the \prm{code} as |#1|. The optional \prm{value} is a default value to be used if the bare key is given without any value. This is useful for boolean switches.
The \prm{datatype} is a the datatype for the option. If omitted, it defaults to <boolean>. For example, with a definition like the following:

\begin{ltxexample}
\DeclareBibliographyOption[boolean]{somekey}[true]{...}
\end{ltxexample}
%
giving <\texttt{somekey}> without a value is equivalent to <\kvopt{somekey}{true}>. Valid \prm{datatype} values are defined in the default \biber Datamodel as:

\begin{ltxexample}
\DeclareDatamodelConstant[type=list]{optiondatatypes}{boolean,integer,string,xml}
\end{ltxexample}

\cmditem{DeclareTypeOption}[datatype]{key}[value]{code}

Similar to \cmd{DeclareBibliographyOption} but defines options which are settable on a per"=type basis using the optional argument of \cmd{ExecuteBibliographyOptions} (see \secref{use:cfg:opt}). The \prm{code} is executed whenever \biblatex prepares the data of an entry of the type for which the option has been set for use by a citation command or a bibliography driver.

\cmditem{DeclareEntryOption}[datatype]{key}[value]{code}

Similar to \cmd{DeclareBibliographyOption} but defines options which are settable on a per"=entry basis in the \bibfield{options} field from \secref{bib:fld:spc}. The \prm{code} is executed whenever \biblatex prepares the data of the entry for use by a citation command or a bibliography driver.

\cmditem{DeclareBiblatexOption}{scope,\dots}[datatype]{key}[value]{code}

This command is a convenient interface to declare an option for several scopes at once. The \prm{scope} argument may be a comma"=separated list of scopes for which the option will be declared. Currently the scopes \opt{global}, \opt{type}, \opt{entry}, \opt{namelist} and \opt{name} are supported, the first three of which are equivalent to defining the option with \cmd{DeclareBibliographyOption}, \cmd{DeclareTypeOption} and \cmd{DeclareEntryOption}, respectively.

\end{ltxsyntax}

\subsubsection{Bibliography Environments}
\label{aut:bbx:env}

Apart from defining bibliography drivers, the bibliography style is also responsible for the environments which control the layout of the bibliography and bibliography lists. These environments are defined with \cmd{defbibenvironment}. By default, \cmd{printbibliography} uses the environment \opt{bibliography}. Here is a definition suitable for a bibliography style which does not print any labels in the bibliography:

\begin{ltxexample}
\defbibenvironment{bibliography}
  {\list
     {}
     {\setlength{\leftmargin}{\bibhang}%
      \setlength{\itemindent}{-\leftmargin}%
      \setlength{\itemsep}{\bibitemsep}%
      \setlength{\parsep}{\bibparsep}}}
  {\endlist}
  {\item}
\end{ltxexample}
%
This definition employs a \env{list} environment with hanging indentation, using the \cmd{bibhang} length register provided by \biblatex. It allows for a certain degree of configurability by using \cmd{bibitemsep} and \cmd{bibparsep}, two length registers provided by \biblatex for this very purpose (see \secref{aut:fmt:len}). The \texttt{authoryear} and \texttt{authortitle} bibliography styles use a definition similar to this example.

\begin{ltxexample}
\defbibenvironment{bibliography}
  {\list
     {\printfield[labelnumberwidth]{labelnumber}}
     {\setlength{\labelwidth}{\labelnumberwidth}%
      \setlength{\leftmargin}{\labelwidth}%
      \setlength{\labelsep}{\biblabelsep}%
      \addtolength{\leftmargin}{\labelsep}%
      \setlength{\itemsep}{\bibitemsep}%
      \setlength{\parsep}{\bibparsep}}%
      \renewcommand*{\makelabel}[1]{\hss##1}}
  {\endlist}
  {\item}
\end{ltxexample}
%
Some bibliography styles print labels in the bibliography. For example, a bibliography style designed for a numeric citation scheme will print the number of every entry such that the bibliography looks like a numbered list. In the first example, the first argument to \cmd{list} was empty. In this example, we need it to insert the number, which is provided by \biblatex in the \bibfield{labelnumber} field. We also employ several length registers and other facilities provided by \biblatex, see \secref{aut:fmt:ich, aut:fmt:ilc} for details. The \texttt{numeric} bibliography style uses the definition given above. The \texttt{alphabetic} style is similar, except that \textsf{\texttt{labelnumber}} is replaced by \texttt{labelalpha} and \texttt{labelnumberwidth} by \texttt{labelalphawidth}.

Bibliography lists are handled in a similar way. \cmd{printbiblist} uses the environment named after the bibliography list by default. A typical example is given below. See \secref{aut:fmt:ich, aut:fmt:ilc} for details on the length registers and facilities used in this example.

\begin{ltxexample}
\defbibenvironment{shorthand}
  {\list
     {\printfield[shorthandwidth]{shorthand}}
     {\setlength{\labelwidth}{\shorthandwidth}%
      \setlength{\leftmargin}{\labelwidth}%
      \setlength{\labelsep}{\biblabelsep}%
      \addtolength{\leftmargin}{\labelsep}%
      \setlength{\itemsep}{\bibitemsep}%
      \setlength{\parsep}{\bibparsep}%
      \renewcommand*{\makelabel}[1]{##1\hss}}}
  {\endlist}
  {\item}
\end{ltxexample}

\subsubsection{Bibliography Drivers}
\label{aut:bbx:drv}

Before we go over the commands which form the data interface of the \biblatex package, it may be instructive to have a look at the structure of a bibliography driver. Note that the example given below is greatly simplified, but still functional. For the sake of readability, we omit several fields which may be part of a \bibtype{book} entry and also simplify the handling of those which are considered. The main point is to give you an idea of how a driver is structured. For information about the mapping of the \bibtex file format fields to \biblatex's data types, see \secref{bib:fld}.

\begin{ltxexample}
\DeclareBibliographyDriver{book}{%
  \printnames{author}%
  \newunit\newblock
  \printfield{title}%
  \newunit\newblock
  \printlist{publisher}%
  \newunit
  \printlist{location}%
  \newunit
  \printfield{year}%
  \finentry}
\end{ltxexample}
%
The standard bibliography styles employ two bibliography macros \texttt{begentry} and \texttt{finentry}:

\begin{ltxexample}
\DeclareBibliographyDriver{<<entrytype>>}{%
  \usebibmacro{begentry}
  ...
  \usebibmacro{finentry}}
\end{ltxexample}
%
with the default definitions
\begin{ltxexample}
\newbibmacro*{begentry}{}
\newbibmacro*{finentry}{\finentry}
\end{ltxexample}
%
Use of these macros is recommended for easy hooks into the beginning and end of the driver.

Returning to the driver for the \texttt{book} entrytype above, there is still one piece missing: the formatting directives used by \cmd{printnames}, \cmd{printlist}, and \cmd{printfield}. To give you an idea of what a formatting directive looks like, here are some fictional ones used by our sample driver. Field formats are straightforward, the value of the field is passed to the formatting directive as an argument which may be formatted as desired. The following directive will simply wrap its argument in an \cmd{emph} command:

\begin{ltxexample}
\DeclareFieldFormat{title}{\emph{#1}}
\end{ltxexample}
%
List formats are slightly more complex. After splitting up the list into individual items, \biblatex will execute the formatting directive once for every item in the list. The item is passed to the directive as an argument. The separator to be inserted between the individual items in the list is also handled by the corresponding directive, hence we have to check whether we are in the middle of the list or at the end when inserting it.

\begin{ltxexample}
\DeclareListFormat{location}{%
  #1%
  \ifthenelse{\value{listcount}<\value{liststop}}
    {\addcomma\space}
    {}}
\end{ltxexample}
%
Formatting directives for names are similar to those for literal lists.

Names depend on the datamodel constant <nameparts> which has the default definition:

\begin{ltxexample}
\DeclareDatamodelConstant[type=list]{nameparts}
                                    {prefix,family,suffix,given}
\end{ltxexample}
%
This can be customised to add more name parts to deal with things like patronymics (see the example file \file{93-nameparts.tex}). This needs an extended name format for data sources since the standard \bibtex name format is very limited. \biblatexml (\secref{apx:biblatexml}) handles this natively and there is an extended name format which can handle custom nameparts when using \biber (see \secref{use:enf}).

Inside name formats, the <nameparts> constant declaration makes available two or three macros for each name part defined in the datamodel:

\begin{ltxexample}
\namepart<namepart>   \% The full <namepart>
\namepart<namepart>i  \% The initials of the <namepart>
\namepart<namepart>un \% Numeric value indicating uniqueness level for <namepart>
\end{ltxexample}
%
\cmd{namepart<namepart>un} only exists if the package option \opt{uniquename} is not set to <false> and can take the following values.

\begin{argumentlist}{00}
\item[0] <namepart> was not used in disambiguating the name (because \opt{disambiguation=none} was set in \cmd{DeclareUniquenameTemplate}, see \secref{aut:cav:amb}). In this case the style should decide what to print for this <namepart>
\item[1] Initials only should be printed for <namepart> to ensure uniqueness according to the \opt{uniquename} package option setting
\item[2] The full <namepart> should be printed to ensure uniqueness according to the \opt{uniquename} package option setting
\end{argumentlist}

Note these per-namepart uniqueness macros are essentially an override of the \opt{uniquename} value (see \secref{aut:aux:tst}) for the name as a whole. Styles can choose to use either the less granular \opt{uniquename} value or the more detailed per-namepart values. Usually the general \opt{uniquename} value is enough for ordinary Western names but the more granular information per-namepart is provided to allow sophisticated name uniqueness processing for more complex name schemata.

The name formatting directive is executed once for each name in the name list. Here is a simplified example---the standard name formats are more intricate:

\begin{ltxexample}
\DeclareNameFormat{author}{%
  \ifthenelse{\value{listcount}=1}
    {\namepartfamily%
     \ifdefvoid{\namepartgiven}{}{\addcomma\space\namepartgiven}}
    {\ifdefvoid{\namepartgiven}{}{\namepartgiven\space}%
     \namepartfamily}%
  \ifthenelse{\value{listcount}<\value{liststop}}
    {\addcomma\space}
    {}}
\end{ltxexample}
%
The above directive reverses the name of the first author («Family, Given») and prints the remaining names in their regular sequence («Given Family»). Note that the only component which is guaranteed to be available is the family name, hence we have to check which parts of the name are actually present. If a certain name part is not available, the corresponding macro will be empty. As with directives for literal lists, the separator to be inserted between the individual items in the name list is also handled by the formatting directive, hence we have to check whether we are in the middle of the list or at the end when inserting it. This is what the second \cmd{ifthenelse} test does. See also \secref{aut:bib:fmt}.

A similar output that also respects the \cmd{multinamedelim} and \cmd{finalnamedelim} commands as well as the <prefix> and <suffix> name parts can be achieved with
\begin{ltxexample}
\DeclareNameAlias{author}{family-given/given-family}
\end{ltxexample}

\subsubsection{Special Fields}
\label{aut:bbx:fld}

The following lists and fields are used by \biblatex to pass data to bibliography drivers and citation commands. They are not used in \file{bib} files but defined automatically by the package. From the perspective of a bibliography or citation style, they are not different from the fields in a \file{bib} file.

\paragraph{Generic Fields}
\label{aut:bbx:fld:gen}

\begin{fieldlist}

\fielditem{$<$datetype$>$dateunspecified}{string}

If $<$datetype$>$date held an \acr{ISO8601-2} 4.3 <unspecified>, this field will be set to one of \opt{yearindecade}, \opt{yearincentury}, \opt{monthinyear}, \opt{dayinmonth} or \opt{dayinyear} which specifies the granularity of the unspecified information. These strings can be tested for and along with the date ranges which are automatically created for such <unspecified> dates, a style may choose to format the date in a special way. See \secref{bib:use:dat}. For example, an entry with dates such as:

\begin{lstlisting}[style=bibtex]{}
@book{key,
  date     = {19uu},
  origdate = {199u}
}
\end{lstlisting}
%
would result in the same information in the \file{.bbl} as:
\begin{lstlisting}[style=bibtex]{}
@book{key,
  date     = {1900/1999},
  origdate = {1990/1999}
}
\end{lstlisting}
%
but would additionally have the field \bibfield{dateunspecified} set to <yearincentury> and \bibfield{origdateunspecified} set to <yearindecade>. This information could be used to render the \bibfield{date} as perhaps <20th century> and \bibfield{origdate} as <The 1990s>, information which cannot be derived from the date ranges alone. Since such auto-generated ranges have known values, given the <unspecified> meta-information, it is relatively easy to use the range values to format special cases. While the standard styles do not do this, examples are given in the file \file{96-dates.tex}.

\fielditem{entrykey}{string}

The entry key of an item in the \file{bib} file. This is the string used by \biblatex and the backend to identify an entry in the \file{bib} file.

Note that the set of characters allowed and usable in the string for \bibfield{entrykey} depends on the backend (\biber, \bibtex) as well as the \latex engine (\pdflatex, \lualatex, \xelatex).
Generally, ASCII-letters (\texttt{a-z}, \texttt{A-Z}) and numbers (\texttt{0-9}) are safe, so are the punctuation characters full stop (\texttt{.}) and solidus (\texttt{/}). The punctuation characters \texttt{-\_:;!?} are also safe even if they are made active by \sty{babel}/\sty{polyglossia}. If a Unicode engine is used, non-ASCII characters are also acceptable.
Curly braces (\texttt{\{\}}), commas, spaces, backslashes (\texttt{\textbackslash}), hashes (\texttt{\#}), percent characters (\texttt{\%}) and tildes (\texttt{\textasciitilde}) are always forbidden. \biber additionally forbids round brackets (\texttt{()}), quotation marks (\texttt{\textquotedbl}, \texttt{\textquotesingle}), and the equals sign (\texttt{=}).
The \bibfield{entrykey} is case sensitive, but it is not recommended to exploit that fact too much by introducing two different entries whose key differs only in capitalisation (\eg\ \texttt{sigfridsson} and \texttt{Sigfridsson}).
For full portability it is advisable to stick to a scheme of lowercase (and if so desired uppercase) ASCII-letters, numbers and a small set of acceptable punctuation characters, say \texttt{.:-}.

\fielditem{childentrykey}{string}

This field is no longer necessary or recommended.\DeprecatedMark For backwards
compatibility, it is merely a copy of the \bibfield{entrykey} field in any
set children.

\fielditem{labelnamesource}{literal}

Holds the name of the field used to populate \bibfield{labelname},
determined by \cmd{DeclareLabelname}.

\fielditem{labeltitlesource}{literal}

Holds the name of the field used to populate \bibfield{labeltitle},
determined by \cmd{DeclareLabeltitle}.

\fielditem{labeldatesource}{literal}

Holds one of:

\begin{itemize}
\item The prefix coming before <date> of the date field name chosen by
  \cmd{DeclareLabeldate}
\item The name of a field
\item A literal or localisation string
\end{itemize}
%
Normally holds the prefix coming before <date> of the date field name chosen by \cmd{DeclareLabeldate}. For example, if the labeldate field is \bibfield{eventdate}, then \bibfield{labeldatesource} will be <event>. In case \cmd{DeclareLabeldate} selects the \bibfield{date} field, then \bibfield{labeldatesource} will be defined but will be an empty string as the prefix coming before <date> in the date label name is empty. This is so that the contents of \bibfield{labeldatesource} can be used in constructing references to the field which \cmd{DeclareLabeldate} chooses. Since \cmd{DeclareLabeldate} can also select literal strings for fallbacks, \bibfield{labeldatesource} may not refer to a field or may be undefined. Bear in mind that \cmd{DeclareLabeldate} can also be used to select non-date fields as a fallback and so \bibfield{labeldatesource} might contain a field name. So, in summary, the rules are

\begin{ltxexample}
\iffieldundef{labeldatesource}
  {}% labeldate package option is not set
  {\iffieldundef{\thefield{labeldatesource}year}
    % \DeclareLabeldate resolved to either a literal/localisation
    % string or a non-date field since
    % if a date is defined by a date field, there is
    % at least a year
    {\iffieldundef{\thefield{labeldatesource}}
       {}% \DeclareLabeldate resolved to a literal/localisation string
       {}% \DeclareLabeldate resolved to a non-date field
    }
    {} % \DeclareLabeldate resolved a date field name prefix like "" or "orig"
  }
\end{ltxexample}

\fielditem{entrytype}{string}

The entry type (\bibtype{book}, \bibtype{inbook}, etc.), given in lowercase letters.

\fielditem{childentrytype}{string}

This field is no longer necessary or recommended.\DeprecatedMark For backwards
compatibility, it is merely a copy of the \bibfield{entrytype} field in any
set children.

\fielditem{entrysetcount}{integer}

This field holds an integer indicating the position of a set member in the entry set (starting at \texttt{1}). This field is only available in the subentries of an entry set.

\fielditem{hash}{string}

This field is special in that it is only available locally in name formatting directives. It holds a hash string which uniquely identifies individual names in a name list. This information is available for all names in all name lists. See also \bibfield{namehash} and \bibfield{fullhash}.

\fielditem{namehash}{string}

A hash string which uniquely identifies the \bibfield{labelname} list. This is useful for recurrence checks. For example, a citation style which replaces recurrent authors or editors with a string like <idem> could save the \bibfield{namehash} field with \cmd{savefield} and use it in a comparison with \cmd{iffieldequals} later (see \secref{aut:aux:dat, aut:aux:tst}). The \bibfield{namehash} is derived from the truncated \bibfield{labelname} list, \ie it is responsive to \opt{maxcitenames} and \opt{mincitenames}. See also \bibfield{hash} and \bibfield{fullhash}.

\fielditem{bibnamehash}{string}

As \bibfield{namehash} but responsive to \opt{maxbibnames} and
\opt{minbibnames}. This is not used in standard styles but may be used to
make tests in bibliography lists (such as those used to determine whether
dashes should replace repeated authors) behave differently.

\fielditem{$<$namelist$>$namehash}{string}

As \bibfield{namehash} for the name list called <namelist>.

\fielditem{$<$namelist$>$bibnamehash}{string}

As \bibfield{bibnamehash} for the name list called <namelist>.

\fielditem{fullhash}{string}

A hash string which uniquely identifies the \bibfield{labelname} list. This fields differs from \bibfield{namehash} in two details: 1) The \bibfield{shortauthor} and \bibfield{shorteditor} lists are ignored when generating the hash. 2) The hash always refers to the full list, ignoring \opt{maxnames} and \opt{minnames}. See also \bibfield{hash} and \bibfield{namehash}.

\fielditem{$<$namelist$>$fullhash}{string}

As \bibfield{fullhash} for the name list called <namelist>.

\listitem{pageref}{literal}

If the \opt{backref} package option is enabled, this list holds the page numbers of the pages on which the respective bibliography entry is cited. If there are \env{refsection} environments in the document, the back references are local to the reference sections.

\fielditem{sortinit}{literal}

This field holds the initial character of the information used during sorting.

\fielditem{sortinithash}{string}

This field holds a hash of the (locale-specific) Unicode Collation Algorithm primary weight of the first extended grapheme cluster (essentially the first character) of the string used during sorting. This is useful when subdividing the bibliography alphabetically and is used internally by \cmd{bibinitsep} (see \secref{use:fmt:len}).

\fielditem{clonesourcekey}{string}

This field holds the entry key of the entry from which an entry was cloned. Clones are created for
entries which are mentioned in \bibfield{related} fields as part of related entry processing, for example.

\fielditem{urlraw}{verbatim}

This is the unencoded, raw version of any \bibfield{url}. This is intended for use when the display version and clickable link version of a URL are different. This can be the case when the URL contains special or Unicode characters. In the case where no such characters occur, may be identical to the \bibfield{url}.
\end{fieldlist}

\paragraph{Fields for Use in Citation Labels}
\label{aut:bbx:fld:lab}

\begin{fieldlist}

\fielditem{labelalpha}{literal}

A label similar to the labels generated by the \path{alpha.bst} style of traditional \bibtex. This default label consists of initials drawn from the \bibfield{labelname} list plus the last two digits of the publication year. The \bibfield{label} field may be used to override its non"=numeric portion. If the \bibfield{label} field is defined, \biblatex will use its value and append the last two digits of the publication year when generating \bibfield{labelalpha}. The \bibfield{shorthand} field may be used to override the entire label. If defined, \bibfield{labelalpha} is the \bibfield{shorthand} rather than an automatically generated label. Users can specify a template used to construct the alphabetic label (see \secref{aut:ctm:lab}) and the default template mirrors the format mentioned for bibtex above. A complete <alphabetic> label consists of the fields \bibfield{labelalpha} plus \bibfield{extraalpha}. Note that the \bibfield{labelalpha} and \bibfield{extraalpha} fields need to be requested with the package option \opt{labelalpha} (\secref{use:opt:pre:int}). See also \bibfield{extraalpha} as well as \cmd{labelalphaothers} in \secref{use:fmt:fmt}.

\fielditem{extraalpha}{integer}

The <alphabetic> citation scheme usually requires a letter to be appended to the label if the bibliography contains two or more works by the same author which were all published in the same year. In this case, the \bibfield{extraalpha} field holds an integer which may be converted to a letter with \cmd{mknumalph} or formatted in some other way. This field is similar to the role of \bibfield{extradate} in the author"=year scheme. A complete <alphabetic> label consists of the fields \bibfield{labelalpha} plus \bibfield{extraalpha}. Note that the \bibfield{labelalpha} and \bibfield{extraalpha} fields need to be requested with the package option \opt{labelalpha}, see \secref{use:opt:pre:int} for details. See also \bibfield{labelalpha} as well as \cmd{labelalphaothers} in \secref{use:fmt:fmt}. Table \ref{use:opt:tab1} summarises the various \opt{extra*} disambiguation counters and what they track.

\listitem{labelname}{name}

The name to be printed in citations. This list is a copy of either the \bibfield{shortauthor}, the \bibfield{author}, the \bibfield{shorteditor}, the \bibfield{editor}, or the \bibfield{translator} list, which are normally checked for in this order. If no authors and editors are available, this list is undefined. Note that this list is also responsive to the \opt{use$<$name$>$}, options, see \secref{use:opt:bib}. Citation styles should use this list when printing the name in a citation. This list is provided for convenience only and does not carry any additional meaning.
This field may be customized. See \secref{aut:ctm:fld} for details.

\fielditem{extraname}{integer}

Holds a count of the number of bibliography entries within a refsection which derive from the same \bibfield{labelname} list. This counter takes account of \opt{uniquename} settings (see \secref{use:opt:pre:int}). While not used by any standard styles, this field is useful in styles which wish to number bibliography entries on a per-\bibfield{labelname} basis. This field will only exist if there is a \bibfield{labelname}. The \bibfield{extraname} counter is related to, but conceptually different from \cmd{ifsingletitle} (see \secref{use:opt:pre:int} and \secref{aut:aux:tst}).

\fielditem{labelnumber}{literal}

The number of the bibliography entry, as required by numeric citation schemes. If the \bibfield{shorthand} field is defined, \biblatex does not assign a number to the respective entry. In this case \bibfield{labelnumber} is the shorthand rather than a number. Numeric styles must use the value of this field instead of a counter. Note that this field needs to be requested with the package option \opt{labelnumber}, see \secref{use:opt:pre:int} for details. Also see the package option \opt{defernumbers} in \secref{use:opt:pre:gen}.

\fielditem{labelprefix}{literal}

If the \opt{labelprefix} option of \cmd{newrefcontext} has been set in order to prefix all entries in a subbibliography with a fixed string, this string is available in the \bibfield{labelprefix} field of all affected entries. If no prefix has been set, the \bibfield{labelprefix} field of the respective entry is undefined. See the \opt{labelprefix} option of \cmd{newrefcontext} in \secref{use:bib:context} for details. If the \bibfield{shorthand} field is defined, \biblatex does not assign the prefix to the \bibfield{labelprefix} field of the respective entry. In this case, the \bibfield{labelprefix} field is undefined.

\fielditem{labeltitle}{literal}

The printable title of a work. In some circumstances, a style might need to choose a title from a list of a possible title fields. For example, citation styles printing short titles may want to print the \bibfield{shorttitle} field if it exists but otherwise print the \bibfield{title} field. The list of fields to be considered when constructing \bibfield{labeltitle} may be customized. See \secref{aut:ctm:fld} for details. Note that the \bibfield{extratitle} field needs to be requested with the package option \opt{labeltitle}, see \secref{use:opt:pre:int} for details. See also \bibfield{extratitle}. Note also that the \bibfield{extratitleyear} field needs to be requested with the package option \opt{labeltitleyear}. See also \bibfield{extratitleyear}.

\fielditem{extratitle}{integer}

It is sometimes useful, for example in author"=title citation schemes, to be able to disambiguate works with the same title. For works by the same \bibfield{labelname} with the same \bibfield{labeltitle}, the \bibfield{extratitle} field holds an integer which may be converted to a letter with \cmd{mknumalph} or formatted in some other way (or it can be merely used as a flag to say that some other field such as a date should be used in conjunction with the \bibfield{labeltitle} field). This field is undefined if there is only one work with the same \bibfield{labeltitle} by the same \bibfield{labelname} in the bibliography. Note that the \bibfield{extratitle} field needs to be requested with the package option \opt{labeltitle}, see \secref{use:opt:pre:int} for details. See also \bibfield{labeltitle}. Table \ref{use:opt:tab1} summarises the various \opt{extra*} disambiguation counters and what they track.

\fielditem{extratitleyear}{integer}

It is sometimes useful, for example in author"=title citation schemes, to be able to disambiguate works with the same title in the same year but with no author. For works with the same \bibfield{labeltitle} and with the same \bibfield{labelyear}, the \bibfield{extratitleyear} field holds an integer which may be converted to a letter with \cmd{mknumalph} or formatted in some other way (or it can be merely used as a flag to say that some other field such as a publisher should be used in conjunction with the \bibfield{labelyear} field). This field is undefined if there is only one work with the same \bibfield{labeltitle} and \bibfield{labelyear} in the bibliography. Note that the \bibfield{extratitleyear} field needs to be requested with the package option \opt{labeltitleyear}, see \secref{use:opt:pre:int} for details. See also \bibfield{labeltitleyear}. Table \ref{use:opt:tab1} summarises the various \opt{extra*} disambiguation counters and what they track.

\fielditem{labelyear}{literal}

The year of the date field selected by \cmd{DeclareLabeldate} (\secref{aut:ctm:fld}) or the \bibfield{year} field, for use in author-year labels. A complete author-year label consists of the fields \bibfield{labelyear} plus \bibfield{extradate}. Note that the \bibfield{labelyear} and \bibfield{extradate} fields need to be requested with the package option \opt{labeldateparts}, see \secref{use:opt:pre:int} for details. See also \bibfield{extradate}.

\fielditem{labelendyear}{literal}

The end year of the date field selected by \cmd{DeclareLabeldate} (\secref{aut:ctm:fld}) if the selected date is a range.

\fielditem{labelmonth}{datepart}

The month of the date field selected by \cmd{DeclareLabeldate} (\secref{aut:ctm:fld}), or the \bibfield{month} field for use in author-year labels. Note that the \bibfield{labelmonth} field needs to be requested with the package option \opt{labeldateparts}, see \secref{use:opt:pre:int} for details.

\fielditem{labelendmonth}{datepart}

The end month of the date field selected by \cmd{DeclareLabeldate} (\secref{aut:ctm:fld}) if the selected date is a range.

\fielditem{labelday}{datepart}

The month of the date field selected by \cmd{DeclareLabeldate} (\secref{aut:ctm:fld}) for use in author-year labels. Note that the \bibfield{labelday} field needs to be requested with the package option \opt{labeldateparts}, see \secref{use:opt:pre:int} for details.

\fielditem{labelendday}{datepart}

The end day of the date field selected by \cmd{DeclareLabeldate} (\secref{aut:ctm:fld}) if the selected date is a range.

\fielditem{extradate}{integer}

The author"=year citation scheme usually requires a letter to be appended to the year if the bibliography contains two or more works by the same author which were all published in the same year. In this case, the \bibfield{extradate} field holds an integer which may be converted to a letter with \cmd{mknumalph} or formatted in some other way. This field is undefined if there is only one work by the author in the bibliography or if all works by the author have different publication years. A complete author-year label consists of the fields \bibfield{labelyear} plus \bibfield{extradate}. Note that the \bibfield{labelyear} and \bibfield{extradate} fields need to be requested with the package option \opt{labeldateparts}, see \secref{use:opt:pre:int} for details. See also \bibfield{labelyear}. Table \ref{use:opt:tab1} summarises the various \opt{extra*} disambiguation counters and what they track.

\fielditem{extradatescope}{literal}

This field contains the name of the most specific field which determined the value of \bibfield{extradate}. It is not used by the standard styles but may be useful in controlling the placement of the \bibfield{extradate} field value.

\end{fieldlist}

\paragraph{Date Component Fields}
\label{aut:bbx:fld:dat}

Note that it is possible to define new date fields in the datamodel which behave exactly like the default data model date fields described in this section.

See \tabref{aut:bbx:fld:tab1} for an overview of how the date fields in \file{bib} files are related to the date fields provided by the style interface. When testing for a field like \bibfield{origdate} in a style, use code like:

\begin{ltxcode}
<<\iffieldundef>>{orig<<year>>}{...}{...}
\end{ltxcode}
%
This will tell you if the corresponding date is defined at all. This test:

\begin{ltxcode}
<<\iffieldundef>>{orig<<endyear>>}{...}{...}
\end{ltxcode}
%
will tell you if the corresponding date is defined and a (fully specified) range. This test:

\begin{ltxcode}
<<\iffieldequalstr>>{orig<<endyear>>}{}{...}{...}
\end{ltxcode}
%
will tell you if the corresponding date is defined and an open"=ended range. Open"=ended ranges are indicated by an empty \texttt{endyear} component (as opposed to an undefined \texttt{endyear} component). See \secref{bib:use:dat} and \tabref{bib:use:tab1} on page~\pageref{bib:use:tab1} for further examples.

\begingroup
\tablesetup
\begin{longtable}[l]{%
  @{}V{0.15\textwidth}%
  @{}V{0.4\textwidth}%
  @{}V{0.3\textwidth}%
  @{}V{0.2\textwidth}@{}}
\caption{Date Interface}
\label{aut:bbx:fld:tab1}
\endfirsthead
\caption[]{Date Interface (cont'd)}
\endhead
\toprule
\multicolumn{2}{@{}H}{\file{bib} File} &
\multicolumn{2}{H}{Data Interface} \\
\cmidrule(r){1-2}\cmidrule(l){3-4}
\multicolumn{1}{@{}H}{Field} &
\multicolumn{1}{H}{Value (Example)} &
\multicolumn{1}{H}{Field} &
\multicolumn{1}{H}{Value (Example)} \\
\cmidrule{1-1}\cmidrule(lr){2-2}\cmidrule(l){3-3}\cmidrule(l){4-4}
date		& 1988			& day		& undefined \\
		&			& month		& undefined \\
		&			& year		& 1988 \\
		&			& season  & undefined \\
		&			& endday	& undefined \\
		&			& endmonth	& undefined \\
		&			& endyear	& undefined \\
		&			& endseason  & undefined \\
		&			& hour	& undefined \\
		&			& minute	& undefined \\
		&			& second	& undefined \\
		&			& timezone & undefined \\
		&			& endhour	& undefined \\
		&			& endminute	& undefined \\
		&			& endsecond	& undefined \\
		&			& endtimezone & undefined \\
date		& 1997/			& day		& undefined \\
		&			& month		& undefined \\
		&			& year		& 1997 \\
		&			& season  & undefined \\
		&			& endday	& undefined \\
		&			& endmonth	& undefined \\
		&			& endyear	& empty \\
		&			& endseason  & undefined \\
		&			& hour	& undefined \\
		&			& minute	& undefined \\
		&			& second	& undefined \\
		&			& timezone & undefined \\
		&			& endhour	& undefined \\
		&			& endminute	& undefined \\
		&			& endsecond	& undefined \\
		&			& endtimezone & undefined \\
urldate		& 2009-01-31		& urlday	& 31 \\
		&			& urlmonth	& 01 \\
		&			& urlyear	& 2009 \\
		&			& urlseason  & undefined \\
		&			& urlendday	& undefined \\
		&			& urlendmonth	& undefined \\
		&			& urlendyear	& undefined \\
		&			& urlendseason  & undefined \\
		&			& urlhour	& undefined \\
		&			& urlminute	& undefined \\
		&			& urlsecond	& undefined \\
		&			& urltimezone & undefined \\
		&			& urlendhour	& undefined \\
		&			& urlendminute	& undefined \\
		&			& urlendsecond	& undefined \\
		&			& urlendtimezone & undefined \\
urldate		& 2009-01-31T15:34:04Z		& urlday	& 31 \\
		&			& urlmonth	& 01 \\
		&			& urlyear	& 2009 \\
		&			& urlseason  & undefined \\
		&			& urlendday	& undefined \\
		&			& urlendmonth	& undefined \\
		&			& urlendyear	& undefined \\
		&			& urlendseason  & undefined \\
		&			& urlhour	& 15 \\
		&			& urlminute	& 34 \\
		&			& urlsecond	& 04 \\
		&			& urltimezone & Z \\
		&			& urlendhour	& undefined \\
		&			& urlendminute	& undefined \\
		&			& urlendsecond	& undefined \\
		&			& urlendtimezone & undefined \\
urldate		& 2009-01-31T15:34:04+05:00		& urlday	& 31 \\
		&			& urlmonth	& 01 \\
		&			& urlyear	& 2009 \\
		&			& urlseason  & undefined \\
		&			& urlendday	& undefined \\
		&			& urlendmonth	& undefined \\
		&			& urlendyear	& undefined \\
		&			& urlendseason  & undefined \\
		&			& urlhour	& 15 \\
		&			& urlminute	& 34 \\
		&			& urlsecond	& 04 \\
		&			& urltimezone & +0500 \\
		&			& urlendhour	& undefined \\
		&			& urlendminute	& undefined \\
		&			& urlendsecond	& undefined \\
		&			& urlendtimezone & undefined \\
urldate		& \parbox[t]{0.4\textwidth}{2009-01-31T15:34:04/\\2009-01-31T16:04:34}& urlday	& 31 \\
		&			& urlmonth	& 1 \\
		&			& urlyear	& 2009 \\
		&			& urlseason  & undefined \\
		&			& urlendday	& 31 \\
		&			& urlendmonth	& 1 \\
		&			& urlendyear	& 2009 \\
		&			& urlendseason  & undefined \\
		&			& urlhour	& 15 \\
		&			& urlminute	& 34 \\
		&			& urlsecond	& 4 \\
		&			& urltimezone & floating \\
		&			& urlendhour	& 16 \\
		&			& urlendminute	& 4 \\
		&			& urlendsecond	& 34 \\
		&			& urlendtimezone & floating \\
origdate	& 2002-21/2002-23	& origday	& undefined \\
		&			& origmonth	& 01 \\
		&			& origyear	& 2002 \\
		&			& origseason  & spring \\
		&			& origendday	& undefined \\
		&			& origendmonth	& 02 \\
		&			& origendyear	& 2002 \\
		&			& origendseason  & autumn \\
		&			& orighour	& undefined \\
		&			& origminute	& undefined \\
		&			& origsecond	& undefined \\
		&			& origtimezone & undefined \\
		&			& origendhour	& undefined \\
		&			& origendminute	& undefined \\
		&			& origendsecond	& undefined \\
		&			& origendtimezone & undefined \\
eventdate	& 1995-01-31/1995-02-05	& eventday	& 31 \\
		&			& eventmonth	& 01 \\
		&			& eventyear	& 1995 \\
		&			& eventseason  & undefined \\
		&			& eventendday	& 05 \\
		&			& eventendmonth	& 02 \\
		&			& eventendyear	& 1995 \\
		&			& eventendseason  & undefined \\
		&			& eventhour	& undefined \\
		&			& eventminute	& undefined \\
		&			& eventsecond	& undefined \\
		&			& eventtimezone & undefined \\
		&			& eventendhour	& undefined \\
		&			& eventendminute	& undefined \\
		&			& eventendsecond	& undefined \\
		&			& eventendtimezone & undefined \\
\bottomrule
%\end{tabularx}
\end{longtable}
\endgroup
\begin{fieldlist}

\fielditem{hour}{datepart}

This field holds the hour component of the \bibfield{date} field. If the date is a range, it holds the hour component of the start date.

\fielditem{minute}{datepart}

This field holds the minute component of the \bibfield{date} field. If the date is a range, it holds the minute component of the start date.

\fielditem{second}{datepart}

This field holds the second component of the \bibfield{date} field. If the date is a range, it holds the second component of the start date.

\fielditem{timezone}{datepart}

This field holds the timezone component of the \bibfield{date} field. If the date is a range, it holds the timezone component of the start date.

\fielditem{day}{datepart}

This field holds the day component of the \bibfield{date} field. If the date is a range, it holds the day component of the start date.

\fielditem{month}{datepart}

This field is the \bibfield{month} as given in the database file or it holds the month component of the \bibfield{date} field. If the date is a range, it holds the month component of the start date.

\fielditem{year}{datepart}

This field is the \bibfield{year} as given in the database file or it holds the year component of the \bibfield{date} field. If the date is a range, it holds the year component of the start date.

\fielditem{season}{datepart}

This field holds the season component of the \bibfield{date} field as specified by \acr{ISO8601-2} 4.7 (\secref{bib:use:dat}). It contains a season localisation string (\secref{aut:lng:key:dt}). If the date is a range, it holds the season component of the start date.

\fielditem{endhour}{datepart}

If the date specification in the \bibfield{date} field is a range, this field holds the hour component of the end date.

\fielditem{endminute}{datepart}

If the date specification in the \bibfield{date} field is a range, this field holds the minute component of the end date.

\fielditem{endsecond}{datepart}

If the date specification in the \bibfield{date} field is a range, this field holds the second component of the end date.

\fielditem{endtimezone}{datepart}

If the date specification in the \bibfield{date} field is a range, this field holds the timezone component of the end date.

\fielditem{endday}{datepart}

If the date specification in the \bibfield{date} field is a range, this field holds the day component of the end date.

\fielditem{endmonth}{datepart}

If the date specification in the \bibfield{date} field is a range, this field holds the month component of the end date.

\fielditem{endyear}{datepart}

If the date specification in the \bibfield{date} field is a range, this field holds the year component of the end date. A blank (but defined) \bibfield{endyear} component indicates an open ended \bibfield{date} range.

\fielditem{endseason}{datepart}

If the date specification in the \bibfield{date} field is a range, this field holds the season component of the end date as specified by \acr{ISO8601-2} 4.7 (\secref{bib:use:dat}). It contains a season localisation string (\secref{aut:lng:key:dt}). A blank (but defined) \bibfield{endseason} component indicates an open ended \bibfield{date} range.

\fielditem{orighour}{datepart}

This field holds the hour component of the \bibfield{origdate} field. If the date is a range, it holds the hour component of the start date.

\fielditem{origminute}{datepart}

This field holds the minute component of the \bibfield{origdate} field. If the date is a range, it holds the minute component of the start date.

\fielditem{origsecond}{datepart}

This field holds the second component of the \bibfield{origdate} field. If the date is a range, it holds the second component of the start date.

\fielditem{origtimezone}{datepart}

This field holds the timezone component of the \bibfield{origdate} field. If the date is a range, it holds the timezone component of the start date.

\fielditem{origday}{datepart}

This field holds the day component of the \bibfield{origdate} field. If the date is a range, it holds the day component of the start date.

\fielditem{origmonth}{datepart}

This field holds the month component of the \bibfield{origdate} field. If the date is a range, it holds the month component of the start date.

\fielditem{origyear}{datepart}

This field holds the year component of the \bibfield{origdate} field. If the date is a range, it holds the year component of the start date.

\fielditem{origseason}{datepart}

This field holds the season component of the \bibfield{origdate} field as specified by \acr{ISO8601-2} 4.7 (\secref{bib:use:dat}). It contains a season localisation string (\secref{aut:lng:key:dt}). If the date is a range, it holds the season component of the start date.

\fielditem{origendhour}{datepart}

If the date specification in the \bibfield{origdate} field is a range, this field holds the hour component of the end date.

\fielditem{origendminute}{datepart}

If the date specification in the \bibfield{origdate} field is a range, this field holds the minute component of the end date.

\fielditem{origendsecond}{datepart}

If the date specification in the \bibfield{origdate} field is a range, this field holds the second component of the end date.

\fielditem{origendtimezone}{datepart}

If the date specification in the \bibfield{origdate} field is a range, this field holds the timezone component of the end date.

\fielditem{origendday}{datepart}

If the date specification in the \bibfield{origdate} field is a range, this field holds the day component of the end date.

\fielditem{origendmonth}{datepart}

If the date specification in the \bibfield{origdate} field is a range, this field holds the month component of the end date.

\fielditem{origendyear}{datepart}

If the date specification in the \bibfield{origdate} field is a range, this field holds the year component of the end date. A blank (but defined) \bibfield{origendyear} component indicates an open ended \bibfield{origdate} range.

\fielditem{origendseason}{datepart}

If the date specification in the \bibfield{origdate} field is a range, this field holds the season component of the end date as specified by \acr{ISO8601-2} 4.7 (\secref{bib:use:dat}). It contains a season localisation string (\secref{aut:lng:key:dt}). A blank (but defined) \bibfield{origendseason} component indicates an open ended \bibfield{origdate} range.

\fielditem{eventhour}{datepart}

This field holds the hour component of the \bibfield{eventdate} field. If the date is a range, it holds the hour component of the start date.

\fielditem{eventminute}{datepart}

This field holds the minute component of the \bibfield{eventdate} field. If the date is a range, it holds the minute component of the start date.

\fielditem{eventsecond}{datepart}

This field holds the second component of the \bibfield{eventdate} field. If the date is a range, it holds the second component of the start date.

\fielditem{eventtimezone}{datepart}

This field holds the timezone component of the \bibfield{eventdate} field. If the date is a range, it holds the timezone component of the start date.

\fielditem{eventday}{datepart}

This field holds the day component of the \bibfield{eventdate} field. If the date is a range, it holds the day component of the start date.

\fielditem{eventmonth}{datepart}

This field holds the month component of the \bibfield{eventdate} field. If the date is a range, it holds the month component of the start date.

\fielditem{eventyear}{datepart}

This field holds the year component of the \bibfield{eventdate} field. If the date is a range, it holds the year component of the start date.

\fielditem{eventseason}{datepart}

This field holds the season component of the \bibfield{eventdate} field as specified by \acr{ISO8601-2} 4.7 (\secref{bib:use:dat}). It contains a season localisation string (\secref{aut:lng:key:dt}). If the date is a range, it holds the season component of the start date.

\fielditem{eventendhour}{datepart}

If the date specification in the \bibfield{eventdate} field is a range, this field holds the hour component of the end date.

\fielditem{eventendminute}{datepart}

If the date specification in the \bibfield{eventdate} field is a range, this field holds the minute component of the end date.

\fielditem{eventendsecond}{datepart}

If the date specification in the \bibfield{eventdate} field is a range, this field holds the second component of the end date.

\fielditem{eventendtimezone}{datepart}

If the date specification in the \bibfield{eventdate} field is a range, this field holds the timezone component of the end date.

\fielditem{eventendday}{datepart}

If the date specification in the \bibfield{eventdate} field is a range, this field holds the day component of the end date.

\fielditem{eventendmonth}{datepart}

If the date specification in the \bibfield{eventdate} field is a range, this field holds the month component of the end date.

\fielditem{eventendyear}{datepart}

If the date specification in the \bibfield{eventdate} field is a range, this field holds the year component of the end date. A blank (but defined) \bibfield{eventendyear} component indicates an open ended \bibfield{eventdate} range.

\fielditem{eventendseason}{datepart}

If the date specification in the \bibfield{eventdate} field is a range, this field holds the season component of the end date as specified by \acr{ISO8601-2} 4.7 (\secref{bib:use:dat}). It contains a season localisation string (\secref{aut:lng:key:dt}). A blank (but defined) \bibfield{eventendseason} component indicates an open ended \bibfield{eventdate} range.

\fielditem{urlhour}{datepart}

This field holds the hour component of the \bibfield{urldate} field. If the date is a range, it holds the hour component of the start date.

\fielditem{urlminute}{datepart}

This field holds the minute component of the \bibfield{urldate} field. If the date is a range, it holds the minute component of the start date.

\fielditem{urlsecond}{datepart}

This field holds the second component of the \bibfield{urldate} field. If the date is a range, it holds the second component of the start date.

\fielditem{timezone}{urldatepart}

This field holds the timezone component of the \bibfield{urldate} field. If the date is a range, it holds the timezone component of the start date.

\fielditem{urlday}{datepart}

This field holds the day component of the \bibfield{urldate} field.

\fielditem{urlmonth}{datepart}

This field holds the month component of the \bibfield{urldate} field.

\fielditem{urlyear}{datepart}

This field holds the year component of the \bibfield{urldate} field.

\fielditem{urlseason}{datepart}

This field holds the season component of the \bibfield{urldate} field as specified by \acr{ISO8601-2} 4.7 (\secref{bib:use:dat}). It contains a season localisation string (\secref{aut:lng:key:dt}). If the date is a range, it holds the season component of the start date.

\fielditem{urlendhour}{datepart}

If the date specification in the \bibfield{urldate} field is a range, this field holds the hour component of the end date.

\fielditem{urlendminute}{datepart}

If the date specification in the \bibfield{urldate} field is a range, this field holds the minute component of the end date.

\fielditem{urlendsecond}{datepart}

If the date specification in the \bibfield{urldate} field is a range, this field holds the second component of the end date.

\fielditem{urlendtimezone}{datepart}

If the date specification in the \bibfield{urldate} field is a range, this field holds the timezone component of the end date.

\fielditem{urlendday}{datepart}

If the date specification in the \bibfield{urldate} field is a range, this field holds the day component of the end date.

\fielditem{urlendmonth}{datepart}

If the date specification in the \bibfield{urldate} field is a range, this field holds the month component of the end date.

\fielditem{urlendyear}{datepart}

If the date specification in the \bibfield{urldate} field is a range, this field holds the year component of the end date. A blank (but defined) \bibfield{urlendyear} component indicates an open ended \bibfield{urldate} range.

\fielditem{urlendseason}{datepart}

If the date specification in the \bibfield{urldate} field is a range, this field holds the season component of the end date as specified by \acr{ISO8601-2} 4.7 (\secref{bib:use:dat}). It contains a season localisation string (\secref{aut:lng:key:dt}). A blank (but defined) \bibfield{urlendseason} component indicates an open ended \bibfield{urldate} range.

\end{fieldlist}

\subsection{Citation Styles}
\label{aut:cbx}

A citation style is a set of commands such as \cmd{cite} which print different types of citations. Such styles are defined in files with the suffix \file{cbx}. The \biblatex package loads the selected citation style file at the end of the package. Note that a small repertory of frequently used macros shared by several of the standard citation styles is also included in \path{biblatex.def}. This file is loaded at the end of the package as well, prior to the selected citation style. It also contains the definitions of the commands from \secref{use:cit:txt}.

\subsubsection{Citation Style Files}
\label{aut:cbx:cbx}

Before we go over the individual commands available in citation style files, consider this example of the overall structure of a typical \file{cbx} file:

\begin{ltxexample}
\ProvidesFile{example.cbx}[2006/03/15 v1.0 biblatex citation style]

\DeclareCiteCommand{\cite}{...}{...}{...}{...}
\DeclareCiteCommand{\parencite}[\mkbibparens]{...}{...}{...}{...}
\DeclareCiteCommand{\footcite}[\mkbibfootnote]{...}{...}{...}{...}
\DeclareCiteCommand{\textcite}{...}{...}{...}{...}
\endinput
\end{ltxexample}

\begin{ltxsyntax}

\cmditem{RequireCitationStyle}{style}

This command is optional and intended for specialized citation styles built on top of a more generic style. It loads the citation style \path{style.cbx}.

\cmditem{InitializeCitationStyle}{code}

Specifies arbitrary \prm{code} required to initialize or reset the citation style. This hook will be executed once at package load-time and every time the \cmd{citereset} command from \secref{use:cit:msc} is used. The \cmd{citereset} command also resets the internal citation trackers of this package. The reset will affect the \cmd{ifciteseen}, \cmd{ifentryseen}, \cmd{ifciteibid}, and \cmd{ifciteidem} tests discussed in \secref{aut:aux:tst}. When used in a \env{refsection} environment, the reset of the citation tracker is local to the current \env{refsection} environment.

\cmditem{OnManualCitation}{code}

Specifies arbitrary \prm{code} required for a partial reset of the citation style. This hook will be executed every time the \cmd{mancite} command from \secref{use:cit:msc} is used. It is particularly useful in citation styles which replace repeated citations by abbreviations like <ibidem> or <op. cit.> which may get ambiguous if automatically generated and manual citations are mixed. The \cmd{mancite} command also resets the internal <ibidem> and <idem> trackers of this package. The reset will affect the \cmd{ifciteibid} and \cmd{ifciteidem} tests discussed in \secref{aut:aux:tst}.

\cmditem{DeclareCiteCommand}{command}[wrapper]{precode}{loopcode}{sepcode}{postcode} \cmditem*{DeclareCiteCommand*}{command}[wrapper]{precode}{loopcode}{sepcode}{postcode}

This is the core command used to define all citation commands. It takes one optional and five mandatory arguments. The \prm{command} is the command to be defined, for example \cmd{cite}. If the optional \prm{wrapper} argument is given, the entire citation will be passed to the \prm{wrapper} as an argument, \ie the wrapper command must take one mandatory argument.\footnote{Typical examples of wrapper commands are \cmd{mkbibparens} and \cmd{mkbibfootnote}.} The \prm{precode} is arbitrary code to be executed at the beginning of the citation. It will typically handle the \prm{prenote} argument which is available in the \bibfield{prenote} field. It may also be used to initialize macros required by the \prm{loopcode}. The \prm{loopcode} is arbitrary code to be executed for each entry key passed to the \prm{command}. This is the core code which prints the citation labels or any other data. The \prm{sepcode} is arbitrary code to be executed after each iteration of the \prm{loopcode}. It will only be executed if a list of entry keys is passed to the \prm{command}. The \prm{sepcode} will usually insert some kind of separator, such as a comma or a semicolon. The \prm{postcode} is arbitrary code to be executed at the end of the citation. The \prm{postcode} will typically handle the \prm{postnote} argument which is available in the \bibfield{postnote} field.\footnote{The bibliographic data available to the \prm{loopcode} is the data of the entry currently being processed. In addition to that, the data of the first entry is available to the \prm{precode} and the data of the last one is available to the \prm{postcode}. <First> and <last> refer to the order in which the citations are printed. If the \opt{sortcites} package option is active, this is the order of the list after sorting. Note that no bibliographic data is available to the \prm{sepcode}.} The starred variant of \cmd{DeclareCiteCommand} defines a starred \prm{command}. For example, |\DeclareCiteCommand*{cite}| would define |\cite*|.\footnote{Note that the regular variant of \cmd{DeclareCiteCommand} defines a starred version of the \prm{command} implicitly, unless the starred version has been defined before. This is intended as a fallback. The implicit definition is an alias for the regular variant.}

\cmditem{DeclareMultiCiteCommand}{command}[wrapper]{cite}{delimiter}

This command defines <multicite> commands (\secref{use:cit:mlt}). The \prm{command} is the multicite command to be defined, for example \cmd{cites}. It is automatically made robust. Multicite commands are built on top of backend commands defined with \cmd{DeclareCiteCommand} and the \prm{cite} argument specifies the name of the backend command to be used. Note that the wrapper of the backend command (\ie the \prm{wrapper} argument passed to \cmd{DeclareCiteCommand}) is ignored. Use the optional \prm{wrapper} argument to specify an alternative wrapper. The \prm{delimiter} is the string to be printed as a separator between the individual citations in the list. This will typically be \cmd{multicitedelim}. The following examples are real definitions taken from \path{biblatex.def}:

\begin{ltxexample}
\DeclareMultiCiteCommand{\cites}%
	{\cite}{\multicitedelim}
\DeclareMultiCiteCommand{\parencites}[\mkbibparens]%
	{\parencite}{\multicitedelim}
\DeclareMultiCiteCommand{\footcites}[\mkbibfootnote]%
	{\footcite}{\multicitedelim}
\end{ltxexample}

\cmditem{DeclareAutoCiteCommand}{name}[position]{cite}{multicite}

This command provides definitions for the \cmd{autocite} and \cmd{autocites} commands from \secref{use:cit:aut}. The definitions are enabled with the \opt{autocite} package option from \secref{use:opt:pre:gen}. The \prm{name} is an identifier which serves as the value passed to the package option. The autocite commands are built on top of backend commands like \cmd{parencite} and \cmd{parencites}. The arguments \prm{cite} and \prm{multicite} specify the backend commands to use. The \prm{cite} argument refers to \cmd{autocite} and \prm{multicite} refers to \cmd{autocites}. The \prm{position} argument controls the handling of any punctuation marks after the citation. Possible values are \texttt{l}, \texttt{r}, \texttt{f}. \texttt{r} means that the punctuation is placed to the right of the citation, \ie it will not be moved around. \texttt{l} means that any punctuation after the citation is moved to the left of the citation. \texttt{f} is like \texttt{r} in a footnote and like \texttt{l} otherwise. This argument is optional and defaults to \texttt{r}. See also \cmd{DeclareAutoPunctuation} in \secref{aut:pct:cfg} and the \opt{autopunct} package option in \secref{use:opt:pre:gen}. The following examples are real definitions taken from \path{biblatex.def}:

\begin{ltxexample}
\DeclareAutoCiteCommand{plain}{\cite}{\cites}
\DeclareAutoCiteCommand{inline}{\parencite}{\parencites}
\DeclareAutoCiteCommand{footnote}[l]{\footcite}{\footcites}
\DeclareAutoCiteCommand{footnote}[f]{\smartcite}{\smartcites}
\end{ltxexample}
%
A definition provided in the document preamble can be subsequently adopted with the following: (see \secref{use:cfg:opt}).

\begin{ltxexample}
\ExecuteBibliographyOptions{autocite=<<name>>}
\end{ltxexample}

\cmditem{DeclareCitePunctuationPosition}{command}{position}

Set up the cite command \prm{command} to move punctuation marks after the citation like \cmd{autocite}. The \prm{position} argument can take the values \opt{r}, \opt{l}, \opt{f}, \opt{c}, \opt{o} and \opt{d}.
If an unknown \prm{position} identifier is used, it defaults to \opt{o}.
\begin{valuelist}
\item[r] The punctuation mark is not moved and remains to the right of the citation.
\item[l] The punctuation mark is moved to the left of the citation and thus appears before it.
\item[f] Like \opt{r} in footnotes and like \opt{l} otherwise.
\item[c] Pass the punctuation on to the internal implementation of the citation commands. It will then be executed within the \prm{wrapper} command if given.
\item[o] Retain the default setup of \opt{c} for citation defined commands without \prm{wrapper} command and \opt{l} for citation commands defined with a \prm{wrapper} command.
\item[d] Drop the explicit punctuation mark. It will only be available as the field \bibfield{postpunct}.
\end{valuelist}
%
This command can not be used for \cmd{autocite}, to configure \cmd{autocite} use the optional \prm{position} argument for \cmd{DeclareAutoCiteCommand}.
\end{ltxsyntax}

\subsubsection{Special Fields}
\label{aut:cbx:fld}

The following fields are used by \biblatex to pass data to citation commands. They are not used in \file{bib} files but defined automatically by the package. From the perspective of a citation style, they are not different from the fields in a \file{bib} file. See also \secref{aut:bbx:fld}.

\begin{fieldlist}

\fielditem{prenote}{literal}

The \prm{prenote} argument passed to a citation command. This field is specific to citations and not available in the bibliography. If the \prm{prenote} argument is missing or empty, this field is undefined.

\fielditem{postnote}{literal}

The \prm{postnote} argument passed to a citation command. This field is specific to citations and not available in the bibliography. If the \prm{postnote} argument is missing or empty, this field is undefined.

\fielditem{multiprenote}{literal}

The \prm{multiprenote} argument passed to a multicite command. This field is specific to citations and not available in the bibliography. If the \prm{multiprenote} argument is missing or empty, this field is undefined.

\fielditem{multipostnote}{literal}

The \prm{multipostnote} argument passed to a multicite command. This field is specific to citations and not available in the bibliography. If the \prm{multipostnote} argument is missing or empty, this field is undefined.

\fielditem{volcitevolume}{literal}

The \prm{volume} argument passed to \cmd{volcite} or a related citation command (\secref{use:cit:spc}). This field is specific to \cmd{volcite} citations and not available in the bibliography or other citations.

\fielditem{volcitepages}{literal}

The \prm{pages} argument passed to \cmd{volcite} or a related citation command (\secref{use:cit:spc}). This field is specific to \cmd{volcite} citations and not available in the bibliography or other citations. If the \prm{pages} argument is missing or empty, this field is undefined.

\fielditem{postpunct}{punctuation command}

The trailing punctuation argument implicitly passed to a citation command. This field is specific to citations and not available in the bibliography. If the character following a given citation command is not specified in \cmd{DeclareAutoPunctuation} (\secref{aut:pct:cfg}), this field is undefined.

\end{fieldlist}

\subsection{Data Interface}
\label{aut:bib}

The data interface are the facilities used to format and print all bibliographic data. These facilities are available in both bibliography and citation styles.

\subsubsection{Data Commands}
\label{aut:bib:dat}

This section introduces the main data interface of the \biblatex package. These are the commands doing most of the work, \ie they actually print the data provided in lists and fields.

\begin{ltxsyntax}

\cmditem{DeprecateField}{field}{message}
\cmditem{DeprecateList}{list}{message}
\cmditem{DeprecateName}{name}{message}

When an attempt is made to print \prm{field}, \prm{list}, \prm{name}, a
deprecation warning is issued with the additional \prm{message}.  This aids
style authors who are changing field names in their style. Note that the
deprecated item must no longer be defined in the datamodel for this work;
\prm{field}, \prm{list} or \prm{name} cannot be listed anywhere as an
argument to \cmd{DeclareDatamodelFields}.

\cmditem{DeprecateFieldWithReplacement}{field}{replacement}
\cmditem{DeprecateListWithReplacement}{list}{replacement}
\cmditem{DeprecateNameWithReplacement}{name}{replacement}

Similar to \cmd{DeprecateField}, \cmd{DeprecateList} and \cmd{DeprecateName}.
The commands do not only issue a deprecation warning,
they try to define a replacement for the deprecated field
that is printed in its stead.
The \cmd{replacement} must be of the same type as the deprecated
\prm{field}, \prm{list} or \prm{name}.
If the formatting of \prm{replacement} should be applied when printing
the deprecated field, that needs to be requested with \cmd{DeclareFieldAlias}
(see \secref{aut:bib:fmt}).
Note that the deprecated item must no longer be defined in the datamodel
for this work; \prm{field}, \prm{list} or \prm{name} cannot be listed
anywhere as an argument to \cmd{DeclareDatamodelFields}.

\cmditem{printfield}[format]{field}

This command prints a \prm{field} using the formatting directive \prm{format}, as defined with \cmd{DeclareFieldFormat}. If a type"=specific \prm{format} has been declared, the type"=specific formatting directive takes precedence over the generic one. If the \prm{field} is undefined, nothing is printed. If the \prm{format} is omitted, \cmd{printfield} tries using the name of the field as a format name. For example, if the \bibfield{title} field is to be printed and the \prm{format} is not specified, it will try to use the field format \texttt{title}.\footnote{In other words, \texttt{\textbackslash printfield\{title\}} is equivalent to \texttt{\textbackslash printfield[title]\{title\}}.} In this case, any type"=specific formatting directive will also take precedence over the generic one. If all of these formats are undefined, it falls back to \texttt{default} as a last resort. Note that \cmd{printfield} provides the name of the field currently being processed in \cmd{currentfield} for use in field formatting directives.

\cmditem{printlist}[format][start\ensuremath\rangle--\ensuremath\langle stop]{literal list}

This command loops over all items in a \prm{literal list}, starting at item number \prm{start} and stopping at item number \prm{stop}, including \prm{start} and \prm{stop} (all lists are numbered starting at~1). Each item is printed using the formatting directive \prm{format}, as defined with \cmd{DeclareListFormat}. If a type"=specific \prm{format} has been declared, the type"=specific formatting directive takes precedence over the generic one. If the \prm{literal list} is undefined, nothing is printed. If the \prm{format} is omitted, \cmd{printlist} tries using the name of the list as a format name. In this case, any type"=specific formatting directive will also take precedence over the generic one. If all of these formats are undefined, it falls back to \texttt{default} as a last resort. The \prm{start} argument defaults to 1; \prm{stop} defaults to the total number of items in the list. If the total number is greater than \prm{maxitems}, \prm{stop} defaults to \prm{minitems} (see \secref{use:opt:pre:gen}). See \cmd{printnames} for further details. Note that \cmd{printlist} provides the name of the literal list currently being processed in \cmd{currentlist} for use in list formatting directives.

\cmditem{printnames}[format][start\ensuremath\rangle--\ensuremath\langle stop]{name list}

This command loops over all items in a \prm{name list}, starting at item number \prm{start} and stopping at item number \prm{stop}, including \prm{start} and \prm{stop} (all lists are numbered starting at~1). Each item is printed using the formatting directive \prm{format}, as defined with \cmd{DeclareNameFormat}. If a type"=specific \prm{format} has been declared, the type"=specific formatting directive takes precedence over the generic one. If the \prm{name list} is undefined, nothing is printed. If the \prm{format} is omitted, \cmd{printnames} tries using the name of the list as a format name. In this case, any type"=specific formatting directive will also take precedence over the generic one. If all of these formats are undefined, it falls back to \texttt{default} as a last resort. The \prm{start} argument defaults to 1; \prm{stop} defaults to the total number of items in the list. If the total number is greater than \prm{maxnames}, \prm{stop} defaults to \prm{minnames} (see \secref{use:opt:pre:gen}). If you want to select a range but use the default list format, the first optional argument must still be given, but is left empty:

\begin{ltxexample}
\printnames[][1-3]{...}
\end{ltxexample}

One of \prm{start} and \prm{stop} may be omitted, hence the following arguments are all valid:

\begin{ltxexample}
\printnames[...][-1]{...}
\printnames[...][2-]{...}
\printnames[...][1-3]{...}
\end{ltxexample}

If you want to override \prm{maxnames} and \prm{minnames} and force printing of the entire list, you may refer to the \cnt{listtotal} counter in the second optional argument:

\begin{ltxexample}
\printnames[...][-\value{listtotal}]{...}
\end{ltxexample}

Whenever \cmd{printnames} and \cmd{printlist} process a list, information concerning the current state is accessible by way of four counters: the \cnt{listtotal} counter holds the total number of items in the current list, \cnt{listcount} holds the number of the item currently being processed, \cnt{liststart} is the \prm{start} argument passed to \cmd{printnames} or \cmd{printlist}, \cnt{liststop} is the \prm{stop} argument. These counters are intended for use in list formatting directives. \cnt{listtotal} may also be used in the second optional argument to \cmd{printnames} and \cmd{printlist}. Note that these counters are local to list formatting directives and do not hold meaningful values when used anywhere else. For every list, there is also a counter by the same name which holds the total number of items in the corresponding list. For example, the \cnt{author} counter holds the total number of items in the \bibfield{author} list. These counters are similar to \cnt{listtotal} except that they may also be used independently of list formatting directives. There are also \cnt{maxnames} and \cnt{minnames} as well as \cnt{maxitems} and \cnt{minitems} counters which hold the values of the corresponding package options. See \secref{aut:fmt:ilc} for a complete list of such internal counters. Note that \cmd{printnames} provides the name of the name list currently being processed in \cmd{currentname} for use in name formatting directives.

\cmditem{printtext}[format]{text}

This command prints \prm{text}, which may be printable text or arbitrary code generating printable text. It clears the punctuation buffer before inserting \prm{text} and informs \biblatex that printable text has been inserted. This ensures that all preceding and following \cmd{newblock} and \cmd{newunit} commands have the desired effect. \cmd{printfield} and \cmd{printnames} as well as \cmd{bibstring} and its companion commands (see \secref{aut:str}) do that automatically. Using this command is required if a bibliography styles inserts literal text (including the commands from \secref{aut:pct:pct, aut:pct:spc}) to ensure that block and unit punctuation works as advertised in \secref{aut:pct:new}. The optional \prm{format} argument specifies a field formatting directive to be used to format \prm{text}. This may also be useful when several fields are to be printed as one chunk, for example, by enclosing the entire chunk in parentheses or quotation marks. If a type"=specific \prm{format} has been declared, the type"=specific formatting directive takes precedence over the generic one. If the \prm{format} is omitted, the \prm{text} is printed as is. See also \secref{aut:cav:pct} for some practical hints.

\cmditem{printfile}[format]{file}

This command is similar to \cmd{printtext} except that the second argument is a file name rather than literal text. The \prm{file} argument must be the name of a valid \latex file found in \tex's search path. \cmd{printfile} will use \cmd{input} to load this \prm{file}. If there is no such file, \cmd{printfile} does nothing. The optional \prm{format} argument specifies a field formatting directive to be applied to the \prm{file}. If a type"=specific \prm{format} has been declared, the type"=specific formatting directive takes precedence over the generic one. If the \prm{format} is omitted, the \prm{file} is printed as is. Note that this feature needs to be enabled explicitly by setting the package option \opt{loadfiles} from \secref{use:opt:pre:gen}. By default, \cmd{printfile} will not input any files.

\csitem{printdate}

This command prints the date of the entry, as specified in the fields \bibfield{date} or \bibfield{month}\slash \bibfield{year}. The date format is controlled by the package option \opt{date} from \secref{use:opt:pre:gen}. Additional formatting (fonts etc.) may be applied by adjusting the field format \texttt{date} (\secref{aut:fmt:ich}). Note that this command interfaces with the punctuation tracker. There is no need to wrap it in a \cmd{printtext} command.

\csitem{printdateextra}

Similar to \cmd{printdate} but incorporates the \bibfield{extradate} field in the date specification. This is useful for bibliography styles designed for author-year citations.

\csitem{printlabeldate}

Similar to \cmd{printdate} but prints the date field determined by \cmd{DeclareLabeldate}. The date format is controlled by the package option \opt{labeldate} from \secref{use:opt:pre:gen}. Additional formatting may be applied by adjusting the field format \texttt{labeldate} (\secref{aut:fmt:ich}).

\csitem{printlabeldateextra}

Similar to \cmd{printlabeldate} but incorporates the \bibfield{extradate} field in the date specification. This is useful for bibliography styles designed for author-year citations.

\csitem{print$<$datetype$>$date}

As \cmd{printdate} but prints the \bibfield{$<$datetype$>$date} of the entry. The date format is controlled by the package option \opt{$<$datetype$>$date} from \secref{use:opt:pre:gen}. Additional formatting may be applied by adjusting the field format \texttt{$<$datetype$>$date} (\secref{aut:fmt:ich}). The $<$datetype$>$s in the default data model are <> (for the main \bibfield{date} field), <orig>, <event> and <url>.

\csitem{printtime}

This command prints the time range of the entry, as specified in the \bibfield{date} field (see \secref{bib:use:dat}). The time format is controlled by the package option \opt{time} from \secref{use:opt:pre:gen}. Additional formatting (fonts etc.) may be applied by adjusting the field format \texttt{time} (\secref{aut:fmt:ich}). Relevant to time formatting are the \opt{timezeros} option and the \cmd{bibtimesep} and \cmd{bibtimezonesep} macros (\secref{use:fmt:lng}). Note that this command interfaces with the punctuation tracker. There is no need to wrap it in a \cmd{printtext} command. Note that this command prints a stand-alone time range apart from the date elements. With the \opt{$<$datepart$>$dateusetime} option, you can have the printed along with a date when printing a date range instead of printing the time range completely separately, which is what this command allows for.

\csitem{print$<$datetype$>$time}

As \cmd{printtime} but prints the \bibfield{$<$datetype$>$time} of the entry. The time format is controlled by the package option \opt{$<$datetype$>$time} from \secref{use:opt:pre:gen}. Additional formatting may be applied by adjusting the field format \texttt{$<$datetype$>$time} (\secref{aut:fmt:ich}). The $<$datetype$>$s in the default data model are <> (for the main \bibfield{date} field), <orig>, <event> and <url>.

\cmditem{indexfield}[format]{field}

This command is similar to \cmd{printfield} except that the \prm{field} is not printed but added to the index using the formatting directive \prm{format}, as defined with \cmd{DeclareIndexFieldFormat}. If a type"=specific \prm{format} has been declared, it takes precedence over the generic one. If the \prm{field} is undefined, this command does nothing. If the \prm{format} is omitted, \cmd{indexfield} tries using the name of the field as a format name. In this case, any type"=specific formatting directive will also take precedence over the generic one. If all of these formats are undefined, it falls back to \texttt{default} as a last resort.

\cmditem{indexlist}[format][start\ensuremath\rangle--\ensuremath\langle stop]{literal list}

This command is similar to \cmd{printlist} except that the items in the list are not printed but added to the index using the formatting directive \prm{format}, as defined with \cmd{DeclareIndexListFormat}. If a type"=specific \prm{format} has been declared, the type"=specific formatting directive takes precedence over the generic one. If the \prm{literal list} is undefined, this command does nothing. If the \prm{format} is omitted, \cmd{indexlist} tries using the name of the list as a format name. In this case, any type"=specific formatting directive will also take precedence over the generic one. If all of these formats are undefined, it falls back to \texttt{default} as a last resort.

\cmditem{indexnames}[format][start\ensuremath\rangle--\ensuremath\langle stop]{name list}

This command is similar to \cmd{printnames} except that the items in the list are not printed but added to the index using the formatting directive \prm{format}, as defined with \cmd{DeclareIndexNameFormat}. If a type"=specific \prm{format} has been declared, the type"=specific formatting directive takes precedence over the generic one. If the \prm{name list} is undefined, this command does nothing. If the \prm{format} is omitted, \cmd{indexnames} tries using the name of the list as a format name. In this case, any type"=specific formatting directive will also take precedence over the generic one. If all of these formats are undefined, it falls back to \texttt{default} as a last resort.

\cmditem{entrydata}{key}{code}
\cmditem*{entrydata*}{key}{code}

Data commands like \cmd{printfield} normally use the data of the entry currently being processed. You may use \cmd{entrydata} to switch contexts locally. The \prm{key} is the entry key of the entry to use locally. The \prm{code} is arbitrary code to be executed in this context. This code will be executed in a group. See \secref{aut:cav:mif} for an example. Note that this command will automatically switch languages if the \opt{autolang} package option is enabled. The starred version \cmd{entrydata*} will clone all fields of the enclosing entry, using field, counter, and other resource names prefixed with the string <\texttt{saved}>. This is useful when comparing two data sets. For example, inside the \prm{code} argument, the \bibfield{author} field holds the author of entry \prm{key} and the author of the enclosing entry is available as \bibfield{savedauthor}. The \cnt{author} counter holds the number of names in the \bibfield{author} field of \prm{key}; the \bibfield{savedauthor} counter refers to the author count of the enclosing entry.

\cmditem{entryset}{precode}{postcode}

This command is intended for use in bibliography drivers handling \bibtype{set} entries. It will loop over all members of the set, as indicated by the \bibfield{entryset} field, and execute the appropriate driver for the respective set member. This is similar to executing the \cmd{usedriver} command from \secref{aut:aux:msc} for each set member. The \prm{precode} is arbitrary code to be executed prior to processing each item in the set. The \prm{postcode} is arbitrary code to be executed immediately after processing each item. Both arguments are mandatory in terms of the syntax but may be left empty. See \secref{aut:cav:set} for usage examples.

\cmditem{DeclareFieldInputHandler}{field}{code}

This command can be used to define a data input handler for \prm{field} when it is read from the \file{.bbl}. The \prm{code} is passed one argument (\lstinline{#1}), which contains the input field value, it should then redefine the command \cmd{NewValue}, which holds the desired output field value. For example, to ignore the \bibfield{volumes} field when it appears, you could do

\begin{ltxexample}
\DeclareFieldInputHandler{volumes}{\def\NewValue{}}
\end{ltxexample}
%
Generally, you would want to use \cmd{DeclareSourcemap} (see \secref{aut:ctm:map}) to remove and modify fields but this alternative method may be useful in some circumstances when the emphasis is on appearance rather than data since the \prm{code} can be arbitraty \tex.

In general, \cmd{DeclareFieldInputHandler} should not be used to apply formatting to a field, since that should happen with \cmd{DeclareFieldFormat}, so the following is just a toy example that shows how \cmd{DeclareFieldInputHandler} works.

\begin{ltxexample}
\DeclareFieldInputHandler{volumes}{\def\NewValue{\textbf{#1}}}
\end{ltxexample}

\cmditem{DeclareListInputHandler}{list}{code}

As \cmd{DeclareFieldInputHandler} but for lists. Within the \prm{code}, the macro \cmd{NewValue}
contains the value of the list and \cmd{NewCount} contains the number of items in the list.
Note that \cmd{NewValue} as well as the single argument to \prm{code} contain the internal representation of the list.

\cmditem{DeclareNameInputHandler}{name}{code}

As \cmd{DeclareFieldInputHandler} but for names. Within the \prm{code}, the macro \cmd{NewValue}
contains the value of the name, \cmd{NewCount} contains the number of individual names in the name and \cmd{NewOption} contains any per-name options passed in the \file{.bbl}.
Note that \cmd{NewValue} as well as the single argument to \prm{code} contain the internal representation of the name list.

\end{ltxsyntax}

\subsubsection{Formatting Directives}
\label{aut:bib:fmt}

This section introduces the commands used to define the formatting directives required by the data commands from \secref{aut:bib:dat}. Note that all standard formats are defined in \path{biblatex.def}.

\begin{ltxsyntax}

\cmditem{DeclareFieldFormat}[entrytype, \dots]{format}{code}
\cmditem*{DeclareFieldFormat}*{format}{code}

Defines the field format \prm{format}. This formatting directive is arbitrary \prm{code} to be executed by \cmd{printfield}. The value of the field will be passed to the \prm{code} as its first and only argument. The name of the field currently being processed is available to the \prm{code} as \cmd{currentfield}. If an \prm{entrytype} is specified, the format is specific to that type. The \prm{entrytype} argument may be a comma"=separated list of values. The starred variant of this command is similar to the regular version, except that all type-specific formats are cleared.

\cmditem{DeclareListFormat}[entrytype, \dots]{format}{code}
\cmditem*{DeclareListFormat}*{format}{code}

Defines the literal list format \prm{format}. This formatting directive is arbitrary \prm{code} to be executed for every item in a list processed by \cmd{printlist}. The current item will be passed to the \prm{code} as its first and only argument. The name of the literal list currently being processed is available to the \prm{code} as \cmd{currentlist}. If an \prm{entrytype} is specified, the format is specific to that type. The \prm{entrytype} argument may be a comma"=separated list of values. Note that the formatting directive also handles the punctuation to be inserted between the individual items in the list. You need to check whether you are in the middle of or at the end of the list, \ie whether \cnt{listcount} is smaller than or equal to \cnt{liststop}. The starred variant of this command is similar to the regular version, except that all type-specific formats are cleared.


\cmditem{DeclareNameFormat}[entrytype, \dots]{format}{code}
\cmditem*{DeclareNameFormat}*{format}{code}

Defines the name list format \prm{format}. This formatting directive is arbitrary \prm{code} to be executed for every name in a list processed by \cmd{printnames}. If an \prm{entrytype} is specified, the format is specific to that type. The \prm{entrytype} argument may be a comma"=separated list of values. The individual parts of a name will be available in automatically created macros (see below). The default data mode defines four name part which correspond to the standard \bibtex name parts arguments:

\begin{argumentlist}{00}
\item[family] The family name(s), know as <last> in \bibtex.  If a name consists of a single part only (for example, <Aristotle>), this part will be treated as the family name.
\item[given] The given name(s). Note that given names are referred to as the <first> names in the \bibtex file format documentation.
\item[prefix] Any name prefices, for example von, van, of, da, de, del, della, etc. Note that name prefices are referred to as the <von> part of the name in the \bibtex file format documentation.
\item[suffix] Any name suffices, for example Jr, Sr. Note that name suffices are referred to as the <Jr> part of the name in the \bibtex file format documentation.
\end{argumentlist}
%
The value of the datamodel <nameparts> constant (see \secref{aut:bbx:drv}) creates two macros for each name part in the datamodel for the name. So, for example, in the default data model, name formats will have defined the following macros:

\begin{ltxexample}
\namepartprefix
\namepartprefixi
\namepartfamily
\namepartfamilyi
\namepartsuffix
\namepartsuffixi
\namepartgiven
\namepartgiveni
\end{ltxexample}
%
If a certain part of a name is not available, the corresponding macro will be empty, hence you may use, for example, the \sty{etoolbox} tests like \cmd{ifdefvoid} to check for the individual parts of a name. The name of the name list currently being processed is available to the \prm{code} as \cmd{currentname}. Note that the formatting directive also handles the punctuation to be inserted between separate names and between the individual parts of a name. You need to check whether you are in the middle of or at the end of the list, \ie whether \cnt{listcount} is smaller than or equal to \cnt{liststop}. See also \secref{use:cav:nam}. The starred variant of this command is similar to the regular version, except that all type-specific formats are cleared.

\cmditem{DeclareListWrapperFormat}[entrytype, \dots]{format}{code}
\cmditem*{DeclareListWrapperFormat}*{format}{code}

Defines the list wrapper format \prm{format}. This formatting directive is arbitrary \prm{code} to be executed once for the entire list processed by \cmd{printlist}. The name of the literal list currently being processed is available to the \prm{code} as \cmd{currentlist}. If an \prm{entrytype} is specified, the format is specific to that type. The \prm{entrytype} argument may be a comma"=separated list of values. The starred variant of this command is similar to the regular version, except that all type-specific formats are cleared.

\cmditem{DeclareNameWrapperFormat}[entrytype, \dots]{format}{code}
\cmditem*{DeclareNameWrapperFormat}*{format}{code}

Defines the list wrapper format \prm{format}. This formatting directive is arbitrary \prm{code} to be executed once for the entire name list processed by \cmd{printnames}. The name of the literal list currently being processed is available to the \prm{code} as \cmd{currentname}. If an \prm{entrytype} is specified, the format is specific to that type. The \prm{entrytype} argument may be a comma"=separated list of values. The starred variant of this command is similar to the regular version, except that all type-specific formats are cleared.

\cmditem{DeclareIndexFieldFormat}[entrytype, \dots]{format}{code}
\cmditem*{DeclareIndexFieldFormat}*{format}{code}

Defines the field format \prm{format}. This formatting directive is arbitrary \prm{code} to be executed by \cmd{indexfield}. The value of the field will be passed to the \prm{code} as its first and only argument. The name of the field currently being processed is available to the \prm{code} as \cmd{currentfield}. If an \prm{entrytype} is specified, the format is specific to that type. The \prm{entrytype} argument may be a comma"=separated list of values. This command is similar to \cmd{DeclareFieldFormat} except that the data handled by the \prm{code} is not intended to be printed but written to the index. Note that \cmd{indexfield} will execute the \prm{code} as is, \ie the \prm{code} must include \cmd{index} or a similar command. The starred variant of this command is similar to the regular version, except that all type-specific formats are cleared.

\cmditem{DeclareIndexListFormat}[entrytype, \dots]{format}{code}
\cmditem*{DeclareIndexListFormat}*{format}{code}

Defines the literal list format \prm{format}. This formatting directive is arbitrary \prm{code} to be executed for every item in a list processed by \cmd{indexlist}. The current item will be passed to the \prm{code} as its only argument. The name of the literal list currently being processed is available to the \prm{code} as \cmd{currentlist}. If an \prm{entrytype} is specified, the format is specific to that type. The \prm{entrytype} argument may be a comma"=separated list of values. This command is similar to \cmd{DeclareListFormat} except that the data handled by the \prm{code} is not intended to be printed but written to the index. Note that \cmd{indexlist} will execute the \prm{code} as is, \ie the \prm{code} must include \cmd{index} or a similar command. The starred variant of this command is similar to the regular version, except that all type-specific formats are cleared.

\cmditem{DeclareIndexNameFormat}[entrytype, \dots]{format}{code}
\cmditem*{DeclareIndexNameFormat}*{format}{code}

Defines the name list format \prm{format}. This formatting directive is arbitrary \prm{code} to be executed for every name in a list processed by \cmd{indexnames}. The name of the name list currently being processed is available to the \prm{code} as \cmd{currentname}. If an \prm{entrytype} is specified, the format is specific to that type. The \prm{entrytype} argument may be a comma"=separated list of values. The parts of the name will be passed to the \prm{code} as separate arguments. This command is very similar to \cmd{DeclareNameFormat} except that the data handled by the \prm{code} is not intended to be printed but written to the index. Note that \cmd{indexnames} will execute the \prm{code} as is, \ie the \prm{code} must include \cmd{index} or a similar command. The starred variant of this command is similar to the regular version, except that all type-specific formats are cleared.

\cmditem{DeclareFieldAlias}[entry type]{alias}[format entry type]{format}

Declares \prm{alias} to be an alias for the field format \prm{format}. If an \prm{entrytype} is specified, the alias is specific to that type. The \prm{format entry type} is the entry type of the backend format. This is only required when declaring an alias for a type"=specific formatting directive.

\cmditem{DeclareListAlias}[entry type]{alias}[format entry type]{format}

Declares \prm{alias} to be an alias for the literal list format \prm{format}. If an \prm{entrytype} is specified, the alias is specific to that type. The \prm{format entry type} is the entry type of the backend format. This is only required when declaring an alias for a type"=specific formatting directive.

\cmditem{DeclareNameAlias}[entry type]{alias}[format entry type]{format}

Declares \prm{alias} to be an alias for the name list format \prm{format}. If an \prm{entrytype} is specified, the alias is specific to that type. The \prm{format entry type} is the entry type of the backend format. This is only required when declaring an alias for a type"=specific formatting directive.

\cmditem{DeclareListWrapperAlias}[entry type]{alias}[format entry type]{format}

Declares \prm{alias} to be an alias for the outer list format \prm{format}. If an \prm{entrytype} is specified, the alias is specific to that type. The \prm{format entry type} is the entry type of the backend format. This is only required when declaring an alias for a type"=specific formatting directive.

\cmditem{DeclareNameWrapperAlias}[entry type]{alias}[format entry type]{format}

Declares \prm{alias} to be an alias for the outer name list format \prm{format}. If an \prm{entrytype} is specified, the alias is specific to that type. The \prm{format entry type} is the entry type of the backend format. This is only required when declaring an alias for a type"=specific formatting directive.

\cmditem{DeclareIndexFieldAlias}[entry type]{alias}[format entry type]{format}

Declares \prm{alias} to be an alias for the field format \prm{format}. If an \prm{entrytype} is specified, the alias is specific to that type. The \prm{format entry type} is the entry type of the backend format. This is only required when declaring an alias for a type"=specific formatting directive.

\cmditem{DeclareIndexListAlias}[entry type]{alias}[format entry type]{format}

Declares \prm{alias} to be an alias for the literal list format \prm{format}. If an \prm{entrytype} is specified, the alias is specific to that type. The \prm{format entry type} is the entry type of the backend format. This is only required when declaring an alias for a type"=specific formatting directive.

\cmditem{DeclareIndexNameAlias}[entry type]{alias}[format entry type]{format}

Declares \prm{alias} to be an alias for the name list format \prm{format}. If an \prm{entrytype} is specified, the alias is specific to that type. The \prm{format entry type} is the entry type of the backend format. This is only required when declaring an alias for a type"=specific formatting directive.

\cmditem{DeprecateFieldFormatWithReplacement}[entry type]{alias}[format entry type]{format}

Declares \prm{alias} to be an alias for the name list format \prm{format} and issue a deprecation warning. If an \prm{entrytype} is specified, the alias is specific to that type. The \prm{format entry type} is the entry type of the backend format. This is only required when declaring an alias for a type"=specific formatting directive.

\cmditem{DeprecateListFormatWithReplacement}[entry type]{alias}[format entry type]{format}

Similar to \cmd{DeprecateFieldFormatWithReplacement} but for list formats.

\cmditem{DeprecateNameFormatWithReplacement}[entry type]{alias}[format entry type]{format}

Similar to \cmd{DeprecateFieldFormatWithReplacement} but for name formats.

\cmditem{DeprecateListWrapperFormatWithReplacement}[entry type]{alias}[format entry type]{format}

Similar to \cmd{DeprecateFieldFormatWithReplacement} but for outer list formats.

\cmditem{DeprecateNameWrapperFormatWithReplacement}[entry type]{alias}[format entry type]{format}

Similar to \cmd{DeprecateFieldFormatWithReplacement} but for outer name formats.

\cmditem{DeprecateIndexFieldFormatWithReplacement}[entry type]{alias}[format entry type]{format}

Similar to \cmd{DeprecateFieldFormatWithReplacement} but for index field formats.

\cmditem{DeprecateIndexListFormatWithReplacement}[entry type]{alias}[format entry type]{format}

Similar to \cmd{DeprecateFieldFormatWithReplacement} but for index list formats.

\cmditem{DeprecateIndexNameFormatWithReplacement}[entry type]{alias}[format entry type]{format}

Similar to \cmd{DeprecateFieldFormatWithReplacement} but for index name formats.

\end{ltxsyntax}

\subsection{Customization}
\label{aut:ctm}

\subsubsection{Related Entries}
\label{aut:ctm:rel}
The related entries feature comprises the following components:
\begin{itemize}
\item Special fields in an entry to set up and describe relationships
\item Optionally, localisation strings to prefix the related data
\item Macros to extract and print the related data
\item Formats to format the localisation string and related data
\end{itemize}
%
The special fields are \bibfield{related}, \bibfield{relatedtype}, \bibfield{relatedstring} and \bibfield{relatedoptions}:
\begin{keymarglist}
\item[related] A separated list of keys of entries which are related to this entry in some way. Note the order of the keys is important. The data from multiple related entries is printed in the order of the keys listed in this field.
\item[relatedtype] The type of relationship. This serves three purposes. If the value of this field resolves to a localisation string identifier, then the resulting localised string is printed before the data from the related entries. Secondly, if there is a macro called \texttt{related:\prm{relatedtype}}, this is used to format the data from the related entries. If no such macro exists, then the macro \texttt{related:default} is used. Lastly, if there is a format named \texttt{related:\prm{relatedtype}}, then it is used to format both the localised string and related entry data. If there is no related type specific format, the \texttt{related} format is used.
\item[relatedstring] If an entry contains this field, then if value of the field resolves to a localisation string identifier, the localisation key value specified is printed before data from the related entries. If the field does not specify a localisation key, its value is printed literally. If both \bibfield{relatedtype} and \bibfield{relatedstring} are present in an entry, \bibfield{relatedstring} is used for the pre-data string (but \bibfield{relatedtype} is still used to determine the macro and format to use when printing the data).
\item[relatedoptions] A list of per"=entry options to set on the related entry (actually on the clone of the related entry which is used as a data source---the actual related entry is not modified because it might be cited directly itself).
\end{keymarglist}

The related entry feature is enabled by default by the package option \opt{related} from \secref{use:opt:pre:gen}. The related information entry data from the related entries is included via a \cmd{usebibmacro\{related\}} call. Standard styles call this macro towards the end of each driver. Style authors should ensure the existence of (or take note of existing) localisation strings which are useful as values for the \bibfield{relatedtype} field, such as \texttt{translationof} or perhaps \texttt{translatedas}. A plural variant can be identified with the localisation key \prm{relatedtype}\texttt{s}. This key's corresponding string is printed whenever more than one entry is specified in \bibfield{related}. Bibliography macros and formatting directives for printing entries related by \prm{relatedtype} should be defined using the name \texttt{related:\prm{relatedtype}}. The file \path{biblatex.def} contains macros and formats for some common relation types which can be used as templates. In particular, the \cmd{entrydata*} command is essential in such macros in order to make the data of the related entries available. Examples of entries using this feature can be found in the \biblatex distribution examples file \path{biblatex-examples.bib}. There are some specific formatting macros for this feature which control delimiters and separators in related entry information, see \secref{aut:fmt:fmt}.

\subsubsection{Datasource Sets}
\label{aut:ctm:dsets}

It is useful to be able to define named sets of datasource field names for use in loops etc. In addition, \biber can use such sets in order to apply options and perform operations on particular sets of datasource fields. The following macros allow the user to define arbitrary sets of datasource fields, exposed to \biblatex as \sty{etoolbox} lists and to \biber in the \file{.bcf}.


\begin{ltxsyntax}

\cmditem{DeclareDatafieldSet}{name}{specification}

Declare a set of datasource fields with name \prm{name}.

\begin{optionlist*}

\valitem{name}{set name}

The name of the set.

\end{optionlist*}

The \prm{specification} is one or more \cmd{member} items:

\cmditem{member}

\begin{optionlist*}

\valitem{fieldtype}{fieldtype}
\valitem{datatype}{datatype}
\valitem{field}{fieldname}

\end{optionlist*}

A \cmd{member} specification appends fields to the set. Fields can be specified by datamodel \prm{fieldtype} and/or \prm{datatype} (see \secref{aut:ctm:dm}). Alternatively, fields can be explicitly added by name using the \prm{field} option. Once defined, the set is available as an \sty{etoolbox} list called \cmd{datafieldset<setname>} and is also passed via the \file{.bcf} to \biber.

For example, here are the default sets defined by \biblatex for name fields and title fields:

\end{ltxsyntax}

\begin{ltxexample}[style=latex]
\DeclareDatafieldSet{setnames}{
  \member[datatype=name, fieldtype=list]
}

\DeclareDatafieldSet{settitles}{
  \member[field=title]
  \member[field=booktitle]
  \member[field=eventtitle]
  \member[field=issuetitle]
  \member[field=journaltitle]
  \member[field=maintitle]
  \member[field=origtitle]
}
\end{ltxexample}
%
This defines the macros \cmd{datafieldsetsetnames} and \cmd{datafieldsetsettitles} as \sty{etoolbox} lists containing the names of the member datasource fields specified.

\subsubsection{Dynamic Modification of Data}
\label{aut:ctm:map}

Bibliographic data sources which are automatically generated or which you have no control over can be a problem if you need to edit them in some way. For this reason, \biber has the ability to modify data as it is read so that you can apply modifications to the source data stream without actually changing it. The modification can be defined in \biber's config file (see \biber docs), or via \biblatex macros in which case you can apply the modification only for specific documents, styles or globally.

Source mapping happens during data parsing and therefore before any other operation such as inheritance and sorting.

Source mappings can be defined at different «levels» which are applied
in a defined order. See the \biblatex\ manual regarding these macros:\\[2ex]

\noindent \texttt{user}-level maps defined with \cmd{DeclareSourcemap}$\rightarrow$\\
\hspace*{1em}\texttt{user}-level maps defined in the \biber config file (see \biber docs)$\rightarrow$\\
\hspace*{2em}\texttt{style}-level maps defined with \cmd{DeclareStyleSourcemap}$\rightarrow$\\
\hspace*{3em}\texttt{driver}-level maps defined with \cmd{DeclareDriverSourcemap}\\[2ex]

\begin{ltxsyntax}

\cmditem{DeclareSourcemap}{specification}

Defines source data modification (mapping) rules which can be used to perform any combination of the following tasks:

\begin{itemize}
\item Map data source entrytypes to different entrytypes
\item Map datasource fields to different fields
\item Add new fields to an entry
\item Remove fields from an entry
\item Modify the contents of a field using standard Perl regular expression
  match and replace\footnote{See for example \url{https://perldoc.perl.org/perlretut.html}, \url{https://perldoc.perl.org/perlrequick.html} and \url{https://perldoc.perl.org/perlre.html}. There are many more resources available about regular expessions in Perl.}
\item Restrict any of the above operations to entries coming from
  particular datasources which you defined in \cmd{addresource} macros
\item Restrict any of the above operations to entries only of a certain
  entrytype
\item Restrict any of the above operations to entries in a particular
  reference section
\end{itemize}

The \prm{specification} is an undelimited list of \cmd{maps} directives which specify containers for mappings rules applying to a particular data source type (\secref{use:bib:res}). Spaces, tabs, and line endings may be used freely to visually arrange the \prm{specification}. Blank lines are not permissible. This command may only be used in the preamble and can be used multiple times, the maps being run in order of definition.

\cmditem{maps}[options]{elements}

Contains an ordered set of \cmd{map} elements each of which is a logically related set of mapping steps to apply to the data source. The \prm{options} are:

\begin{optionlist*}

\choitem[bibtex]{datatype}{bibtex, biblatexml}

Data source type to which the contained \cmd{map} directives apply (\secref{use:bib:res}).

\boolitem[false]{overwrite}

Specify whether a mapping rule is allowed to overwrite already existing data in an entry. If this option is not specified, the default is \texttt{false}. The short form \opt{overwrite} is equivalent to \kvopt{overwrite}{true}.

\end{optionlist*}

\cmditem{map}[options]{restrictions,steps}

A container for an ordered set of map \cmd{step}s, optionally restricted to particular entrytypes or data sources. This is a grouping element to allow a set of mapping steps to apply only to specific entrytypes or data sources. Mapping steps must always be contained within a \cmd{map} element. The \prm{options} are:

\begin{optionlist*}

\boolitem{overwrite}

As the same option on the parent \cmd{maps} element. This option allows an override on a per-map group basis. If this option is not specified, the default is the parent \cmd{maps} element option value. The short form \opt{overwrite} is equivalent to \kvopt{overwrite}{true}.

\valitem{foreach}{loopval}

Loop over all \cmd{step}s in this \cmd{map}, setting the special variable |$MAPLOOP| %$
to each of the comma-separated values contained in \prm{loopval}. \prm{loopval} can either be the name of a datafield set defined with \cmd{DeclareDatafieldSet} (see \secref{aut:ctm:dsets}), a datasource field which is fetched and parsed as a comma"=separated values list or an explicit comma"=separated values list. \prm{loopval} is determined in this order. This allows the user to repeat a group of \cmd{step}s for each value \prm{loopval}. Using regexp maps, it is possible to create a CSV field for use with this functionality. The special variable |$MAPUNIQ| %$
may also be used in the \cmd{step}s to generate a random unique string. This can be useful when creating keys for new entries. An example:

\begin{ltxexample}[style=latex]
\DeclareSourcemap{
  \maps[datatype=bibtex]{
    \map[overwrite, foreach={author,editor, translator}]{
      \step[fieldsource=\regexp{$MAPLOOP}, match={Smith}, replace={Jones}]
    }
  }
}
\end{ltxexample}
%$<- to stop emacs highlighting breaking

\intitem{refsection}

Only apply the contained \cmd{step} commands to entries in the reference section with number \prm{refsection}.

\end{optionlist*}

\cmditem{perdatasource}{datasource}

Restricts all \cmd{step}s in this \cmd{map} element to entries from the named \prm{datasource}. The \prm{datasource} name should be exactly as given in a \cmd{addresource} macro defining a data source for the document. Multiple \cmd{perdatasource} restrictions are allowed within a \cmd{map} element.

\cmditem{pertype}{entrytype}

Restricts all \cmd{step}s in this \cmd{map} element to entries of the named \prm{entrytype}. Multiple \cmd{pertype} restrictions are allowed within a \cmd{map} element.

\cmditem{pernottype}{entrytype}

Restricts all \cmd{step}s in this \cmd{map} element to entries not of the named \prm{entrytype}. Multiple \cmd{pernottype} restrictions are allowed within a \cmd{map} element.

\cmditem{step}[options]

A mapping step. Each step is applied sequentially to every relevant entry where <relevant> means those entries which correspond to the data source type, entrytype and data source name restrictions mentioned above. Each step is applied to the entry as it appears after the application of all previous steps. The mapping performed by the step is determined by the following \prm{option}s:

\begin{optionlist*}

\valitem{typesource}{entrytype}
\valitem{typetarget}{entrytype}
\valitem{fieldsource}{entryfield}
\valitem{notfield}{entryfield}
\valitem{fieldtarget}{entryfield}
\valitem{match}{regexp}
\valitem{matchi}{regexp}
\valitem{notmatch}{regexp}
\valitem{notmatchi}{regexp}
\valitem{replace}{regexp}
\valitem{fieldset}{entryfield}
\valitem{fieldvalue}{string}
\valitem{entryclone}{clonekey}
\valitem{entrynew}{entrynewkey}
\valitem{entrynewtype}{string}
\valitem{entrytarget}{string}
\boolitem[false]{cited}
\boolitem[false]{nocited}
\boolitem[false]{citedornocited}
\boolitem[false]{allnocited}
\boolitem[false]{starnocited}
\boolitem[false]{entrynocite}
\boolitem[false]{entrynull}
\boolitem[false]{append}
\boolitem[false]{appendstrict}
\boolitem[false]{final}
\boolitem[false]{null}
\boolitem[false]{origfield}
\boolitem[false]{origfieldval}
\boolitem[false]{origentrytype}
%
For all boolean \cmd{step} options, the short form \opt{option} is equivalent to \kvopt{option}{true}. The following rules for a mapping step apply:

\renewcommand{\labelitemii}{$\circ$}

Note that the options \opt{cited}, \opt{nocited}, \opt{citedornocited},
\opt{allnocited} and \opt{starnocited} are unique in that they can make the
results of a sourcemap differ depending on the refsection. This is because
a datasource to which source mapping applies may be used in several
refsections and source mappings are applied when fetching the data from the
datasources for a refsection. Citation commands are local to a refsection
and therefore may differ for the same entry from refsection to refsection.
For example, the same entry may be \cmd{cite}d in one refsection but
\cmd{nocite}d in another, resulting in different source map results and
therefore data between the refsections. This can be avoided if desired, by
limiting source maps to specific refsections only (see \opt{refsection}
option to the \cmd{map} command above).

\begin{itemize}
\item If \texttt{entrynew} is set, a new entry is created with the entry key \texttt{entrynewkey} and the entry type given in the option \texttt{entrynewtype}. This
  entry is only in-scope during the processing of the current entry and can be referenced by
  \texttt{entrytarget}.  In \texttt{entrynewkey}, you may use standard Perl regular expression
  backreferences to captures from a  previous \texttt{match} step.
\item When a \texttt{fieldset} step has \texttt{entrytarget} set to the entrykey of an entry
  created by \texttt{entrynew}, the target for the field set will be the \texttt{entrytarget} entry
  rather than the entry being currently processed. This allows users to create new entries and set
  fields in them.
\item If \texttt{entrynocite} is used in a \texttt{entrynew} or
  \texttt{entryclone} step, the new/clone entry will be included in the
  \file{.bbl} as if the entry/clone had been \cmd{nocite}ed in the document.
\item If \texttt{entrynull} is set, processing of the \cmd{map}
  immediately terminates and the current entry is not created. It is
  as if it did not exist in the datasource. Obviously, you should
  select the entries which you want to apply this to using prior
  mapping steps.
\item If \texttt{entryclone} is set, a clone of the entry is created with an entry key
  \texttt{clonekey}. Obviously this may cause labelling problems in author/year styles etc.
  and should be used with care. The cloned entry is in-scope during the processing of the
  current entry and can be modified by passing its key as the value to \texttt{entrytarget}.
  In \texttt{clonekey}, you may use standard Perl regular expression backreferences to
  captures from a previous \texttt{match} step.
\item If \texttt{cited} is used then only apply the step if the entry key
  of an entry was specifically cited via \cmd{cite}.
\item If \texttt{nocited} is used then only apply the step if the entry key
  of an entry was specifically nocited via \cmd{nocite} or was included via \cmd{nocite\{*\}}.
\item If \texttt{citedornocited} is used then only apply the step if the entry key
  of an entry was specifically cited via \cmd{cite} or specifically nocited via \cmd{nocite}.
\item If \texttt{allnocited} is used then only apply the step if the entry key
  of an entry was included via \cmd{nocite\{*\}}.
\item If \texttt{starnocited} is used then only apply the step if the entry key
  of an entry was included solely because of \cmd{nocite\{*\}}. This
  implies that the entry was neither explicitly \cmd{cite}ed nor explicitly \cmd{nocite}ed.
\item Change the \texttt{typesource} \prm{entrytype} to the
  \texttt{typetarget} \prm{entrytype}, if defined. If
  \texttt{final} is \texttt{true} then if the \prm{entrytype} of the entry is not \texttt{typesource}, processing of the parent \cmd{map} immediately terminates.
\item Change the \texttt{fieldsource} \prm{entryfield} to
  \texttt{fieldtarget}, if defined. If
  \texttt{final} is \texttt{true} then if there is no \texttt{fieldsource} \prm{entryfield} in the entry, processing of the parent \cmd{map} immediately terminates.
\item If \texttt{notfield} is true only if the \prm{entryfield} does not
  exist. Usually used with \texttt{final} so that if an entry does contain
  \prm{entryfield}, the map terminates.
\item If \texttt{match} is defined but
  \texttt{replace} is not, only apply the step if the \texttt{fieldsource} \prm{entryfield} matches the
  \texttt{match} regular expression (logic is reversed if you use \texttt{notmatch} and case-insensitive if you use the versions ending in <i>)\footnote{Regular expressions are full Perl 5.16 regular expressions. This means you may need to deal with special characters, see examples.}. You may use capture parenthesis as usual and refer to these (\$1\ldots\$9) in later \texttt{fieldvalue} specifications. This allows you to pull out parts of some fields and put these parts in other fields.
\item Perform a regular expression match and replace on the value of the \texttt{fieldsource} \prm{entryfield} if \texttt{match} and \texttt{replace} are defined.
\item If \texttt{fieldset} is defined, then its value is \prm{entryfield}
  which will be set to a value specified by further options. If
  \texttt{overwrite} is false for this step and the field to set already
  exists then the map step is ignored. If \texttt{final} is also true for
  this step, then processing of the parent map stops at this point. If
  \texttt{append} is true, then the value to set is appended to the current
  value of \prm{entryfield}. \texttt{appendstrict} only appends to
  \prm{entryfield} if \prm{entryfield} is not empty. The value to set is
  specified by a mandatory one and only one of the following options:
  \begin{itemize}
    \item\ \texttt{fieldvalue} --- The \texttt{fieldset} \prm{entryfield} is set to the \texttt{fieldvalue} \prm{string}
    \item\ \texttt{null} --- The \texttt{fieldset} \prm{entryfield} is ignored, as if it did not exist in the datasource
    \item\ \texttt{origentrytype} --- The \texttt{fieldset} \prm{entryfield} is set to the most recently mentioned \texttt{typesource} \prm{entrytype} name
    \item\ \texttt{origfield} --- The \texttt{fieldset} \prm{entryfield} is set to the most recently mentioned \texttt{fieldsource} \prm{entryfield} name
    \item\ \texttt{origfieldval} --- The \texttt{fieldset} \prm{entryfield} is set to the most recently mentioned \texttt{fieldsource} value
  \end{itemize}
\end{itemize}

\end{optionlist*}

\end{ltxsyntax}

\noindent With \bibtex\ datasources, you may specify the
pseudo-field \bibfield{entrykey} for \texttt{fieldsource}
which is the citation key of the entry. With \biblatexml\ the \bibfield{entrykey} is a normal attribute and can be reference like any other attribute. Naturally, this <field> cannot
be changed (used as \texttt{fieldset}, \texttt{fieldtarget} or changed using \texttt{replace}).

Macros used in \cmd{step} are expanded. Unexpandable contents should be protected with \cmd{detokenize}, regular expressions can be escaped using the dedicated \cmd{regexp} command (see the examples below).

\begin{ltxsyntax}

\cmditem{DeclareStyleSourcemap}{specification}

This command sets the source mappings used by a style. Such mappings are conceptually separate from user mappings defined with \cmd{DeclareSourcemap} and are applied directly after user maps. The syntax is identical to \cmd{DeclareSourcemap}. This command is provided for style authors so that any maps defined for the style do not interfere with user maps or the default driver maps defined with \cmd{DeclareDriverSourcemap}. This command is for use in style files and can be used multiple times, the maps being run in order of definition.

\end{ltxsyntax}

\begin{ltxsyntax}

\cmditem{DeclareDriverSourcemap}[datatype=driver]{specification}

This command sets the driver default source mappings for the specified \prm{driver}. Such mappings are conceptually separate from user mappings defined with \cmd{DeclareSourcemap} and style mapping defined with \cmd{DeclareStyleSourcemap}. They consist of mappings which are part of the driver setup. Users should not normally need to change these. Driver default mapping are applied after user mappings (\cmd{DeclareSourcemap}) and style mappings (\cmd{DeclareStyleSourcemap}). These defaults are described in Appendix \secref{apx:maps}. The \prm{specification} is identical to that for \cmd{DeclareSourcemap} but without the \cmd{maps} elements: the \prm{specification} is just a list of \cmd{map} elements since each \cmd{DeclareDriverSourcemap} only applies to one datatype driver. See the default definitions in Appendix \secref{apx:maps} for examples.

\end{ltxsyntax}

Here are some data source mapping examples:

\begin{ltxexample}
\DeclareSourcemap{
  \maps[datatype=bibtex]{
    \map{
      \perdatasource{<<example1.bib>>}
      \perdatasource{<<example2.bib>>}
      \step[fieldset=<<keywords>>, fieldvalue={<<keyw1, keyw2>>}]
      \step[fieldsource=<<entrykey>>]
      \step[fieldset=<<note>>, origfieldval]
    }
  }
}
\end{ltxexample}
%
This would add a \bibfield{keywords} field with value <keyw1, keyw2> and set the \bibfield{note} field to the entry key to all entries which are found in either the
\texttt{examples1.bib} or \texttt{examples2.bib} files.
%
\begin{ltxexample}
\DeclareSourcemap{
  \maps[datatype=bibtex]{
    \map{
      \step[fieldsource=<<title>>]
      \step[fieldset=<<note>>, origfieldval]
    }
  }
}
\end{ltxexample}
%
Copy the \bibfield{title} field to the \bibfield{note} field unless the
\bibfield{note} field already exists.
%
\begin{ltxexample}
\DeclareSourcemap{
  \maps[datatype=bibtex]{
    \map{
      \step[typesource=<<chat>>, typetarget=<<customa>>, final]
      \step[fieldset=<<type>>, origentrytype]
    }
  }
}
\end{ltxexample}
%
Any \bibfield{chat} entrytypes would become \bibfield{customa} entrytypes and
would automatically have a \bibfield{type} field set to
<chat> unless the \bibfield{type} field already exists in the entry (because
\texttt{overwrite} is false by default). This mapping applies only to entries of type
\bibtype{chat} since the first step has \texttt{final} set and so if the
\texttt{typesource} does not match the entry entrytype, processing of this
\cmd{map} immediately terminates.
%
\begin{ltxexample}
\DeclareSourcemap{
  \maps[datatype=bibtex]{
    \map{
      \perdatasource{<<examples.bib>>}
      \pertype{<<article>>}
      \pertype{<<book>>}
       \step[fieldset=<<abstract>>, null]
       \step[fieldset=<<note>>, fieldvalue={<<Auto-created this field>>}]
    }
  }
}
\end{ltxexample}
%
Any entries of entrytype \bibtype{article} or \bibtype{book} from the
\texttt{examples.bib} datasource would have their \bibfield{abstract}
fields removed and a \bibfield{note} field added with value <Auto-created this field>.
%
\begin{ltxexample}
\DeclareSourcemap{
  \maps[datatype=bibtex]{
    \map{
       \step[fieldset=<<abstract>>, null]
       \step[fieldsource=<<conductor>>, fieldtarget=<<namea>>]
       \step[fieldsource=<<gps>>, fieldtarget=<<usera>>]
    }
  }
}
\end{ltxexample}
%
This removes \bibfield{abstract} fields from any entry, changes
\bibfield{conductor} fields to \bibfield{namea} fields and changes \bibfield{gps}
fields to \bibfield{usera} fields.
%
\begin{ltxexample}
\DeclareSourcemap{
  \maps[datatype=bibtex]{
    \map{
       \step[fieldsource=<<pubmedid>>, fieldtarget=<<eprint>>, final]
       \step[fieldset=<<eprinttype>>, origfield]
       \step[fieldset=<<userd>>, fieldvalue={<<Some string of things>>}]
    }
  }
}
\end{ltxexample}
%
Applies only to entries with \bibfield{pubmed} fields and maps
\bibfield{pubmedid} fields to \bibfield{eprint} fields, sets the \bibfield{eprinttype}
field to <pubmedid> and also sets the \bibfield{userd} field to the string
<Some string of things>.
%
\begin{ltxexample}
\DeclareSourcemap{
  \maps[datatype=bibtex]{
    \map{
       \step[fieldsource=<<series>>,
             match=\regexp{<<\A\d*(.+)>>},
             replace=\regexp{<<\L$1>>}]
    }
  }
}
\end{ltxexample}
%$<- to stop emacs highlighting breaking
Here, the contents of the \bibfield{series} field have leading numbers stripped and the remainder of the contents lowercased. Since regular expressions usually contain all sort of special characters, it is best to enclose them in the provided \cmd{regexp} macro as shown---this will pass the expression through to \biber\ correctly.
%
\begin{ltxexample}
\DeclareSourcemap{
  \maps[datatype=bibtex]{
    \map{
       \step[fieldsource=<<maintitle>>,
             match=\regexp{<<Collected\s+Works.+Freud>>},
             final]
       \step[fieldset=<<keywords>>, fieldvalue=<<freud>>]
    }
  }
}
\end{ltxexample}
%$<- to stop emacs highlighting breaking
Here, if for an entry, the \bibfield{maintitle} field matches a particular regular expression, we set a special keyword so we can, for example, make a references section just for certain items.
%
\begin{ltxexample}
\DeclareSourcemap{
  \maps[datatype=bibtex]{
    \map{
       \step[fieldsource=<<lista>>, match=\regexp{<<regexp>>}, final]
       \step[fieldset=<<lista>>, null]
    }
  }
}
\end{ltxexample}
%
If an entry has a \bibfield{lista} field which matches regular expression <regexp>, then it is removed.
%
\begin{ltxexample}
\DeclareSourcemap{
  \maps[datatype=bibtex]{
    \map[overwrite=false]{
       \step[fieldsource=<<author>>]
       \step[fieldset=<<editor>>, origfieldval, final]
       \step[fieldsource=<<editor>>, match=\regexp{<<\A(.+?)\s+and.*>>}, replace={<<$1>>}]
    }
  }
}
\end{ltxexample}
%$<- to stop emacs highlighting breaking
For any entry with an \bibfield{author} field, try to set
\bibfield{editor} to the same as \bibfield{author}. If this fails because
\bibfield{editor} already exists, stop, otherwise truncate
\bibfield{editor} to just the first name in the name list.
%
\begin{ltxexample}
\DeclareSourcemap{
  \maps[datatype=bibtex]{
    \map{
       \step[fieldsource=<<author>>,
             match={<<Smith, Bill>>},
             replace={<<Smith, William>>}]
       \step[fieldsource=<<author>>,
             match={<<Jones, Baz>>},
             replace={<<Jones, Barry>>}]
    }
  }
}
\end{ltxexample}
%
Here, we use multiple match/replace for the same field to regularise some inconstant name variants. Bear in mind that \cmd{step} processing within a \opt{map} element is sequential and so the changes from a previous \cmd{step}s are already committed. Note that we don't need the \cmd{regexp} macro to protect the regular expressions in this example as they contain no characters which need special escaping. Please note that due to the difficulty of protecting regular expressions in \LaTeX, there should be no literal spaces in the argument to \cmd{regexp}. Please use escape code equivalents if spaces are needed. For example, this example, if using \cmd{regexp}, should be:
%
\begin{ltxexample}
\DeclareSourcemap{
  \maps[datatype=bibtex]{
    \map{
       \step[fieldsource=<<author>>,
             match=\regexp{<<Smith,\s+Bill>>},
             replace=\regexp{<<Smith,\x20William>>}]
       \step[fieldsource=<<author>>,
             match=\regexp{<<Jones,\s+Baz>>},
             replace=\regexp{<<Jones,\x20Barry>>}]
    }
  }
}
\end{ltxexample}
%
Here, we have used the hexadecimal escape sequence <|\x20|> in place of literal spaces in the replacement strings.
%
\begin{ltxexample}
\DeclareSourcemap{
  \maps[datatype=bibtex]{
    \map[overwrite]{
       \step[fieldsource=<<author>>, match={<<Doe,>>}, final]
       \step[fieldset=<<shortauthor>>, origfieldval]
       \step[fieldset=<<sortname>>, origfieldval]
       \step[fieldsource=<<shortauthor>>,
             match=\regexp{<<Doe,\s*(?:\.|ohn)(?:[-]*)(?:P\.|Paul)*>>},
             replace={<<Doe, John Paul>>}]
       \step[fieldsource=<<sortname>>,
             match=\regexp{<<Doe,\s*(?:\.|ohn)(?:[-]*)(?:P\.|Paul)*>>},
             replace={<<Doe, John Paul>>}]
    }
  }
}
\end{ltxexample}
%
Only applies to entries with an \bibfield{author} field matching <Doe,>. First the \bibfield{author} field is copied to both the \bibfield{shortauthor} and \bibfield{sortname} fields, overwriting them if they already exist. Then, these two new fields are modified to canonicalise a particular name, which presumably has some variants in the data source.
%
\begin{ltxexample}
\DeclareSourcemap{
  \maps[datatype=bibtex]{
    \map[overwrite]{
      \step[fieldsource=<<verba>>, final]
      \step[fieldset=<<verbb>>, fieldvalue=<</>>, append]
      \step[fieldset=<<verbb>>, origfieldval, append]
      \step[fieldsource=<<verbb>>, final]
      \step[fieldset=<<verbc>>, fieldvalue=<</>>, append]
      \step[fieldset=<<verbc>>, origfieldval, append]
    }
  }
}
\end{ltxexample}
%
This example demonstrates the sequential nature of the step processing and the \opt{append} option. If an entry has a \bibfield{verba} field then first, a forward slash is appended to the \bibfield{verbb} field. Then, the contents of \bibfield{verba} are appended to the \bibfield{verbb} field. A slash is then appended to the \bibfield{verbc} field and the contents of \bibfield{verbb} are appended to the \bibfield{verbc} field.
%
\begin{ltxexample}
\DeclareSourcemap{
  \maps[datatype=bibtex]{
    \map[overwrite]{
      \step[fieldset=<<autourl>>, fieldvalue={<<http://scholar.google.com/scholar?q=">>}]
      \step[fieldsource=<<title>>]
      \step[fieldset=<<autourl>>, origfieldval, append]
      \step[fieldset=<<autourl>>, fieldvalue={<<"+author:>>}, append]
      \step[fieldsource=<<author>>, match=\regexp{<<\A([^,]+)\s*,>>}]
      \step[fieldset=<<autourl>>, fieldvalue={<<$1>>}, append]
      \step[fieldset=<<autourl>>, fieldvalue={<<&as_ylo=>>}, append]
      \step[fieldsource=<<year>>]
      \step[fieldset=<<autourl>>, origfieldval, append]
      \step[fieldset=<<autourl>>, fieldvalue={<<&as_yhi=>>}, append]
      \step[fieldset=<<autourl>>, origfieldval, append]
    }
  }
}
\end{ltxexample}%$ <- keep AucTeX highlighting happy
%
This example assumes you have created a field called \bibfield{autourl} using the datamodel macros from \secref{aut:ctm:dm} in order to hold, for example, a Google Scholar query URL auto-created from elements of the entry. The example progressively extracts information from the entry, constructing the URL as it goes. It demonstrates that it is possible to refer to parenthetical matches from the most recent \texttt{match} in any following \texttt{fieldvalue} which allows extracting the family name from the \bibfield{author}, assuming a <family, given> format. The resulting field could then be used as a hyperlink from, for example, the title of the work in the bibliography.
%
\begin{ltxexample}
\DeclareSourcemap{
  \maps[datatype=bibtex]{
    \map{
      \step[fieldsource=<<title>>, match={A Title}, final]
      \step[entrynull]
    }
  }
}
\end{ltxexample}
%
Any entry with a \bibfield{title} field matching <A Title> will be completely ignored.
%
\begin{ltxexample}
\DeclareSourcemap{
  \maps[datatype=bibtex]{
    \map{
      \pernottype{book}
      \pernottype{article}
      \step[entrynull]
    }
  }
}
\end{ltxexample}
%
Any entry which is not a \bibtype{book} or \bibtype{article} will be ignored.
%
\begin{ltxexample}
\DeclareSourcemap{
  \maps[datatype=bibtex]{
    \map{
      \perdatasource{<<biblatex-examples.bib>>}
      \step[entryclone={rel-}]
    }
  }
}
\end{ltxexample}
%
Here, a clone of an entry from the specified data source will be created. The entry key of the clone will be the same as the original but prefixed by the value of the \texttt{entryclone} parameter. The cloned entry would still need to be cited in the document using its new entry key. This type of mapping step should be used with care as it may produce labelling problems in authoryear styles which use, for example, \opt{extradate}. One use case is for numeric styles which contain multiple bibliographies containing the same entry. In this case, you may need different bibliography number labeld for the same entry and this is very tricky when there is only one entry which needs different labels. Creating clones with different entry keys solves this problem.

\biblatexml\ datasources are more structured than \bibtex\ since they are XML. Sourcemapping is possible with \biblatexml\ too but the specifications of source and target fields etc. also support XPath 1.0 paths in order to be able to work with the structured data. Fields can be specified as per the \bibtex\ examples above and these are converted into XPath 1.0 queries internally as necessary. For example:

\begin{ltxexample}
\DeclareSourcemap{
  \maps[datatype=biblatexml]{
    \map{
   \step[fieldsource=\regexp{./bltx:names[@type='author']/bltx:name[2]/bltx:namepart[@type='family']},
      match=\regexp{\ASmith},
      replace={Jones}]
    }
    \map{
      \step[fieldsource=editor, fieldtarget=translator]
    }
    \map{
      \step[fieldsource=\regexp{./bltx:names[@type='editor']},
            fieldtarget=\regexp{./bltx:names[@type='translator']}]
    }
    \map{
      \step[fieldset=\regexp{./bltx:names[@type='author']/bltx:name[2]/@useprefix},
            fieldvalue={false}]
    }
  }
}
\end{ltxexample}
%
These maps, respectively,

\begin{itemize}
\item Replace the family name <Smith> of the second \bibfield{author} name with <Jones>
\item Move the \bibfield{editor} to \bibfield{translator}
\item Move the \bibfield{editor} to \bibfield{translator} but with explicit XPaths
\item Set the per-namelist \opt{useprefix} option on the \bibfield{author} name list to <false>
\end{itemize}

\subsubsection{Data Model Specification}
\label{aut:ctm:dm}

The data model which \biblatex uses consists of four main elements:

\begin{itemize}
\item Specification of constant strings and lists of strings
\item Specification of valid Entrytypes
\item Specification of valid Fields along with their type, datatype and any special flags
\item Specification of which Fields are valid in which Entrytypes
\item Specification of constraints which can be used to validate data against the data model
\end{itemize}

The default data model is defined in the core \biblatex file \file{blx-dm.def} using the macros described in this section. The default data model is described in detail in \secref{bib}. The data model is used internally by \biblatex and also by the backend. In practice, changing the data model means that you can define the entrytypes and fields for your datasources and validate your data against the data model. Naturally, this is not much use unless your style supports any new entrytypes or fields and it raises issues of portability between styles (although this can be mitigated by using the dynamic data modification functionality described in \secref{aut:ctm:map}).

Note that while the \biber/\bibtex input site is not case sensitive when it comes to entry types and field names (Perl's Unicode case folding is used to normalise field names and entry types), the \latex side is case sensitive and uses the exact capitalisation from the data model.

Validation against the data model means that after mapping your data sources into the data model, \biber (using its \path{--validate_datamodel} option) can check:

\begin{itemize}
\item Whether all entrytypes are valid entrytypes
\item Whether all fields are valid fields for their entrytype
\item Whether the fields obey various constraints on their format which you specify
\end{itemize}
%
Redefining the data model can be done in several places. Style authors can create a \file{.dbx} file which contains the data model macros required and this will be loaded automatically when using the \biblatex package \opt{style} option by looking for a file named after the style with a \file{.dbx} extension (just like the \file{.cbx} and \file{.bbx} files for a style). If the \opt{style} option is not used but rather the \opt{citestyle} and \opt{bibstyle} options, then the package will try to load \file{.dbx} files called \file{$<$citestyle$>$.dbx} and \file{$<$bibstyle$>$.dbx}.
Alternatively, the name of the data model file can be different from any of the style option names by specifying the name (without \file{.dbx} extension) to the package \opt{datamodel} option. After loading the style data model file, \biblatex then loads, if present, a users \file{biblatex-dm.cfg} which should be put somewhere \biblatex can find it, just like the main configuration file \sty{biblatex.cfg}. To summarise, the data model is determined by adding to the data model from each of these locations, in order:\\

\noindent\file{blx-dm.def}$\rightarrow$\\
\hspace*{1em}\file{$<$datamodel option$>$.dbx} $\rightarrow$\\
\hspace*{2em}\file{$<$style option$>$.dbx} $\rightarrow$\\
\hspace*{3em}\file{$<$citestyle option$>$.dbx} and \file{$<$bibstyle option$>$.dbx} $\rightarrow$\\
\hspace*{4em}\file{biblatex-dm.cfg}\\

\noindent It is not possible to add to a loaded data model by using the macros below in your preamble as the preamble is read after \biblatex has defined critical internal macros based on the data model. If any data model macro is used in a document, it will be ignored and a warning will be generated. The data model is defined using the following macros:

\begin{ltxsyntax}

\cmditem{DeclareDatamodelConstant}[options]{name}{constantdef}

Declares the \prm{name} as a datamodel constant with definition \prm{constantdef}. Such constants are typically used internally by \biber.

\begin{optionlist*}

\choitem[string]{type}{string, list}

A constant can be a simple string (default if the \prm{type} option is omitted) or a comma"=separated list of strings.

\end{optionlist*}

\cmditem{DeclareDatamodelEntrytypes}[options]{entrytypes}

Declares the comma"=separated list of \prm{entrytypes} to be valid entrytypes in the data model. As usual in \tex csv lists, make sure each element is immediately followed by a comma or the closing brace---no extraneous whitespace.

\begin{optionlist*}

\boolitem[false]{skipout}

This entrytype is not output to the \file{.bbl}. Typically used for special entrytypes which are processed and consumed by the backend such as \bibtype{xdata}.

\end{optionlist*}

\cmditem{DeclareDatamodelFields}[options]{fields}

Declares the comma"=separated list of \prm{fields} to be valid fields in the data model with associated comma"=separated \prm{options}. The \prm{type} and \prm{datatype} options are mandatory. All valid \prm{options} are:

\begin{optionlist*}

\valitem{type}{field type}

Set the type of the field as described in \secref{bib:fld:typ}, typically <field> or <list>.

\valitem{format}{field format}

Any special format of the field. Normally unspecified but can take the value <xsv> which tells \biber that this field has a separated values format. The exact separator can be controlled with the \biber option \opt{xsvsep} and defaults to the expected comma surrounded by optional whitespace.

\valitem{datatype}{field datatype}

Set the datatype of the field as described in \secref{bib:fld:typ}. For example, <name> or <literal>.

\boolitem[false]{nullok}

The field is allowed to be defined but empty.

\boolitem[false]{skipout}

The field is not output to the \file{.bbl} and is therefore not present during \biblatex style processing. As usual in \tex csv lists, make sure each element is immediately followed by a comma or the closing brace---no extraneous whitespace.

\boolitem[false]{label}

The field can be used as a label in a bibliography or bibliography list. Specifying this causes \biblatex to create several helper macros for the field so that there are some internal lengths and headings etc. defined.

\end{optionlist*}

\cmditem{DeclareDatamodelEntryfields}[entrytypes]{fields}

Declares that the comma"=separated list of \prm{fields} is valid for the comma"=separated list of \prm{entrytypes}. If \prm{entrytypes} is not given, the fields are valid for all entrytypes. As usual in \tex csv lists, make sure each element is immediately followed by a comma or the closing brace---no extraneous whitespace.

\cmditem{DeclareDatamodelConstraints}[entrytypes]{specification}

If a comma"=separated list of \prm{entrytypes} is given, the constraints apply only to those entrytypes. The \prm{specification} is an undelimited list of \cmd{constraint} directives which specify a constraint. Spaces, tabs, and line endings may be used freely to visually arrange the \prm{specification}. Blank lines are not permissible.

\cmditem{constraint}[type=constrainttype]{elements}

Specifies a constraint of type \prm{constrainttype}. Valid constraint types are:

\begin{optionlist*}

\choitem{type}{data, mandatory, conditional}

Constraints of type <data> put restrictions on the value of a field. Constraints of type <mandatory> specify which fields or combinations of fields an entrytype should have. Constraints of type <conditional> allow more sophisticated conditional and quantified field constraints.

\choitem{datatype}{integer, isbn, issn, ismn, datepart, pattern}

For constraints of type \prm{data}, constrain field values to be the given datatype.

\valitem{rangemin}{num}

For constraints of \prm{type} <data> and \prm{datatype} <integer>, constrain field values to be at least \prm{num}.

\valitem{rangemax}{num}

For constraints of \prm{type} <data> and \prm{datatype} <integer>, constrain field values to be at most \prm{num}.

\valitem{pattern}{patt}

For constraints of \prm{type} <data> and \prm{datatype} <pattern>, constrain field values to match regular expression pattern \prm{patt}. It is best to wrap any regular expression in the macro \cmd{regexp}, see \secref{aut:ctm:map}.

\end{optionlist*}

A \cmd{constraint} macro may contain any of the following:

\cmditem{constraintfieldsor}{fields}

For constraints of \prm{type} <mandatory>, specifies that an entry must contain a boolean OR of the \cmd{constraintfield}s.

\cmditem{constraintfieldsxor}{fields}

For constraints of \prm{type} <mandatory>, specifies that an entry must contain a boolean XOR of the \cmd{constraintfield}s.

\cmditem{antecedent}[quantifier=quantspec]{fields}

For constraints of \prm{type} <conditional>, specifies a quantified set of \cmd{constraintfield}s which must be satisfied before the \cmd{consequent} of the constraint is checked. \prm{quantspec} should have one of the following values:

\begin{optionlist*}

\choitem{quantifier}{all, one, none}

Specifies how many of the \cmd{constrainfield}'s inside the \cmd{antecedent} have to be present to satisfy the antecedent of the conditional constraint.

\end{optionlist*}

\cmditem{consequent}[quantifier=quantspec]{fields}

For constraints of \prm{type} <conditional>, specifies a quantified set of \cmd{constraintfield}s which must be satisfied if the preceding \cmd{antecedent} of the constraint was satisfied. \prm{quantspec} should have one of the following values:

\begin{optionlist*}

\choitem{quantifier}{all, one, none}

Specifies how many of the \cmd{constraintfield}'s inside the \cmd{consequent} have to be present to satisfy the consequent of the conditional constraint.

\end{optionlist*}

\cmditem{constraintfield}{field}

For constraints of \prm{type} <data>, the constraint applies to this \prm{field}. For constraints of \prm{type} <mandatory>, the entry must contain this \prm{field}.

The data model declaration macros may be used multiple times as they append to the previous definitions. In order to replace, change or remove existing definitions (such as the default model which is loaded with \biblatex), you should reset (clear) the current definition and then set what you want using the following macros. Typically, these macros will be the first things in any \file{biblatex-dm.cfg} file:

\cmditem{ResetDatamodelEntrytypes}

Clear all data model entrytype information.

\cmditem{ResetDatamodelFields}

Clear all data model field information.

\cmditem{ResetDatamodelEntryfields}

Clear all data model fields for entrytypes information.

\cmditem{ResetDatamodelConstraints}

Clear all data model fields Constraints information.

\end{ltxsyntax}

Here is an example of a simple data model. Refer to the core \biblatex file \file{blx-dm.def} for the default data model specification.

\begin{ltxexample}
\ResetDatamodelEntrytypes
\ResetDatamodelFields
\ResetDatamodelEntryfields
\ResetDatamodelConstraints

\DeclareDatamodelEntrytypes{<<entrytype1, entrytype2>>}

\DeclareDatamodelFields[type=field, datatype=literal]{<<field1,field2,field3,field4>>}

\DeclareDatamodelEntryfields{<<field1>>}
\DeclareDatamodelEntryfields[entrytype1]{<<field2,field3>>}
\DeclareDatamodelEntryfields[entrytype2]{<<field2,field3,field4>>}

\DeclareDatamodelConstraints[<<entrytype1>>]{
  \constraint[type=data, datatype=integer, rangemin=3, rangemax=10]{
    \constraintfield{<<field1>>}
  }
  \constraint[type=mandatory]{
    \constraintfield{<<field1>>}
    \constraintfieldsxor{
      \constraintfield{<<field2>>}
      \constraintfield{<<field3>>}
    }
  }
}
\DeclareDatamodelConstraints{
  \constraint[type=conditional]{
    \antecedent[quantifier=none]{
      \constraintfield{<<field2>>}
    }
    \consequent[quantifier=all]{
      \constraintfield{<<field3>>}
      \constraintfield{<<field4>>}
    }
  }
}
\end{ltxexample}
%
This model specifies:

\begin{itemize}
\item Clear the default data model completely
\item Two valid entry types \bibtype{entrytype1} and \bibtype{entrytype2}
\item Four valid literal field fields
\item \bibfield{field1} is valid for all entrytypes
\item \bibfield{field2} and \bibfield{field3} are valid for \bibfield{entrytype1}
\item \bibfield{field2}, \bibfield{field3} and \bibfield{field4} are valid for \bibtype{entrytype2}
\item For \bibtype{entrytype1}:
  \begin{itemize}
  \item \bibfield{field1} must be an integer between 3 and 10
  \item \bibfield{field1} must be present
  \item One and only one of \bibfield{field2} or \bibfield{field3} must be present
  \end{itemize}
\item For any entrytype, if \bibfield{field2} is not present, \bibfield{field3} and \bibfield{field4} must be present
\end{itemize}

\subsubsection{Labels}
\label{aut:ctm:lab}

Alphabetic styles use a label to identify bibliography entries. This label is constructed from components of the entry using a template which describes how to build the label. The template can be customised on a global or per-type basis. A separate template is used to specify how to extract parts of name fields for labels, since names can be quite complex fields.

\begin{ltxsyntax}

\cmditem{DeclareLabelalphaTemplate}[]{specification}

Defines the alphabetic label template for the given entrytypes. If no entrytypes are specified in the first argument, then the global label template is defined. The \prm{specification} is an undelimited list of \cmd{labelelement} directives which specify the elements used to build the label. Spaces, tabs, and line endings may be used freely to visually arrange the \prm{specification}. Blank lines are not permissible. This command may only be used in the preamble.

\cmditem{labelelement}{elements}

Specifies the elements used to build the label. The \prm{elements} are an undelimited list of \cmd{field} or \cmd{literal} commands which are evaluated in the order in which they are given. The first \cmd{field} or \cmd{literal} which expands to a non-empty string is used as the \cmd{labelelement} expansion and the next \cmd{labelelement}, if any, is then processed.

\cmditem{field}[options]{field}

If \prm{field} is non-empty, use it as the current label \cmd{labelelement}, subject to the options below. Useful values for \prm{field} are typically the name list type fields, date fields, and title fields. You may also use the `citekey' or `entrykey` pseudo-fields to specify the citation/entry key as part of the label. Name list fields are treated specially and when a name list field is specified, the template defined with \cmd{DeclareLabelalphaNameTemplate} is used to extract parts from the name which then returns the string that the \cmd{field} option uses.

\begin{optionlist*}

\boolitem[false]{final}

This option marks a \cmd{field} directive as the final one in the \prm{specification}. If the \prm{field} is non-empty, then this field is used for the label and the remainder of the \prm{specification} will be ignored. The short form \opt{final} is equivalent to \kvopt{final}{true}.

\boolitem[false]{lowercase}

Forces the label part derived from the field to lowercase. By default, the case is taken from the field source and not modified.

\intitem[1]{strwidth}

The number of characters of the \prm{field} to use. This setting may be overridden by an individual name part when extracting characters from a name. See \cmd{DeclareLabelalphaNameTemplate} below.

\choitem[left]{strside}{left, right}

The side of the string from which to take the \texttt{strwidth} number of characters. This setting may be overridden by an individual name part when extracting characters from a name. See \cmd{DeclareLabelalphaNameTemplate} below.

\choitem[right]{padside}{left, right}

Side to pad the label part when using the \texttt{padchar} option. Only for use with fixed-width label strings (\texttt{strwidth}).

\valitem{padchar}{character}

If present, pads the label part on the \texttt{padside} side with the specified character to the length of \texttt{strwidth}. Only for use with fixed-width label strings (\texttt{strwidth}).

\boolitem[false]{uppercase}

Forces the label part derived from the field to uppercase. By default, the case is taken from the field source and not modified.

\boolitem[false]{varwidth}

Use a variable width, left-side substring of characters from the string returned for \prm{field}. The length of the string is determined by the minimum length needed to disambiguate the substring from all other \prm{field} elements in the same position in the label. For name list fields, this means that each name substring is disambiguated from all other name substrings which occur in the same position in the name list (see examples below). This option overrides \texttt{strwidth} if both are used. The short form \opt{varwidth} is equivalent to \kvopt{varwidth}{true}. For name list fields, the \cmd{namepart}s with the \opt{pre} option set are prepended to the string returned from this disambiguation.

\boolitem[false]{varwidthnorm}

As \texttt{varwidth} but will force the disambiguated substrings for the \prm{field} to be the same length as the longest disambiguated substring. This can be used to regularise the format of the labels if desired. This option overrides \texttt{strwidth} if both are used. The short form \opt{varwidthnorm} is equivalent to \kvopt{varwidthnorm}{true}.

\boolitem[false]{varwidthlist}

Alternative method of automatic label disambiguation where the field as a whole is disambiguated from all other fields in the same label position. For non-name list fields, this is equivalent to \texttt{varwidth}. For name list fields, names in a name list are not disambiguated from other names in the same position in their name lists but instead the entire name list is disambiguated as a whole from other name lists (see examples below). This option overrides \texttt{strwidth} if both are used. The short form \opt{varwidthlist} is equivalent to \kvopt{varwidthlist}{true}.  For name list fields, the \cmd{namepart}s with the \opt{pre} option set are prepended to the string returned from this disambiguation.

\intitem{strwidthmax}

When using \texttt{varwidth}, this option sets a limit (in number of characters) on the length of variable width substrings. This option can be used to regularise the label.

\intitem[1]{strfixedcount}

When using \texttt{varwidthnorm}, there must be at least \texttt{strfixedcount} disambiguated substrings with the same, maximal length to trigger the forcing of all disambiguated substrings to this same maximal length.

\valitem{ifnames}{range}

Only use this \cmd{field} specification if it is a name list field with a number of names matching the \texttt{ifnames} range value. This allows a \cmd{labelelement} to be conditionalised on name length (see below). The range can specified as in the following examples:

\begin{lstlisting}[language=xml]
ifnames=3     -> Only apply to name lists containing exactly 3 names
ifnames={2-4} -> Only apply to name lists containing minimum 2 and maximum 4 names
ifnames={-3}  -> Only apply to name lists containing at most 3 names
ifnames={2-}  -> Only apply to name lists containing at least 2 names
\end{lstlisting}

\valitem{names}{range}

By default, for name list fields, the names used range from the first name to the \cnt{maxalphanames}\slash \cnt{minalphanames} truncation. This option can be used to override this with an explicit range of names to consider. The plus <+> sign is a special end of range marker denoting the truncation point of max/minalphanames. The range separator can be any number of characters with the Unicode Dash property. For example:

\begin{lstlisting}[language=xml]
names=3     -> Use first 3 names in the name list
names={2-3} -> Use second and thirds names only
names={-3}  -> Same as 1-3
names={2-}  -> Use all names starting with the second name (ignoring max/minalphanames truncation)
names={2-+} -> Use all names starting with the second name (respecting max/minalphanames truncation)
\end{lstlisting}

\valitem[empty]{namessep}{string}

An arbitrary string separator to put between names in a namelist.

\boolitem[false]{noalphaothers}

By default, \cmd{labelalphaothers} is appended to label parts derived from name lists if there are more names in the list than are shown in the label part. This option can be used to disable the default behaviour.

\end{optionlist*}

\cmditem{literal}{characters}

Insert the literal \prm{characters} into the label at this point.

\end{ltxsyntax}
%
When a name list \cmd{field} is specified, the method of extracting the string is specified by a separate template specified by the following command:

\begin{ltxsyntax}

\cmditem{DeclareLabelalphaNameTemplate}[name]{specification}

Defines the \opt{labelalphaname} template \prm{name}. The \prm{name} is optional and defaults to \prm{<global>}.

Such templates specify how to extract a label string from a name list when a \cmd{field} specification in \cmd{DeclareLabelalphaTemplate} contains a name list.

\cmditem{namepart}[options]{namepart}

\prm{namepart} is one of the datamodel nameparts defined with the \cmd{DeclareDatamodelConstant} command (see \secref{aut:bbx:drv}). The \prm{options} are:

\begin{optionlist*}

\boolitem[false]{use}

Only use the \prm{namepart} in constructing the label information if there is a corresponding option \opt{use<namepart>} and that option is true.

\boolitem[false]{pre}

When constructing label strings from names, the \cmd{namepart} \emph{without} a \opt{pre} option will be used to construct label string, passing through disambiguation, substring etc. operations as specified by the \cmd{field} options in \cmd{DeclareLabelalpaTemplate}. Then the \cmd{namepart} options \emph{with} the \opt{pre} option set will be prepended to the result, (in the order given, if there are more than one such \cmd{namepart}s). This allows to unconditionally prepend certain namepart information to name label strings, like name prefices. Note that the \opt{uppercase} and \opt{lowercase} options of \cmd{field} in \cmd{DeclareLabelalphaTemplate} are applied to the entire label returned from \cmd{DeclareLabelalphaTemplate}, both \opt{pre} parts and non \opt{pre}.

\boolitem[false]{compound}

For static (non-varwidth) disambiguation in \cmd{DeclareLabelalphaTemplate}, treat nameparts separated by whitespace or hyphens (compound names) as separate names for label generation. This means that when forming a label out of, for example the surname <Ballam Forsyth> with a 1 character, left-side substring, this name would give <BF> with \kvopt{compound}{true} and <B> with \kvopt{compound}{false}. The short form \opt{compound} is equivalent to \kvopt{compound}{true}.

\intitem[1]{strwidth}

The number of characters of the \prm{namepart} to use.

\choitem[left]{strside}{left, right}

The side of the string from which to take the \texttt{strwidth} number of characters.

\end{optionlist*}

\end{ltxsyntax}

Note that the templates for labels can be defined per-type and you should be aware of this when using the automatically disambiguated label functionality. Disambiguation is not per-type as this might lead to ambiguity due to different label formats for different types being isolated from each others disambiguation process. Normally, you will want to use very different label formats for different types to make the type obvious by the label.

Here are some examples. The default global \biblatex alphabetic label template is defined below. Firstly, \bibfield{shorthand} has \kvopt{final}{true} and so if there is a \bibfield{shorthand} field, it is used as the label and nothing more of the template is considered. Next, the \bibfield{label} field is used as the first label element if it exists. Otherwise, if there is only one name (\kvopt{ifnames}{1}) in the \bibfield{labelname} list, then three characters from the left side of the family name in the \bibfield{labelname} are used as the first label element. If the \bibfield{labelname} has more than one name in it, one character from the left side of each family name is used as the first label element. The second label element consists of 2 characters from the right side of the \bibfield{year} field.

The default template for constructing labels from names is also shown. This prepends the first character from the left side of any prefix (if the \opt{useprefix} option is true) to a label extracted from the family name (according to the options on the calling \cmd{field} option from \cmd{DeclareLabelalphaTemplate}), allowing for compound family names.

\begin{ltxexample}
\DeclareLabelalphaTemplate{
  \labelelement{
    \field[final]{<<shorthand>>}
    \field{<<label>>}
    \field[strwidth=3,strside=left,ifnames=1]{<<labelname>>}
    \field[strwidth=1,strside=left]{<<labelname>>}
  }
  \labelelement{
    \field[strwidth=2,strside=right]{<<year>>}
  }
}

\DeclareLabelalphaNameTemplate{
  \namepart[use=true, pre=true, strwidth=1, compound=true]{prefix}
  \namepart{family}
}

\end{ltxexample}
%
To get an idea of how the label automatic disambiguation works, consider the following author lists:

\begin{lstlisting}{}
Agassi, Chang, Laver   (2000)
Agassi, Connors, Lendl (2001)
Agassi, Courier, Laver (2002)
Borg, Connors, Edberg  (2003)
Borg, Connors, Emerson (2004)
\end{lstlisting}
%
Assuming a template declaration such as:

\begin{ltxexample}
\DeclareLabelalphaTemplate{
  \labelelement{
    \field[varwidth]{<<labelname>>}
  }
}
\end{ltxexample}
%
Then the labels would be:

\begin{lstlisting}{}
Agassi, Chang, Laver    [AChLa]
Agassi, Connors, Lendl  [AConLe]
Agassi, Courier, Laver  [ACouLa]
Borg, Connors, Edberg   [BConEd]
Borg, Connors, Emerson  [BConEm]
\end{lstlisting}
%
With normalised variable width labels defined:

\begin{ltxexample}
\DeclareLabelalphaTemplate{
  \labelelement{
    \field[varwidthnorm]{<<labelname>>}
  }
}
\end{ltxexample}
%
You would get the following as the substrings of names in each position are extended to the length of the longest substring in that same position:

\begin{lstlisting}{}
Agassi, Chang, Laver    [AChaLa]
Agassi, Connors, Lendl  [AConLe]
Agassi, Courier, Laver  [ACouLa]
Borg, Connors, Edberg   [BConEd]
Borg, Connors, Emerson  [BConEm]
\end{lstlisting}
%
With a restriction to two characters for the name components of the label element defined like this:

\begin{ltxexample}
\DeclareLabelalphaTemplate{
  \labelelement{
    \field[varwidthnorm,strwidthmax=2]{<<labelname>>}
  }
}
\end{ltxexample}
%
This would be the result (note that the individual family name label parts are no longer unambiguous):

\begin{lstlisting}{}
Agassi, Chang, Laver    [AChLa]
Agassi, Connors, Lendl  [ACoLe]
Agassi, Courier, Laver  [ACoLa]
Borg, Connors, Edberg   [BCoEd]
Borg, Connors, Emerson  [BCoEm]
\end{lstlisting}
%
Alternatively, you could choose to disambiguate the name lists as a whole with:

\begin{ltxexample}
\DeclareLabelalphaTemplate{
  \labelelement{
    \field[varwidthlist]{<<labelname>>}
  }
}
\end{ltxexample}
%
Which would result in:

\begin{lstlisting}{}
Agassi, Chang, Laver    [AChL]
Agassi, Connors, Lendl  [ACoL]
Agassi, Courier, Laver  [ACL]
Borg, Connors, Edberg   [BCEd]
Borg, Connors, Emerson  [BCE]
\end{lstlisting}
%
Perhaps you only want to consider at most two names for label generation but disambiguate at the whole name list level:
\begin{ltxexample}
\DeclareLabelalphaTemplate{
  \labelelement{
    \field[varwidthlist,names=2]{<<labelname>>}
  }
}
\end{ltxexample}
%
Which would result in:
\begin{lstlisting}{}
Agassi, Chang, Laver    [ACh+]
Agassi, Connors, Lendl  [ACo+]
Agassi, Courier, Laver  [AC+]
Borg, Connors, Edberg   [BC+a]
Borg, Connors, Emerson  [BC+b]
\end{lstlisting}
%
In this last example, you can see \cmd{labelalphaothers} has been appended to show that there are more names. The last two labels now require disambiguating with \cmd{extraalpha} as there is no way of disambiguating this label name list with only two names.

Finally, here is an example using multiple label elements:

\begin{ltxexample}
\DeclareLabelalphaTemplate{
  \labelelement{
    \field[varwidthlist]{<<labelname>>}
  }
  \labelelement{
    \literal{-}
  }
  \labelelement{
    \field[strwidth=3,strside=right]{<<labelyear>>}
  }
}
\end{ltxexample}
%
Which would result in:
\begin{lstlisting}{}
Agassi, Chang, Laver    [AChL-000]
Agassi, Connors, Lendl  [AConL-001]
Agassi, Courier, Laver  [ACouL-002]
Borg, Connors, Edberg   [BCEd-003]
Borg, Connors, Emerson  [BCEm-004]
\end{lstlisting}
%
Here is another rather contrived example showing that you don't need to specially quote \latex special characters (apart from <\%>, obviously) when specifying padding characters and literals:

\begin{ltxexample}
\DeclareLabelalphaTemplate{
  \labelelement{
    \literal{>}
  }
  \labelelement{
    \literal{\%}
  }
  \labelelement{
    \field[namessep={/}, strwidth=4, padchar=_]{<<labelname>>}
  }
  \labelelement{
    \field[strwidth=3, padchar=&, padside=left]{title}
  }
  \labelelement{
    \field[strwidth=2,strside=right]{<<year>>}
  }
}
\end{ltxexample}
%
which given:
\begin{lstlisting}[style=bibtex]{}
@Book{test,
  author    = {XXX YY and WWW ZZ},
  title     = {T},
  year      = {2007},
}
\end{lstlisting}
would resulting a label looking like this:
\begin{verbatim}
[>%YY/ZZ__&&T07]
\end{verbatim}

Generating labels from fields may involve some difficulties when you have fields containing diacritics, hyphens, spaces etc. Often, you want to ignore things like separator characters or spaces when generating labels. An option is provided to customise the regular expression(s) to strip from a field before it is passed to the label generation system.

\begin{ltxsyntax}

\cmditem{DeclareNolabel}{specification}

Defines regular expressions to strip from any field before generating a label part for the field. The \prm{specification} is an undelimited list of \cmd{nolabel} directives which specify the regular expressions to remove from fields. Spaces, tabs and line endings may be used freely to visually arrange the \prm{specification}. Blank lines are not permissible. This command may only be used in the preamble.

\cmditem{nolabel}{regexp}

Any number of \cmd{nolabel} commands can be given each of which specifies to remove the \prm{regexp} from the copy of the field which the label generation system sees. Since regular expressions usually contain special characters, it is best to enclose them in the provided \cmd{regexp} macro as shown---this will pass the expression through to \biber correctly.

\end{ltxsyntax}

If there is no \cmd{DeclareNolabel} specification, \biber will default to:

\begin{ltxexample}
\DeclareNolabel{
  % strip punctuation, symbols, separator and control characters
  \nolabel{\regexp{<<[\p{P}\p{S}\p{C}]+>>}}
}
\end{ltxexample}
%
This \biber default strips punctuation, symbol, separator and control characters from fields before passing the field string to the label generation system.

\begin{ltxsyntax}

\cmditem{DeclareNolabelwidthcount}{specification}

Defines regular expressions to ignore from any field when counting characters in fixed-width substrings. The \prm{specification} is an undelimited list of \cmd{nolabelwidthcount} directives which specify the regular expressions to ignore when counting characters for fixed-width substrings. Spaces, tabs and line endings may be used freely to visually arrange the \prm{specification}. Blank lines are not permissible. This command may only be used in the preamble.

\cmditem{nolabelwidthcount}{regexp}

Any number of \cmd{nolabelwidthcount} commands can be given each of which specifies to ignore the \prm{regexp} when generating fixed-width substrings during label generation. Since regular expressions usually contain special characters, it is best to enclose them in the provided \cmd{regexp} macro as shown---this will pass the expression through to \biber correctly.

\end{ltxsyntax}

There is no default \cmd{DeclareNolabelwidthcount} specification. Note that
this setting is only taken into account when using fixed-width substrings
(non-varwidth) during label part generation. See \secref{aut:ctm:lab}.

\subsubsection{Sorting}
\label{aut:ctm:srt}

In addition to the predefined sorting templates discussed in \secref{use:srt}, it is possible to define new ones or modify the default definitions. The sorting process may be customized further by excluding certain fields from sorting on a per-type basis and by automatically populating the \bibfield{presort} field on a per-type basis.

\begin{ltxsyntax}

\cmditem{DeclareSortingTemplate}[options]{name}{specification}

Defines the sorting template \prm{name}. The \prm{name} is the identifier passed to the \opt{sorting} option (\secref{use:opt:pre:gen}) when selecting the sorting template. The \cmd{DeclareSortingTemplate} command supports the following optional arguments:

\begin{optionlist*}

\choitem{locale}{\prm{locale}}

The locale for the sorting template which then overrides the global sorting locale in the \opt{sortlocale} option discussed in \secref{use:opt:pre:gen}.

\end{optionlist*}

The \prm{specification} is an undelimited list of \cmd{sort} directives which specify the elements to be considered in the sorting process. Spaces, tabs, and line endings may be used freely to visually arrange the \prm{specification}. Blank lines are not permissible. This command may only be used in the preamble.

\cmditem{sort}{elements}

Specifies the elements considered in the sorting process. The \prm{elements} are an undelimited list of \cmd{field}, \cmd{literal}, and \cmd{citeorder} commands which are evaluated in the order in which they are given. If an element is defined, it is added to the sort key and the sorting routine skips to the next \cmd{sort} directive. If it is undefined, the next element is evaluated. Since literal strings are always defined, any \cmd{literal} commands should be the sole or the last element in a \cmd{sort} directive. All \prm{elements} should be the same datatype as described in \secref{bib:fld:dat} since they will be potentially compared to any of the other \prm{elements} in other entries.. The \cmd{sort} command supports the following optional arguments:

\begin{optionlist*}

\choitem{locale}{\prm{locale}}

Override the locale used for sorting at the level of a particular set of sorting elements. If specified, the locale overrides the locale set at the level of \cmd{DeclareSortingTemplate} and also the global setting. See also the discussion of the global sorting locale option \opt{sortlocale} in \secref{use:opt:pre:gen}.

\choitem[ascending]{direction}{ascending, descending}

The sort direction, which may be either \texttt{ascending} or \texttt{descending}. The default is ascending order.

\boolitem[false]{final}

This option marks a \cmd{sort} directive as the final one in the \prm{specification}. If one of the \prm{elements} is available, the remainder of the \prm{specification} will be ignored. The short form \opt{final} is equivalent to \kvopt{final}{true}.

\boolitem{sortcase}

Whether or not to sort case"=sensitively. The default setting depends on the global \opt{sortcase} option.

\boolitem{sortupper}

Whether or not to sort in <uppercase before lowercase> (\texttt{true}) or <lowercase before uppercase> order (\texttt{false}). The default setting depends on the global \opt{sortupper} option.

\end{optionlist*}

\cmditem{field}[key=value, \dots]{field}

The \cmd{field} element adds a \prm{field} to the sorting specification. If the \prm{field} is undefined, the element is skipped. The \cmd{field} command supports the following optional arguments:

\begin{optionlist*}

\choitem[left]{padside}{left, right}

Pads a field on the \texttt{left} or \texttt{right} side using \opt{padchar} so that its width is \opt{padwidth}. If no padding option is set, no padding is done at all. If any padding option is specified, then padding is performed and the missing options are assigned built-in default values. If padding and substring matching are both specified, the substring match is performed first.

\intitem[4]{padwidth}

The target width in characters.

\valitem[0]{padchar}{character}

The character to be used when padding the field.

\choitem[left]{strside}{left, right}

Performs a substring match on the \texttt{left} or \texttt{right} side of the field. The number of characters to match is specified by the corresponding \texttt{strwidth} option. If no substring option is set, no substring matching is performed at all. If any substring option is specified, then substring matching is performed and the missing options are assigned built-in default values. If padding and substring matching are both specified, the substring match is performed first.

\intitem[4]{strwidth}

The number of characters to match.

\end{optionlist*}

\cmditem{literal}{string}

The \cmd{literal} element adds a literal \prm{string} to the sorting specification. This is useful as a fallback if some fields are not available.

\csitem{citeorder}

The \cmd{citeorder} element has a special meaning. It requests a sort based on the lexical order of the actual citations. For entries cited within the same citation command like:

\begin{ltxexample}
\cite{one,two}
\end{ltxexample}
%
there is a distinction between the lexical order and the semantic order. Here «one» and «two» have the same semantic order but a unique lexical order. The semantic order only matters if you specify further sorting to disambiguate entries with the same semantic order. For example, this is the definition of the \opt{none} sorting template:

\begin{ltxexample}
\DeclareSortingTemplate{none}{
  \sort{\citeorder}
}
\end{ltxexample}
%
This sorts the bibliography purely lexically by the order of the keys in the citation commands. In the example above, it sorts «one» before «two». However, suppose that you consider «one» and «two» to have the same order (semantic order) since they are cited at the same time and want to further sort these by year. Suppose «two» has an earlier \bibfield{year} than «one»:

\begin{ltxexample}
\DeclareSortingTemplate{noneyear}{
  \sort{\citeorder}
  \sort{<<year>>}
}
\end{ltxexample}
%
This sorts «two» before «one», even though lexically, «one» would sort before «two». This is possible because the semantic order can be disambiguated by the further sorting on year. With the standard \opt{none} sorting template, the lexical order and semantic order are identical because there is nothing further to disambiguate them. This means that you can use \cmd{citeorder} just like any other sorting specification element, choosing how to further sort entries cited at the same time (in the same citation command).

\cmditem{DeclareSortingNamekeyTemplate}[name]{specification}

Defines how the sorting keys for names are constructed. This can change the sorting order of names arbitrarily because you can choose how to put together the name parts when constructing the string to compare when sorting. The sorting key construction template so defined is called \prm{name} which defaults to «global» if this optional parameter is absent. When constructing the sorting key for a name, a sorting key for each name part is constructed and the key for each name is formed into an ordered key list with a special internal separator. The point of this option is to accommodate languages or situations where sorting of names needs to be customised (for example, Icelandic names are sometimes sorted by given names rather than by family names). This macro may be used multiple times to define templates with different names which can then be referred to later. Sorting name key templates can be specified at the following scopes, in order of increasing precedence:

\begin{itemize}
\item The default template defined without the optional name argument
\item Given as the \opt{sortingnamekeytemplate} option to a reference context (see \secref{use:bib:context})
\item Given as a per-entry option \opt{sortnamekeytemplate} in a bibliography data source entry
\item Given as a per-namelist option \opt{sortnamekeytemplate}
\item Given as a per-name option \opt{sortnamekeytemplate}
\end{itemize}

By default there is only a global template which has the following \prm{specification}:

\begin{ltxexample}
\DeclareSortingNamekeyTemplate{
  \keypart{
    \namepart[use=true]{<<prefix>>}
    \namepart{<<family>>}
  }
  \keypart{
    \namepart{<<given>>}
  }
  \keypart{
    \namepart{<<suffix>>}
  }
  \keypart{
    \namepart[use=false]{<<prefix>>}
  }
}
\end{ltxexample}
%
This means that the key is constructed by concatenating, in order, the name prefix (only if the \opt{useprefix} option is true) with the family name(s), the given names(s), the name suffix and then the name prefix (only if the \opt{useprefix} option is false).

\cmditem{keypart}{part}

\prm{part} is an ordered list of of \cmd{namepart} and \cmd{literal} specifications which are concatenated together when constructing a part of the name sorting key. The \cmd{keyparts} are then concatenated together with terminal padding to ensure correct sorting.

\cmditem{literal}{string}

A literal string to insert into the name sorting key.

\cmditem{namepart}{name}

Specifies the \prm{name} of a namepart to use in constructing the name sorting key.

\begin{optionlist*}

\boolitem[true]{use}

Indicates that the namepart \prm{name} is only to be used in this concatenation position if the corresponding \opt{use<name>} option is set to the specified boolean value.

\boolitem[true]{inits}

Indicates that only the initials of namepart \prm{name} are to be used in
constructing the sorting specification.

\end{optionlist*}

\end{ltxsyntax}

As an example, suppose you wanted to be able to sort names by given name rather than family name, you could define a sorting name key template like this:

\begin{ltxexample}
\DeclareSortingNamekeyTemplate[givenfirst]{
  \keypart{
    \namepart{<<given>>}
  }
  \keypart{
    \namepart[use=true]{<<prefix>>}
  }
  \keypart{
    \namepart{<<family>>}
  }
  \keypart{
    \namepart[use=false]{<<prefix>>}
  }
}
\end{ltxexample}
%
You can then use the name \opt{givenfirst} at the appropriate scope in order to make \biber use this template when constructing sorting name keys. For example, you could enable this for one bibliography list like this:

\begin{ltxexample}
\begin{refcontext}[sortnamekeytemplate=givenfirst]
\printbibliography
\end{refcontext}
\end{ltxexample}
%
or perhaps you only want to do this for a particular entry:

\begin{lstlisting}[style=bibtex]{}
@BOOK{key,
  OPTIONS = {sortnamekeytemplate=givenfirst},
  AUTHOR = {Arnar Vigfusson}
}
\end{lstlisting}
%
or just a name list by using the option as a pseudo-name which will be ignored:

\begin{lstlisting}[style=bibtex]{}
@BOOK{key,
  AUTHOR = {sortnamekeytemplate=givenfirst and Arnar Vigfusson}
}
\end{lstlisting}
%
or just a single name by passing the option as part of the extended name information
format which \biber supports (see \secref{use:enf}):

\begin{lstlisting}[style=bibtex]{}
@BOOK{key,
  AUTHOR = {given=Arnar, family=Vigfusson, sortnamekeytemplate=givenfirst}
}
\end{lstlisting}
%
Now we give some examples of sorting templates. In the first example, we define a simple name\slash title\slash year template. The name element may be either the \bibfield{author}, the \bibfield{editor}, or the \bibfield{translator}. Given this specification, the sorting routine will use the first element which is available and continue with the \bibfield{title}. Note that the options \opt{use$<$name$>$} options are considered automatically in the sorting process:

\begin{ltxexample}
\DeclareSortingTemplate{sample}{
  \sort{
    \field{<<author>>}
    \field{<<editor>>}
    \field{<<translator>>}
  }
  \sort{
    \field{<<title>>}
  }
  \sort{
    \field{<<year>>}
  }
}
\end{ltxexample}
%
In the next example, we define the same template in a more elaborate way, considering special fields such as \bibfield{presort}, \bibfield{sortkey}, \bibfield{sortname}, etc. Since the \bibfield{sortkey} field specifies the master sort key, it needs to override all other elements except for \bibfield{presort}. This is indicated by the \opt{final} option. If the \bibfield{sortkey} field is available, processing will stop at this point. If not, the sorting routine continues with the next \cmd{sort} directive. This setup corresponds to the default definition of the \texttt{nty} template:

\begin{ltxexample}
\DeclareSortingTemplate{nty}{
  \sort{
    \field{<<presort>>}
  }
  \sort[final]{
    \field{<<sortkey>>}
  }
  \sort{
    \field{<<sortname>>}
    \field{<<author>>}
    \field{<<editor>>}
    \field{<<translator>>}
    \field{<<sorttitle>>}
    \field{<<title>>}
  }
  \sort{
    \field{<<sorttitle>>}
    \field{<<title>>}
  }
  \sort{
    \field{<<sortyear>>}
    \field{<<year>>}
  }
}
\end{ltxexample}
%
Finally, here is an example of a sorting template which overrides the global sorting locale and additionally overrides again when sorting by the \bibfield{origtitle} field. Note the use in the template-level override of a babel/polyglossia language name instead of a real locale identifier. \biber will map this to a suitable, real locale identifier (in this case, \texttt{sv\_SE}):

\begin{ltxexample}
\DeclareSortingTemplate[locale=swedish]{custom}{
  \sort{
    \field{<<sortname>>}
    \field{<<author>>}
    \field{<<editor>>}
    \field{<<translator>>}
    \field{<<sorttitle>>}
    \field{<<title>>}
  }
  \sort[locale=de_DE_phonebook]{
    \field{<<origtitle>>}
  }
}
\end{ltxexample}

\begin{ltxsyntax}

\cmditem{DeclareSortExclusion}{entrytype, \dots}{field, \dots}

Specifies fields to be excluded from sorting on a per-type basis. The \prm{entrytype} argument and the \prm{field} argument may be a comma"=separated list of values. A blank \prm{field} argument will clear all exclusions for this \prm{entrytype}. A value of <*> for \prm{entrytype} will exclude \prm{field,\dots} for every entrytype. This is equivalent to simply deleting the field from the sorting specification and is only normally used in combination with \cmd{DeclareSortInclusion} when one wishes to exclude a field for all but explicitly included entrytypes. See example in \cmd{DeclareSortInclusion} below. This command may only be used in the preamble.

\cmditem{DeclareSortInclusion}{entrytype, \dots}{field, \dots}

Only used along with \cmd{DeclareSortExclusion}. Specifies fields to be included in sorting on a per-type basis. This allows the user to exclude a field from sorting for all entrytypes and then to override this for certain entrytypes. This is easier sometimes than using \cmd{DeclareSortExclusion} to list exclusions for many entrytypes. The \prm{entrytype} argument and the \prm{field} argument may be a comma"=separated list of values. This command may only be used in the preamble. For example, this would use \bibfield{title} during sorting only for \bibtype{article}s:

\begin{ltxexample}
\DeclareSortExclusion{*}{title}
\DeclareSortInclusion{article}{title}
\end{ltxexample}

\cmditem{DeclarePresort}[entrytype, \dots]{string}

Specifies a string to be used to automatically populate the \bibfield{presort} field of entries without a \bibfield{presort} field. The \bibfield{presort} may be defined globally or on a per-type basis. If the optional \prm{entrytype} argument is given, the \prm{string} applies to the respective entry type. If not, it serves as the global default value. Specifying an \prm{entrytype} in conjunction with a blank \prm{string} will clear the type-specific setting. The \prm{entrytype} argument may be a comma"=separated list of values. This command may only be used in the preamble.

\cmditem{DeclareSortTranslit}[entrytype]{specification}

Languages which can be written in different scripts or alphabets often only have CLDR sorting tailoring for one script and it is expected that you transliterate into the supported script for sorting purposes. A common example is Sanskrit which is often written in academic contexts in IAST
romanised script but which needs to be sorted in the <sa> locale which expects the Devanāgarī script. Another common case is transliteration of Russian Cyrillic into Latin as defined by the ALA-LC standard. Such requirement means that it is necessary to transliterate into the sorting script internally. \cmd{DeclareSortTranslit} declares which parts of an entry you would like to transliterate for sorting purposes. Without the \prm{entrytype} parameter, the \prm{specification} applies to all entrytypes. The \prm{specification} is one or more \cmd{translit} commands:

\cmditem{translit}[langids]{field or fieldset}{from}{to}

Specifies that the data field \bibfield{field} or all fields in a fieldset \prm{fieldset} declared with \cmd{DeclareDatafieldSet} (see \secref{aut:ctm:dsets}) should be transliterated from script \prm{from} to script \prm{to} for sorting purposes. The field/set argument should be <*> to apply transliteration to all fields. The valid \prm{from} and \prm{to} values are given in table \ref{tab:translit}. The optional \prm{langids} parameter is a comma-separated list of \bibfield{langid} fields and the transliteration will apply only to bibliography entries containing one of the \bibfield{langid}s in the list. Note that \biblatex does not aim to support general transliteration, only those which are useful for sorting purposes. Please open a GitHub ticket for \biblatex\ if you think you need additional transliterations.

An example of transliterating titles so that they sort correctly in Sanskrit. This example assumes that entries that should have their title fields transliterated have a \bibfield{langid} field set to <sanskrit>.

\begin{ltxexample}
\DeclareDatafieldSet{settitles}{
  \member[field=title]
  \member[field=booktitle]
  \member[field=eventtitle]
  \member[field=issuetitle]
  \member[field=journaltitle]
  \member[field=maintitle]
  \member[field=origtitle]
}

\DeclareSortTranslit{
  \translit[sanskrit]{settitles}{iast}{devanagari}
}
\end{ltxexample}

\end{ltxsyntax}

\begin{table}
\caption{Valid transliteration pairs}
\label{tab:translit}
\tablesetup\centering
\begin{tabular}{lll}
\toprule
\sffamily\bfseries\spotcolor From
  & \sffamily\bfseries\spotcolor To
  & Description\\
\midrule
iast & devanagari & Sanskrit IAST to Devanāgarī\\
russian & ala-lc & ALA-LC romanisation for Russian\\
russian & bgn/pcgn-standard & BGN/PCGN:1947 (Standard Variant), Cyrillic to Latin, Russian\\
 \bottomrule
\end{tabular}
\end{table}

\subsubsection[Bibliography List Filters]{Bibliography List Filters}
\label{aut:ctm:bibfilt}

When using customisable bibliography lists (See \secref{use:bib:biblist}), usually one wants to return in the \file{.bbl} only those entries which have the particular fields which the bibliography list is summarising. For example, when printing a normal list of shorthands, you want the list returned by \biber in the \file{.bbl} to contain only those entries which have a \bibfield{shorthand} field. This is accomplished by defining a bibliography list filter using the \cmd{DeclareBiblistFilter} command. This differs from the filters defined using \cmd{defbibfilter} (see \secref{use:bib:flt}) since the filters defined by \cmd{defbibfilter} run inside \biblatex after the \file{.bbl} has been generated.

\begin{ltxsyntax}
\cmditem{DeclareBiblistFilter}{name}{specification}

Defines a bibliography list filter with \prm{name}. The \prm{specification} consists of one or more \cmd{filter} or \cmd{filteror} macros, all of which must be satisfied for the entry to pass the filter:

\cmditem{filter}[filterspec]{filter}

Filter entries according to the \prm{filterspec} and \prm{filter}. \prm{filterspec} can be one of:

\end{ltxsyntax}

\begin{description}
\item[type/nottype] Entry is/is not of \bibfield{entrytype} \prm{filter}
\item[subtype/notsubtype] Entry is/is not of \bibfield{subtype} \prm{filter}
\item[keyword/notkeyword] Entry has/does not have \bibfield{keyword} \prm{filter}
\item[field/notfield] Entry has/does not have a field called \prm{filter}
\end{description}

\begin{ltxsyntax}
\cmditem{filteror}{type}{filters}

A wrapper around one or more \cmd{filter} commands specifying that they form a disjunctive set, i.e. any one of the \prm{filters} must be satisfied.

\end{ltxsyntax}

Fields in the datamodel which are marked as <Label fields> (see \secref{aut:ctm:dm}) automatically have a filter defined for them with the same name and which filters out any entries which do no contain the field. For example, \biblatex automatically generates a filter for the \bibfield{shorthand} field:

\begin{ltxexample}
\DeclareBiblistFilter{<<shorthand>>}{
  \filter[type=field,filter=shorthand]
}
\end{ltxexample}

\subsubsection{Controlling Name Initials Generation}
\label{aut:ctm:noinit}

Generating initials for name parts from a given name involves some difficulties when you have names with prefixes, diacritics, hyphens etc. Often, you want to ignore things like prefixes when generating initials so that the initials for «al-Hasan» is just «H» instead of «a-H». This is tricky when you also have names like «Ho-Pun» where you want the initials to be «H-P», for example.

\begin{ltxsyntax}

\cmditem{DeclareNoinit}{specification}

Defines regular expressions to strip from names before generating initials. The \prm{specification} is an undelimited list of \cmd{noinit} directives which specify the regular expressions to remove from the name. Spaces, tabs and line endings may be used freely to visually arrange the \prm{specification}. Blank lines are not permissible. This command may only be used in the preamble.

\cmditem{noinit}{regexp}

Any number of \cmd{noinit} commands can be given each of which specifies to remove the \prm{regexp} from the copy of the name which the initials generation system sees. Since regular expressions usually contain special characters, it is best to enclose them in the provided \cmd{regexp} macro as shown---this will pass the expression through to \biber correctly.

\end{ltxsyntax}

If there is no \cmd{DeclareNoinit} specification, \biber will default to:

\begin{ltxexample}
\DeclareNoinit{
  % strip lowercase prefixes like 'al-' when generating initials from names
  \noinit{\regexp{<<\b\p{Ll}{2}\p{Pd}>>}}
  % strip some common diacritics when generating initials from names
  \noinit{\regexp{<<[\x{2bf}\x{2018}]>>}}
}
\end{ltxexample}
%
This \biber default strips a couple of diacritics and also strips lowercase prefixes from names before generating initials.

\subsubsection{Fine Tuning Sorting}
\label{aut:ctm:nosort}

It can be useful to fine tune sorting so that it ignores certain parts of particular fields.

\begin{ltxsyntax}

\cmditem{DeclareNosort}{specification}

Defines regular expressions to strip from particular fields or types of fields when sorting. The \prm{specification} is an undelimited list of \cmd{nosort} directives which specify the regular expressions to remove from particular fields or type of field. Spaces, tabs and line endings may be used freely to visually arrange the \prm{specification}. Blank lines are not permissible. This command may only be used in the preamble.

\cmditem{nosort}{field or datafield set}{regexp}

Any number of \cmd{nosort} commands can be given each of which specifies to remove the \prm{regexp} from the \prm{field} or \prm{datafield set}. A \prm{datafield set} is simply a convenience grouping of semantically similar fields from which you might want to remove a regexp. See \secref{aut:ctm:dsets} for the avaiable sets, their members and customisation. Since regular expressions usually contain special characters, it is best to enclose them in the provided \cmd{regexp} macro as shown---this will pass the expression through to \biber correctly.

\end{ltxsyntax}

The default is:

\begin{ltxexample}
\DeclareNosort{
  % strip prefixes like 'al-' when sorting names
  \nosort{setnames}{\regexp{<<\A\p{L}{2}\p{Pd}>>}}
  % strip some diacritics when sorting names
  \nosort{setnames}{\regexp{<<[\x{2bf}\x{2018}]>>}}
}
\end{ltxexample}
%
This \biber default strips a couple of diacritics and also strips two-letter prefixes (like «Al-») from names when sorting. Suppose you wanted to ignore «The» at the beginning of the \bibfield{title} field when sorting:

\begin{ltxexample}
\DeclareNosort{
  \nosort{<<title>>}{\regexp{<<\AThe\s+>>}}
}
\end{ltxexample}
%
Or if you wanted to ignore «The» at the beginning of any title field:

\begin{ltxexample}
\DeclareNosort{
  \nosort{<<settitles>>}{\regexp{<<\AThe\s+>>}}
}
\end{ltxexample}

\subsubsection{Special Fields}
\label{aut:ctm:fld}

Some of the automatically generated fields from \secref{aut:bbx:fld:lab} may be customized.

\begin{ltxsyntax}

\cmditem{DeclareLabelname}[entrytype, \dots]{specification}

Defines the fields to consider when generating the \bibfield{labelname} field (see \secref{aut:bbx:fld:lab}). The \prm{specification} is an ordered list of \cmd{field} commands. The fields are checked in the order listed and the first field which is available will be used as \bibfield{labelname}. This is the default definition:

\begin{ltxexample}
\DeclareLabelname{%
  \field{shortauthor}
  \field{author}
  \field{shorteditor}
  \field{editor}
  \field{translator}
}
\end{ltxexample}
%
The \bibfield{labelname} field may be customized globally or on a per-type basis. If the optional \prm{entrytype} argument is given, the specification applies to the respective entry type. If not, it is applied globally. The \prm{entrytype} argument may be a comma"=separated list of values. This command may only be used in the preamble.

\cmditem{DeclareLabeldate}[entrytype, \dots]{specification}

Defines the date components to consider when generating \bibfield{labelyear}, \bibfield{labelmonth}, \bibfield{labelday}, \bibfield{labelendyear}, \bibfield{labelendmonth} and \bibfield{labelendday} fields (see \secref{aut:bbx:fld:lab}). The \prm{specification} is an ordered list of \cmd{field} or \cmd{literal} commands. The items are checked in the order listed and the first item which is available will be used to popluate the mentioned fields. Note that the \cmd{field} items do not have to be datetype <date> in the data model so that you can create pseudo-year labels by, for example, using a \bibfield{pubstate} field contents, if available, as the year label by defining \cmd{DeclareLabeldate} suitably. Note also that a \cmd{literal} command will always be used when found and so this should always be the last thing in the list. If the value of a \cmd{literal} command is a valid localisation string, then this will be resolved in the current language, otherwise the value is used as a literal string as-is. This is the default definition:

\begin{ltxexample}
\DeclareLabeldate{%
  \field{date}
  \field{year}
  \field{eventdate}
  \field{origdate}
  \field{urldate}
  \literal{nodate}
}
\end{ltxexample}
%
Note that the \bibfield{date} field is split by the backend into \bibfield{year}, \bibfield{month} which are also valid fields in the default data model. In order to support legacy data which directly sets \bibfield{year} and/or \bibfield{month}, the specification <\bibfield{date}> in \cmd{DeclareLabeldate} will also match \bibfield{year} and \bibfield{month} fields, if present.
The \bibfield{label*} fields may be customized globally or on a per-type basis. If the optional \prm{entrytype} argument is given, the specification applies to the respective entry type. If not, it is applied globally. The \prm{entrytype} argument may be a comma"=separated list of values. This command may only be used in the preamble. See also \secref{aut:bbx:fld:dat}.

\cmditem{DeclareExtradate}{specification}

Defines how \biber tracks information used to construct the \bibfield{extradate} field. This field (see \secref{aut:bbx:fld:lab}) is printed to disambiguate works by the same author which occur in the same date scope. By default, the date scope is the year and so two works by the same author within the same year will have different \bibfield{extradate} values which are used to disambiguate the works in the bibliography in the usual manner seen in many authoryear type styles. The \prm{specification} is one or more \cmd{scope} specifications which can contain one or more \cmd{field} specifications. Within a \cmd{scope}, the existence of each \cmd{field} will be checked and if found, the first \cmd{field} is used and the rest are ignored. This allows a fallback in case certain fields are not available in all entries. All \cmd{scope}s are used to track information and \cmd{scope}s should be specified in decreasing order of generality (e.g. year then month then day etc) The default definition is:

\begin{ltxexample}
\DeclareExtradate{%
  \scope{
    \field{labelyear}
    \field{year}
  }
}
\end{ltxexample}
%
This means that the \bibfield{labelyear} field only (or \bibfield{year} if this does not exist) will be used to track works by the same author. With the following datasource entries:

\begin{lstlisting}[style=bibtex]{}
@BOOK{extra1,
  AUTHOR = {John Doe},
  DATE   = {2001-01}
}

@BOOK{extra2,
  AUTHOR = {John Doe},
  DATE   = {2001-02}
}
\end{lstlisting}
%
The default definition would result in:

\begin{lstlisting}{}
Doe 2001a
Doe 2001b
\end{lstlisting}
%
Here, \bibfield{extradate} only considers the \bibfield((label)year) information and since this is identical, disambiguation is required. However, consider the following definition:
\begin{ltxexample}
\DeclareExtradate{%
  \scope{
    \field{labelyear}
    \field{year}
  }
  \scope{
    \field{labelmonth}
  }
}
\end{ltxexample}
%
The result would be:
\begin{lstlisting}{}
Doe 2001
Doe 2001
\end{lstlisting}
%
If only years were printed, this would be ambiguous because \bibfield{extradate} now considers \bibfield{labelmonth} and since this differs, no disambiguation is necessary. Care should therefore be taken to synchronise the printed information with the \bibfield{extradate} disambiguation settings. Notice that the second definition is <month-in-year> disambiguation and quite different from:
\begin{ltxexample}
\DeclareExtradate{%
  \scope{
    \field{labelmonth}
  }
}
\end{ltxexample}
%
which is just plain <month> disambiguation which is very unlikely to be what you ever want to do since this disambiguation only based on month and ignores the year entirely. \bibfield{extradate} calculation should almost always be based on all information down to the resolution you require. For example, if you wish to disambiguate right down to the hour level (perhaps useful in large bibliographies of rapidly changing online material), you would specify something like this:
\begin{ltxexample}
\DeclareExtradate{%
  \scope{
    \field{labelyear}
    \field{year}
  }
  \scope{
    \field{labelmonth}
  }
  \scope{
    \field{labelday}
  }
  \scope{
    \field{labelhour}
  }
}
\end{ltxexample}
%
Entries without the specified granularity of information will disambiguate at the lowest granularity they contain, so, for example, with:
\begin{ltxexample}
\DeclareExtradate{%
  \scope{
    \field{labelyear}
    \field{year}
  }
  \scope{
    \field{labelmonth}
  }
}
\end{ltxexample}
%
\begin{lstlisting}[style=bibtex]{}
@BOOK{extra1,
  AUTHOR = {John Doe},
  DATE   = {2001}
}

@BOOK{extra2,
  AUTHOR = {John Doe},
  DATE   = {2001}
}
\end{lstlisting}
%
The result would still be:

\begin{lstlisting}{}
Doe 2001a
Doe 2001b
\end{lstlisting}
%
This command may only be used in the preamble.

\cmditem{DeclareLabeltitle}[entrytype, \dots]{specification}

Defines the fields to consider when generating the \bibfield{labeltitle} field (see \secref{aut:bbx:fld:lab}). The \prm{specification} is an ordered list of \cmd{field} commands. The fields are checked in the order listed and the first field which is available will be used as \bibfield{labeltitle}. This is the default definition:

\begin{ltxexample}
\DeclareLabeltitle{%
  \field{shorttitle}
  \field{title}
}
\end{ltxexample}
%
The \bibfield{labeltitle} field may be customized globally or on a per-type basis. If the optional \prm{entrytype} argument is given, the specification applies to the respective entry type. If not, it is applied globally. The \prm{entrytype} argument may be a comma"=separated list of values. This command may only be used in the preamble.

\end{ltxsyntax}

\subsubsection{Data Inheritance (\bibfield{crossref})}
\label{aut:ctm:ref}

\biber features a highly customizable cross-referencing mechanism with flexible data inheritance rules. This sections deals with the configuration interface. See \apxref{apx:ref} for the default configuration. A note on terminology: the \emph{child} or \emph{target} is the entry with the \bibfield{crossref} field, the \emph{parent} or \emph{source} is the entry the \bibfield{crossref} field points to. The child inherits data from the parent.

\begin{ltxsyntax}

\cmditem{DefaultInheritance}[exceptions]{options}

Configures the default inheritance behavior. This command may only be used in the preamble. The default behavior may be customized be setting the following \prm{options}:

\begin{optionlist*}

\boolitem[true]{all} Whether or not to inherit all fields from the parent by default.

\kvopt{all}{true} means that the child entry inherits all fields from the parent, unless a more specific inheritance rule has been set up with \cmd{DeclareDataInheritance}. If an inheritance rule is defined for a field, data inheritance is controlled by that rule. \kvopt{all}{false} means that no data is inherited from the parent by default and each field to be inherited requires an explicit inheritance rule set up with \cmd{DeclareDataInheritance}. The package default is \kvopt{all}{true}.

\boolitem[false]{override} Whether or not to overwrite target fields with source fields if both are defined. This applies both to automatic inheritance and to explicit inheritance rules. The package default is \kvopt{override}{false}, \ie existing fields of the child entry are not overwritten.

\valitem{ignore}{csv list of uniqueness options}

This option takes a comma-separated list of one of more of <singletitle>, <uniquetitle>, <uniquebaretitle> and/or <uniquework>. The purpose of this option is to ignore tracking information for these three options when the field which would trigger the tracking (\tabref{use:opt:wu}) is inherited. An example---Suppose that you have several \bibtype{book} entries which all crossref a \bibtype{mvbook} from which they get their \bibfield{author} field. You might reasonably want the \cmd{ifsingletitle} test to return <true> for this author as their only <work> is the \bibtype{mvbook}. Similar comments would apply to situations involving the \cmd{ifuniquetitle}, \cmd{ifuniquebaretitle} and \cmd{ifuniquework} tests. The \opt{ignore} option lists which of these should have their tracking information ignored when the fields which would trigger them are inherited. The idea is that the presence of an inherited field does not contribute towards the determination of whether some combination of name/title is unique in the bibliographic data. For example, this modified default setting would ignore \opt{singletitle} and \opt{uniquetitle} tracking:

\begin{ltxexample}
\DefaultInheritance{ignore={singletitle,uniquetitle}, all=true, override=false}
\end{ltxexample}
%
Of course, the ignoring of tracking does nothing if the fields inherited do not play a role in tracking. Only the fields listed in \tabref{use:opt:wu} are relevant to this option.

\end{optionlist*}

The optional \prm{exceptions} are an undelimited list of \cmd{except} directives. Spaces, tabs, and line endings may be used freely to visually arrange the \prm{exceptions}. Blank lines are not permissible.

\cmditem{except}{source}{target}{options}

Defines an exception to the default inheritance rules.

\cmd{DeclareDataInheritance} sets the inheritance \prm{options} for a specific \prm{source} and \prm{target} combination. The \prm{source} and \prm{target} arguments specify the parent and the child entry type. The asterisk matches all types and is permissible in either argument.

\cmditem{DeclareDataInheritance}[options]{source, \dots}{target, \dots}{rules}

Declares inheritance rules. The \prm{source} and \prm{target} arguments specify the parent and the child entry type. Either argument may be a single entry type, a comma"=separated list of types, or an asterisk. The asterisk matches all entry types. The \prm{rules} are an undelimited list of \cmd{inherit} and\slash or \cmd{noinherit} directives. Spaces, tabs, and line endings may be used freely to visually arrange the \prm{rules}. Blank lines are not permissible. This command may only be used in the preamble. The options are:

\begin{optionlist*}

\valitem{ignore}{csv list of uniqueness options}

As the \opt{ignore} option on \cmd{DefaultInheritance} explained above. When set here, it takes precedence over any global options set with \cmd{DefaultInheritance}. For example, this would ignore \opt{singletitle} and \opt{uniquetitle} tracking for a \bibtype{book} inheriting from a \bibtype{mvbook}.

\begin{ltxexample}
\DeclareDataInheritance[ignore={singletitle,uniquetitle}]{mvbook}{book}{<<...>>}
\end{ltxexample}

\end{optionlist*}

\cmditem{inherit}[option]{source}{target}

Defines an inheritance rule by mapping a \prm{source} field to a \prm{target} field. \prm{option} can be one of

\begin{optionlist*}

\boolitem[false]{override}

As the \opt{override} option for \cmd{DefaultInheritance} explained above. When set here, it takes precedence over any global options set with \cmd{DefaultInheritance}.

\end{optionlist*}

\cmditem{noinherit}{source}

Unconditionally prevents inheritance of the \prm{source} field.

\csitem{ResetDataInheritance}

Clears all inheritance rules defined with \cmd{DeclareDataInheritance}. This command may only be used in the preamble.

\end{ltxsyntax}

Here are some practical examples:

\begin{ltxexample}
\DefaultInheritance{<<all=true>>,<<override=false>>}
\end{ltxexample}
%
This example shows how to configure the default inheritance behavior. The above settings are the package defaults.

\begin{ltxexample}
\DefaultInheritance[
  \except{<<*>>}{<<online>>}{<<all=false>>}
]{all=true,override=false}
\end{ltxexample}
%
This example is similar to the one above but adds one exception: entries of type \bibtype{online} will, by default, not inherit any data from any parent.

\begin{ltxexample}
\DeclareDataInheritance{<<collection>>}{<<incollection>>}{
  \inherit{<<title>>}{<<booktitle>>}
  \inherit{<<subtitle>>}{<<booksubtitle>>}
  \inherit{<<titleaddon>>}{<<booktitleaddon>>}
}
\end{ltxexample}
%
So far we have looked at setting up standard inheritance. For example, \kvopt{all}{true} means that the \bibfield{publisher} field of a source entry is copied to the \bibfield{publisher} field of the target entry. In some cases, however, asymmetric mappings are required. They are defined with \cmd{DeclareDataInheritance}. The above example sets up three typical rules for \bibtype{incollection} entries referencing a \bibtype{collection}. We map the \bibfield{title} and related fields of the source to the corresponding \bibfield{booktitle} fields of the target.

\begin{ltxexample}
\DeclareDataInheritance{<<mvbook,book>>}{<<inbook,bookinbook>>}{
  \inherit{<<author>>}{<<author>>}
  \inherit{<<author>>}{<<bookauthor>>}
}
\end{ltxexample}
%
This rule is an example of one-to-many mapping: it maps the \bibfield{author} field of the source to both the \bibfield{author} and the \bibfield{bookauthor} fields of the target in order to allow for compact \bibfield{inbook}\slash \bibfield{bookinbook} entries. The source may be either a \bibtype{mvbook} or a \bibtype{book} entry, the target either an \bibtype{inbook} or a \bibtype{bookinbook} entry.

\begin{ltxexample}
\DeclareDataInheritance{<<*>>}{<<inbook,incollection>>}{
  \noinherit{<<introduction>>}
}
\end{ltxexample}
%
This rule prevents inheritance of the \bibfield{introduction} field. It applies to all targets of type
\bibtype{inbook} or \bibtype{incollection}, regardless of the source entry type.

\begin{ltxexample}
\DeclareDataInheritance{<<*>>}{<<*>>}{
  \noinherit{<<abstract>>}
}
\end{ltxexample}
%
This rule, which applies to all entries, regardless of the source and target entry types, prevents inheritance of the \bibfield{abstract} field.

\begin{ltxexample}
\DefaultInheritance{all=true,override=false}
\ResetDataInheritance
\end{ltxexample}
%
This example demonstrates how to emulate traditional \bibtex's cross"=referencing mechanism. It enables inheritance by default, disables overwriting, and clears all other inheritance rules and mappings.

In a bibliography entry, you can give an option <noinherit> where the value
is a datafield set defined with \cmd{DeclareDatafieldSet}
(\secref{aut:ctm:dsets}). This will block inheritance of the fields in the
set on a per-entry basis. For example:

\begin{ltxexample}
\DeclareDatafieldSet{nobtitle}{
  \member[field=booktitle]
}
\end{ltxexample}

\begin{lstlisting}[style=bibtex]{}
@INBOOK{s1,
  OPTIONS  = {noinherit=nobtitle},
  TITLE    = {Subtitle},
  CROSSREF = {s2}
}

@BOOK{s2,
  TITLE = {Title}
}
\end{lstlisting}
%
Here, \bibfield{s1} will not inherit the \bibfield{TITLE} of \bibfield{s2}
as \bibfield{BOOKTITLE} as this is blocked by the datafield set given as
the value to the \opt{noinherit} option.
%
One important thing to note is that children will never inherit any dateparts of a given type if they already contain a datepart of that type. So, for example:

\begin{lstlisting}[style=bibtex]{}
@INBOOK{b1,
  DATE     = {2004-03-03},
  ORIGDATE = {2004-03},
  CROSSREF = {b2}
}

@BOOK{b2,
  DATE      = {2004-03-03/2005-08-09},
  ORIGDATE  = {2004-03/2005-08},
  EVENTDATE = {2004-03/2005-08},
}
\end{lstlisting}
%
Here, \bibfield{b1} will not inherit any of \bibfield{endyear}, \bibfield{endmonth}, \bibfield{endday}, \bibfield{origendyear} or \bibfield{origendmonth} as this would make a mess of its own dates. It will, given the inheritance defaults, inherit all of the \bibfield{event*} date parts.

\subsection{Auxiliary Commands}
\label{aut:aux}

The facilities in this section are intended for analyzing and saving bibliographic data rather than formatting and printing it.

\subsubsection{Data Commands}
\label{aut:aux:dat}

The commands in this section grant low"=level access to the unformatted bibliographic data. They are not intended for typesetting but rather for things like saving data to a temporary macro so that it may be used in a comparison later.

\begin{ltxsyntax}

\cmditem{thefield}{field}

Expands to the unformatted \prm{field}. If the \prm{field} is undefined, this command expands to an empty string.

\cmditem{strfield}{field}

Similar to \cmd{thefield}, except that the field is automatically sanitized such that its value may safely be used in the formation of a control sequence name.

\cmditem{csfield}{field}

Similar to \cmd{thefield}, but prevents expansion.

\cmditem{usefield}{command}{field}

Executes \prm{command} using the unformatted \prm{field} as its argument.

\cmditem{thelist}{literal list}

Expands to the unformatted \prm{literal list}. If the list is undefined, this command expands to an empty string. Note that this command will dump the \prm{literal list} in the internal format used by this package. This format is not suitable for printing.

\cmditem{strlist}{literal list}

Similar to \cmd{thelist}, except that the list internal representation is automatically sanitized such that its value may safely be used in the formation of a control sequence name.

\cmditem{thefirstlistitem}{literal list}

Expands to the unformatted first item in \prm{literal list}. If the \prm{literal list} is undefined, this command expands to an empty string.

\cmditem{strfirstlistitem}{literal list}

Similar to \cmd{thefirstlistitem}, except that the item is automatically sanitized such that its value may safely be used in the formation of a control sequence name.

\cmditem{usefirstlistitem}{command}{literal list}

Executes \prm{command} using the unformatted first item of \prm{literal list} as its argument.

\cmditem{thename}{name list}

Expands to the unformatted \prm{name list}. If the list is undefined, this command expands to an empty string. Note that this command will dump the \prm{name list} in the internal format used by this package. This format is not suitable for printing.

\cmditem{strname}{name list}

Similar to \cmd{thename}, except that the name internal representation is automatically sanitized such that its value may safely be used in the formation of a control sequence name.

\cmditem{savefield}{field}{macro}
\cmditem*{savefield*}{field}{macro}

Copies an unformatted \prm{field} to a \prm{macro}. The regular variant of this command defines the \prm{macro} globally, the starred one works locally.

\cmditem{savelist}{literal list}{macro}
\cmditem*{savelist*}{literal list}{macro}

Copies an unformatted \prm{literal list} to a \prm{macro}. The regular variant of this command defines the \prm{macro} globally, the starred one works locally.

\cmditem{savename}{name list}{macro}
\cmditem*{savename*}{name list}{macro}

Copies an unformatted \prm{name list} to a \prm{macro}. The regular variant of this command defines the \prm{macro} globally, the starred one works locally.

\cmditem{savefieldcs}{field}{csname}
\cmditem*{savefieldcs*}{field}{csname}

Similar to \cmd{savefield}, but takes the control sequence name \prm{csname} (without a leading backslash) as an argument, rather than a macro name.

\cmditem{savelistcs}{literal list}{csname}
\cmditem*{savelistcs*}{literal list}{csname}

Similar to \cmd{savelist}, but takes the control sequence name \prm{csname} (without a leading backslash) as an argument, rather than a macro name.

\cmditem{savenamecs}{name list}{csname}
\cmditem*{savenamecs*}{name list}{csname}

Similar to \cmd{savename}, but takes the control sequence name \prm{csname} (without a leading backslash) as an argument, rather than a macro name.

\cmditem{restorefield}{field}{macro}

Restores a \prm{field} from a \prm{macro} defined with \cmd{savefield} before. The field is restored within a local scope.

\cmditem{restorelist}{literal list}{macro}

Restores a \prm{literal list} from a \prm{macro} defined with \cmd{savelist} before. The list is restored within a local scope.

\cmditem{restorename}{name list}{macro}

Restores a \prm{name list} from a \prm{macro} defined with \cmd{savename} before. The list is restored within a local scope.

\cmditem{clearfield}{field}

Clears the \prm{field} within a local scope. A field cleared this way is treated as undefined by subsequent data commands.

\cmditem{clearlist}{literal list}

Clears the \prm{literal list} within a local scope. A list cleared this way is treated as undefined by subsequent data commands.

\cmditem{clearname}{name list}

Clears the \prm{name list} within a local scope. A list cleared this way is treated as undefined by subsequent data commands.

\end{ltxsyntax}

\subsubsection{Stand-alone Tests}
\label{aut:aux:tst}

The commands in this section are various kinds of stand"=alone tests for use in bibliography and citation styles.

\begin{ltxsyntax}

\cmditem{if$<$datetype$>$julian}{true}{false}

Expands to \prm{true} if the date <datetype>date (\opt{date}, \opt{urldate}, \opt{eventdate} etc.) Was converted to the Julian Calendar due to the settings of the \opt{julian}and \opt{gregorianstart}  options.

\cmditem{ifdatejulian}{true}{false}

As \cmd{if$<$datetype$>$julian} but for use in \cmd{mkbibdate*} formatting commands (\secref{aut:fmt:lng}) inside which the appropriate \cmd{if$<$datetype$>$julian} command is aliased to this command.

\cmditem{if$<$datetype$>$dateera}{era}{true}{false}

Expands to \prm{true} if the date <datetype>date (\opt{date}, \opt{urldate}, \opt{eventdate} etc.) has an era specification equal to \prm{era} and \prm{false} otherwise.  The supported \prm{era} strings which \biber determines and passes in the \file{.bbl} are:

\begin{description}
\item[bce]~BCE/BC era
\item[ce]~CE/AD era
\end{description}

This command is useful for determining whether to print the location
strings in \secref{aut:lng:key:dt}.

\cmditem{ifdateera}{era}{true}{false}

As \cmd{if$<$datetype$>$dateera} but for use in \cmd{mkbibdate*} formatting commands (\secref{aut:fmt:lng}) inside which the appropriate \cmd{if$<$datetype$>$dateera} command is aliased to this command.

\cmditem{if$<$datetype$>$datecirca}{true}{false}

Expands to \prm{true} if the date <datetype>date (\opt{date}, \opt{urldate}, \opt{eventdate} etc.) had a <circa> marker in the source and \prm{false} otherwise.  See \secref{bib:use:dat}. This command is useful for determining whether to print the location strings in \secref{aut:lng:key:dt}.

\cmditem{ifdatecirca}{true}{false}

As \cmd{if$<$datetype$>$datecirca} but for use in \cmd{mkbibdate*} formatting commands (\secref{aut:fmt:lng}) inside which the appropriate \cmd{if$<$datetype$>$datecirca} command is aliased to this command.

\cmditem{if$<$datetype$>$dateuncertain}{true}{false}

Expands to \prm{true} if the date <datetype>date (\opt{date}, \opt{urldate}, \opt{eventdate} etc.) had an uncertainty marker in the source and \prm{false} otherwise.  See \secref{bib:use:dat}. This command is useful for determining whether to print, for example, a question mark after a year.

\cmditem{ifdateuncertain}{true}{false}

As \cmd{if$<$datetype$>$dateuncertain} but for use in \cmd{mkbibdate*} formatting commands (\secref{aut:fmt:lng}) inside which the appropriate \cmd{if$<$datetype$>$dateuncertain} command is aliased to this command.

\cmditem{ifenddateuncertain}{true}{false}

As \cmd{ifend$<$datetype$>$dateuncertain} but for use in \cmd{mkbibdate*} formatting commands (\secref{aut:fmt:lng}) inside which the appropriate \cmd{ifend$<$datetype$>$dateuncertain} command is aliased to this command.

\cmditem{if$<$datetype$>$dateunknown}{true}{false}

Expands to \prm{true} if the date <datetype>date (\opt{date}, \opt{urldate}, \opt{eventdate} etc.) is marked as unknown (as opposed to open) in the source and \prm{false} otherwise.  See \secref{bib:use:dat}.

\cmditem{ifdateunknown}{true}{false}

As \cmd{if$<$datetype$>$dateunknown} but for use in \cmd{mkbibdate*} formatting commands (\secref{aut:fmt:lng}) inside which the appropriate \cmd{if$<$datetype$>$dateunknown} command is aliased to this command.

\cmditem{ifenddateunknown}{true}{false}

As \cmd{ifend$<$datetype$>$dateunknown} but for use in \cmd{mkbibdate*} formatting commands (\secref{aut:fmt:lng}) inside which the appropriate \cmd{ifend$<$datetype$>$dateunknown} command is aliased to this command.

\cmditem{iflabeldateisdate}{true}{false}

Expands to \prm{true} if labeldate is defined and was obtained from date, and to \prm{false} otherwise.

\cmditem{ifdatehasyearonlyprecision}{datetype}{true}{false}

Expands to \prm{true} if the \prm{datetype}date is defined and would be shown with year precision \cmd{print$<$datetype$>$date}, and to false otherwise.

\cmditem{ifdatehastime}{datetype}{true}{false}

Expands to \prm{true} if the \prm{datetype}date is defined, has a time component and \opt{$<$datetype$>$dateusetime} is true, and to false otherwise.

\cmditem{ifdateshavedifferentprecision}{datetype1}{datetype2}{true}{false}

Expands to \prm{true} if the two dates \prm{datetype1} and \prm{datetype2} would show in different precision when printed with \cmd{print$<$datetype1$>$date} and \cmd{print$<$datetype2$>$date} respectively, and to \prm{false} otherwise.

\cmditem{ifdateyearsequal}{datetype1}{datetype2}{true}{false}

Expands to \prm{true} if the two dates \prm{datetype1} and \prm{datetype2} have the same year and era. Since the sign of the date is saved in the era field, years should be compared using this command to avoid confusion when the two years have opposite signs

\cmditem{ifdatesequal}{datetype1}{datetype2}{true}{false}

Expands to \prm{true} if the two dates \prm{datetype1} and \prm{datetype2} are the same. Here \prm{datetype2} may be the <end> bit of \prm{datetype1} (or vice versa).

\cmditem{ifdaterangesequal}{datetype1}{datetype2}{true}{false}

Expands to \prm{true} if the two date ranges---that is the start and the end date---\prm{datetype1} and \prm{datetype2} are the same.


\cmditem{ifcaselang}[language]{true}{false}

Expands to \prm{true} if the optional \prm{language} is one of those
declared by \cmd{DeclareCaseLangs} (see \secref{aut:aux:msc}) and to
\prm{false} otherwise. Without the optional argument, checks the current
value of \cmd{currentlang}.

\cmditem{ifsortingnamekeytemplatename}{string}{true}{false}

Expands to \prm{true} if the \prm{string} is equal to the current in scope sorting name key template name (see \ref{aut:ctm:srt}), and to \prm{false} otherwise.

\cmditem{ifuniquenametemplatename}{string}{true}{false}

Expands to \prm{true} if the \prm{string} is equal to the current in scope uniqueness name key template name (see \ref{aut:ctm:srt}), and to \prm{false} otherwise.

\cmditem{iflabelalphanametemplatename}{string}{true}{false}

Expands to \prm{true} if the \prm{string} is equal to the current in scope alphabetic label name template name (see \ref{aut:ctm:srt}), and to \prm{false} otherwise.

\cmditem{iffieldundef}{field}{true}{false}

Expands to \prm{true} if the \prm{field} is undefined, and to \prm{false} otherwise.

\cmditem{iflistundef}{literal list}{true}{false}

Expands to \prm{true} if the \prm{literal list} is undefined, and to \prm{false} otherwise.

\cmditem{ifnameundef}{name list}{true}{false}

Expands to \prm{true} if the \prm{name list} is undefined, and to \prm{false} otherwise.

\cmditem{iffieldsequal}{field 1}{field 2}{true}{false}

Expands to \prm{true} if the values of \prm{field 1} and \prm{field 2} are equal, and to \prm{false} otherwise.

\cmditem{iflistsequal}{literal list 1}{literal list 2}{true}{false}

Expands to \prm{true} if the values of \prm{literal list 1} and \prm{literal list 2} are equal, and to \prm{false} otherwise.

\cmditem{ifnamesequal}{name list 1}{name list 2}{true}{false}

Expands to \prm{true} if the values of \prm{name list 1} and \prm{name list 2} are equal, and to \prm{false} otherwise.

\cmditem{iffieldequals}{field}{macro}{true}{false}

Expands to \prm{true} if the value of the \prm{field} is equal to the definition of \prm{macro}, and to \prm{false} otherwise.

\cmditem{iflistequals}{literal list}{macro}{true}{false}

Expands to \prm{true} if the value of the \prm{literal list} is equal to the definition of \prm{macro}, and to \prm{false} otherwise.

\cmditem{ifnameequals}{name list}{macro}{true}{false}

Expands to \prm{true} if the value of the \prm{name list} is equal to the definition of \prm{macro}, and to \prm{false} otherwise.

\cmditem{iffieldequalcs}{field}{csname}{true}{false}

Similar to \cmd{iffieldequals} but takes the control sequence name \prm{csname} (without a leading backslash) as an argument, rather than a macro name.

\cmditem{iflistequalcs}{literal list}{csname}{true}{false}

Similar to \cmd{iflistequals} but takes the control sequence name \prm{csname} (without a leading backslash) as an argument, rather than a macro name.

\cmditem{ifnameequalcs}{name list}{csname}{true}{false}

Similar to \cmd{ifnameequals} but takes the control sequence name \prm{csname} (without a leading backslash) as an argument, rather than a macro name.

\cmditem{iffieldequalstr}{field}{string}{true}{false}

Executes \prm{true} if the value of the \prm{field} is equal to \prm{string}, and \prm{false} otherwise. This command is robust.

\cmditem{iffieldxref}{field}{true}{false}

If the \bibfield{crossref}\slash \bibfield{xref} field of an entry is defined, this command checks if the \prm{field} is related to the cross"=referenced parent entry. It executes \prm{true} if the \prm{field} of the child entry is equal to the corresponding \prm{field} of the parent entry, and \prm{false} otherwise. If the \bibfield{crossref}\slash \bibfield{xref} field is undefined, it always executes \prm{false}. This command is robust. See the description of the \bibfield{crossref} and \bibfield{xref} fields in \secref{bib:fld:spc} as well as \secref{bib:cav:ref} for further information concerning cross"=referencing.

\cmditem{iflistxref}{literal list}{true}{false}

Similar to \cmd{iffieldxref} but checks if a \prm{literal list} is related to the cross"=referenced parent entry. See the description of the \bibfield{crossref} and \bibfield{xref} fields in \secref{bib:fld:spc} as well as \secref{bib:cav:ref} for further information concerning cross"=referencing.

\cmditem{ifnamexref}{name list}{true}{false}

Similar to \cmd{iffieldxref} but checks if a \prm{name list} is related to the cross"=referenced parent entry. See the description of the \bibfield{crossref} and \bibfield{xref} fields in \secref{bib:fld:spc} as well as \secref{bib:cav:ref} for further information concerning cross"=referencing.

\cmditem{ifcurrentfield}{field}{true}{false}

Executes \prm{true} if the current field is \prm{field}, and \prm{false} otherwise. This command is robust. It is intended for use in field formatting directives and always executes \prm{false} when used in any other context.

\cmditem{ifcurrentlist}{literal list}{true}{false}

Executes \prm{true} if the current list is \prm{literal list}, and \prm{false} otherwise. This command is robust. It is intended for use in list formatting directives and always executes \prm{false} when used in any other context.

\cmditem{ifcurrentname}{name list}{true}{false}

Executes \prm{true} if the current list is \prm{name list}, and \prm{false} otherwise. This command is robust. It is intended for use in list formatting directives and always executes \prm{false} when used in any other context.

\cmditem{ifuseprefix}{true}{false}

Expands to \prm{true} if the \opt{useprefix} option is enabled (either globally or for the current entry), and \prm{false} otherwise. See \secref{use:opt:bib} for details on this option.

\cmditem{ifuseauthor}{true}{false}

This is just a particular case of the \cmd{ifuse$<$name$>$} macro below but is mentioned here as \bibfield{author} is part of the default data model. Expands to \prm{true} if the \opt{useauthor} option is enabled (either globally or for the current entry), and \prm{false} otherwise. See \secref{use:opt:bib} for details on this option.

\cmditem{ifuseeditor}{true}{false}

This is just a particular case of the \cmd{ifuse$<$name$>$} macro below but is mentioned here as \bibfield{editor} is part of the default data model. Expands to \prm{true} if the \opt{useeditor} option is enabled (either globally or for the current entry), and \prm{false} otherwise. See \secref{use:opt:bib} for details on this option.

\cmditem{ifusetranslator}{true}{false}

This is just a particular case of the \cmd{ifuse$<$name$>$} macro below but is mentioned here as \bibfield{translator} is part of the default data model. Expands to \prm{true} if the \opt{usetranslator} option is enabled (either globally or for the current entry), and \prm{false} otherwise. See \secref{use:opt:bib} for details on this option.

\cmditem{ifuse$<$name$>$}{true}{false}

Expands to \prm{true} if the \opt{use$<$name$>$} option is enabled (either globally or for the current entry), and \prm{false} otherwise. See \secref{use:opt:bib} for details on this option.

\cmditem{ifcrossrefsource}{true}{false}

Expands to \prm{true} if the entry was inclued in the \file{.bbl} due to being referenced more than \opt{mincrossrefs} times and false otherwise. See \secref{use:opt:pre:gen}. Also expands to false if the entry was directly cited.

\cmditem{ifxrefsource}{true}{false}

Expands to \prm{true} if the entry was inclued in the \file{.bbl} due to being referenced more than \opt{minxrefs} times and false otherwise. See \secref{use:opt:pre:gen}. Also expands to false if the entry was directly cited.

\cmditem{ifsingletitle}{true}{false}

Expands to \prm{true} if there is only one work by the \opt{labelname} name in the bibliography, and to \prm{false} otherwise. If \opt{labelname} is not set for an entry, this will always expand to \prm{false}. Note that this feature needs to be enabled explicitly with the package option \opt{singletitle}.

\cmditem{ifnocite}{true}{false}

Expands to \prm{true} if the entry was \emph{only} included in the \file{.bbl} via \cmd{nocite}. That is, returns \prm{false} if an entry was both \cmd{nocite}'d and \cmd{cite}'d.

\cmditem{ifuniquetitle}{true}{false}

Expands to \prm{true} if there is only one work with the title \opt{labeltitle} and to \prm{false} otherwise. If \opt{labeltitle} is not set for an entry, this will always expand to \prm{false}. Note that this feature needs to be enabled explicitly with the package option \opt{uniquetitle}.

\cmditem{ifuniquebaretitle}{true}{false}

Expands to \prm{true} if \bibfield{labelname} is empty and there is only one work with the title \opt{labeltitle} and to \prm{false} otherwise. If \opt{labeltitle} is not set for an entry, this will always expand to \prm{false}. Note that this feature needs to be enabled explicitly with the package option \opt{uniquebaretitle}.

\cmditem{ifuniquework}{true}{false}

Expands to \prm{true} if there is only one work by the \opt{labelname} name with the \opt{labeltitle} title in the bibliography, and to \prm{false} otherwise. If neither \opt{labelname} nor \opt{labeltitle} are set for an entry, this will always expand to \prm{false}. Note that this feature needs to be enabled explicitly with the package option \opt{uniquework}. If both \bibfield{singletitle} and \bibfield{uniquetitle} are false for the same entry, this could be because another entry has the same \bibfield{labdlname} and yet another, different, entry has the same \bibfield{labeltitle}. \bibfield{uniquework} would let you know that there is another entry that has \emph{both} the same \bibfield{labelname} \emph{and} the same \bibfield{labeltitle}. This could be helpful in cases where multiple people maintain bibliography datasources and there is a risk of adding the same work with different keys without other parties realising this. This test could help to find such duplicates.

\cmditem{ifuniqueprimaryauthor}{true}{false}

Expands to \prm{true} if the primary (first) author name of \opt{labelname} is unique in the bibliography list and to \prm{false} otherwise. This effectively answers the question <is there more than one author with the same base name>. The base name parts are defined by \cmd{DeclareUniquenameTemplate} see \secref{aut:cav:amb}. This is required by some styles (e.g. APA) which mandates primary author disambiguation only and only if there are (different) primary authors with the same family name. If \opt{labelname} is not set for an entry, this will always expand to \prm{false}. Note that this feature needs to be enabled explicitly with the package option \opt{uniqueprimaryauthor}.

\cmditem{ifandothers}{list}{true}{false}

Expands to \prm{true} if the \prm{list} is defined and has been truncated in the \file{bib} file with the keyword <\texttt{and others}>, and to \prm{false} otherwise. The \prm{list} may be a literal list or a name list.

\cmditem{ifmorenames}{true}{false}

Expands to \prm{true} if the current name list has been or will be truncated, and to \prm{false} otherwise. This command is intended for use in formatting directives for name lists. It will always expand to \prm{false} when used elsewhere. This command performs the equivalent of an \cmd{ifandothers} test for the current list. If this test is negative, it also checks if the \cnt{listtotal} counter is larger than \cnt{liststop}. This command may be used in a formatting directive to decide if a note such as «and others» or «et al.» is to be printed at the end of the list. Note that you still need to check whether you are in the middle or at the end of the list, \ie whether \cnt{listcount} is smaller than or equal to \cnt{liststop}, see \secref{aut:bib:dat} for details.

\cmditem{ifmoreitems}{true}{false}

This command is similar to \cmd{ifmorenames} but checks the current literal list. It is intended for use in formatting directives for literal lists. It will always expand to \prm{false} when used elsewhere.

\cmditem{if$<$namepart$>$inits}{true}{false}

Expands to \prm{true} or \prm{false}, depending on the state of the \opt{$<$namepart$>$inits} package option (see \secref{use:opt:pre:int}). This command is intended for use in formatting directives for name lists.

\cmditem{ifterseinits}{true}{false}

Expands to \prm{true} or \prm{false}, depending on the state of the \opt{terseinits} package option (see \secref{use:opt:pre:int}). This command is intended for use in formatting directives for name lists.

\cmditem{ifentrytype}{type}{true}{false}

Executes \prm{true} if the entry type of the entry currently being processed is \prm{type}, and \prm{false} otherwise.

\cmditem{ifkeyword}{keyword}{true}{false}

Executes \prm{true} if the \prm{keyword} is found in the \bibfield{keywords} field of the entry currently being processed, and \prm{false} otherwise.

\cmditem{ifentrykeyword}{entrykey}{keyword}{true}{false}

A variant of \cmd{ifkeyword} which takes an entry key as its first argument. This is useful for testing an entry other than the one currently processed. A user-facing version of this command is available for use in documents see \secref{use:eq}.

\cmditem{ifcategory}{category}{true}{false}

Executes \prm{true} if the entry currently being processed has been assigned to a \prm{category} with \cmd{addtocategory}, and \prm{false} otherwise.

\cmditem{ifentrycategory}{entrykey}{category}{true}{false}

A variant of \cmd{ifcategory} which takes an entry key as its first argument. This is useful for testing an entry other than the one currently processed. A user-facing version of this command is available for use in documents see \secref{use:eq}

\cmditem{ifciteseen}{true}{false}

Executes \prm{true} if the entry currently being processed has been cited before, and \prm{false} otherwise. This command is robust and intended for use in citation styles. If there are any \env{refsection} environments in the document, the citation tracking is local to these environments. Note that the citation tracker needs to be enabled explicitly with the package option \opt{citetracker}. The behavior of this test depends on the mode the citation tracker is operating in, see \secref{use:opt:pre:int} for details. If the citation tracker is disabled, the test always yields \prm{false}. Also see the \cmd{citetrackertrue} and \cmd{citetrackerfalse} switches in \secref{aut:aux:msc}.

\cmditem{ifentryseen}{entrykey}{true}{false}

A variant of \cmd{ifciteseen} which takes an entry key as its first argument. Since the \prm{entrykey} is expanded prior to performing the test, it is possible to test for entry keys in a field such as \bibfield{xref}:

\begin{ltxexample}
\ifentryseen{<<\thefield{xref}>>}{true}{false}
\end{ltxexample}
%
Apart from the additional argument, \cmd{ifentryseen} behaves like \cmd{ifciteseen}. A user-facing version of this command is available for use in documents see \secref{use:eq}.

\cmditem{ifentryinbib}{entrykey}{true}{false}

Executes \prm{true} if the entry \prm{entrykey} appears in the current bibliography, and \prm{false} otherwise. A user-facing version of this command is available for use in documents see \secref{use:eq}.

\cmditem{iffirstcitekey}{true}{false}

Executes \prm{true} if the entry currently being processed is the first one in the citation list, and \prm{false} otherwise. This command relies on the \cnt{citecount}, \cnt{citetotal}, \cnt{multicitecount} and \cnt{multicitetotal} counters (\secref{aut:fmt:ilc}) and thus is intended for use only in the \prm{loopcode} of a citation command defined with \cmd{DeclareCiteCommand}.

\cmditem{iflastcitekey}{true}{false}

Similar \cmd{iffirstcitekey}, but executes \prm{true} if the entry currently being processed is the last one in the citation list, and \prm{false} otherwise.

\cmditem{ifciteibid}{true}{false}

Expands to \prm{true} if the entry currently being processed is the same as the last one, and to \prm{false} otherwise. This command is intended for use in citation styles. If there are any \env{refsection} environments in the document, the tracking is local to these environments. Note that the <ibidem> tracker needs to be enabled explicitly with the package option \opt{ibidtracker}. The behavior of this test depends on the mode the tracker is operating in, see \secref{use:opt:pre:int} for details. If the tracker is disabled, the test always yields \prm{false}. Also see the \cmd{citetrackertrue} and \cmd{citetrackerfalse} switches in \secref{aut:aux:msc}.

\cmditem{ifciteidem}{true}{false}

Expands to \prm{true} if the primary name (\ie the author or editor) in the entry currently being processed is the same as the last one, and to \prm{false} otherwise. This command is intended for use in citation styles. If there are any \env{refsection} environments in the document, the tracking is local to these environments. Note that the <idem> tracker needs to be enabled explicitly with the package option \opt{idemtracker}. The behavior of this test depends on the mode the tracker is operating in, see \secref{use:opt:pre:int} for details. If the tracker is disabled, the test always yields \prm{false}. Also see \cmd{citetrackertrue} and \cmd{citetrackerfalse} in \secref{aut:aux:msc}.

\cmditem{ifopcit}{true}{false}

This command is similar to \cmd{ifciteibid} except that it expands to \prm{true} if the entry currently being processed is the same as the last one \emph{by this author or editor}. Note that the <opcit> tracker needs to be enabled explicitly with the package option \opt{opcittracker}. The behavior of this test depends on the mode the tracker is operating in, see \secref{use:opt:pre:int} for details. If the tracker is disabled, the test always yields \prm{false}. Also see the \cmd{citetrackertrue} and \cmd{citetrackerfalse} switches in \secref{aut:aux:msc}.

\cmditem{ifloccit}{true}{false}

This command is similar to \cmd{ifopcit} except that it also compares the \prm{postnote} arguments and expands to \prm{true} only if they match and are numerical (in the sense of \cmd{ifnumerals} from \secref{aut:aux:tst}), \ie \cmd{ifloccit} will yield \texttt{true} if the citation refers to the same page cited before. Note that the <loccit> tracker needs to be enabled explicitly with the package option \opt{loccittracker}. The behavior of this test depends on the mode the tracker is operating in, see \secref{use:opt:pre:int} for details. If the tracker is disabled, the test always yields \prm{false}. Also see the \cmd{citetrackertrue} and \cmd{citetrackerfalse} switches in \secref{aut:aux:msc}.

\cmditem{iffirstonpage}{true}{false}

The behavior of this command is responsive to the package option \opt{pagetracker}. If the option is set to \texttt{page}, it expands to \prm{true} if the current item is the first one on the page, and to \prm{false} otherwise. If the option is set to \texttt{spread}, it expands to \prm{true} if the current item is the first one on the double-page spread, and to \prm{false} otherwise. If the page tracker is disabled, this test always yields \prm{false}. Depending on the context, the <item> may be a citation or an entry in the bibliography or a bibliography list. Note that this test distinguishes between body text and footnotes. For example, if used in the first footnote on a page, it will expand to \prm{true} even if there is a citation in the body text prior to the footnote. Also see the \cmd{pagetrackertrue} and \cmd{pagetrackerfalse} switches in \secref{aut:aux:msc}.

\cmditem{ifsamepage}{instance 1}{instance 2}{true}{false}

This command expands to \prm{true} if two instances of a reference are located on the same page or double-page spread, and to \prm{false} otherwise. An instance of a reference may be a citation or an entry in the bibliography or a bibliography list. These instances are identified by the value of the \cnt{instcount} counter, see \secref{aut:fmt:ilc}. The behavior of this command is responsive to the package option \opt{pagetracker}. If this option is set to \texttt{spread}, \cmd{ifsamepage} is in fact an <if same spread> test. If the page tracker is disabled, this test always yields \prm{false}. The arguments \prm{instance 1} and \prm{instance 2} are treated as integer expressions in the sense of \etex's \cmd{numexpr}. This implies that it is possible to make calculations within these arguments, for example:

\begin{ltxexample}
\ifsamepage{<<\value>>{instcount}}{<<\value>>{instcount}<<-1>>}{true}{false}
\end{ltxexample}

Note that \cmd{value} is not prefixed by \cmd{the} and that the subtraction is included in the second argument in the above example. If \prm{instance 1} or \prm{instance 2} is an invalid number (for example, a negative one), the test yields \prm{false}. Also note that this test does not distinguish between body text and footnotes. Also see the \cmd{pagetrackertrue} and \cmd{pagetrackerfalse} switches in \secref{aut:aux:msc}.

\cmditem{ifinteger}{string}{true}{false}

Executes \prm{true} if the \prm{string} is a positive integer, and \prm{false} otherwise. This command is robust.

\cmditem{ifnumeral}{string}{true}{false}

Executes \prm{true} if the \prm{string} is an Arabic or Roman numeral, and \prm{false} otherwise. This command is robust. See also \cmd{DeclareNumChars} and \cmd{NumCheckSetup} in \secref{aut:aux:msc}.

\cmditem{ifnumerals}{string}{true}{false}

Executes \prm{true} if the \prm{string} is a range or a list of Arabic or Roman numerals, and \prm{false} otherwise. This command is robust. In contrast to \cmd{ifnumeral}, it will also execute \prm{true} with arguments like «52--58», «14/15», «1,~3,~5», and so on. See also \cmd{DeclareNumChars}, \cmd{DeclareRangeChars}, \cmd{DeclareRangeCommands}, \cmd{NumCheckSetup}, and \cmd{NumsCheckSetup} in \secref{aut:aux:msc}.

\cmditem{ifpages}{string}{true}{false}

Similar to \cmd{ifnumerals}, but also considers \cmd{DeclarePageCommands} and \cmd{PagesCheckSetup} from \secref{aut:aux:msc}.

\cmditem{iffieldint}{field}{true}{false}

Similar to \cmd{ifinteger}, but uses the value of a \prm{field} rather than a literal string in the test. If the \prm{field} is undefined, it executes \prm{false}.

\cmditem{iffieldnum}{field}{true}{false}

Similar to \cmd{ifnumeral}, but uses the value of a \prm{field} rather than a literal string in the test. If the \prm{field} is undefined, it executes \prm{false}.

\cmditem{iffieldnums}{field}{true}{false}

Similar to \cmd{ifnumerals}, but uses the value of a \prm{field} rather than a literal string in the test. If the \prm{field} is undefined, it executes \prm{false}.

\cmditem{iffieldpages}{field}{true}{false}

Similar to \cmd{ifpages}, but uses the value of a \prm{field} rather than a literal string in the test. If the \prm{field} is undefined, it executes \prm{false}.

\cmditem{ifbibstring}{string}{true}{false}

Expands to \prm{true} if the \prm{string} is a known localisation key, and to \prm{false} otherwise. The localisation keys defined by default are listed in \secref{aut:lng:key}. New ones may be defined with \cmd{NewBibliographyString}.

\cmditem{ifbibxstring}{string}{true}{false}

Similar to \cmd{ifbibstring}, but the \prm{string} is expanded.

\cmditem{iffieldbibstring}{field}{true}{false}

Similar to \cmd{ifbibstring}, but uses the value of a \prm{field} rather than a literal string in the test. If the \prm{field} is undefined, it expands to \prm{false}.

\cmditem{iffieldplusstringbibstring}{field}{string}{true}{false}

Similar to \cmd{iffieldbibstring}, but appends \prm{string} to the value of \prm{field} and checks if the resulting string is a known localisation key. Expands to \prm{false} if \prm{field} is undefined.

\cmditem{ifdriver}{entrytype}{true}{false}

Expands to \prm{true} if a driver for the \prm{entrytype} is available, and to \prm{false} otherwise.

\cmditem{ifcapital}{true}{false}

Executes \prm{true} if \biblatex's punctuation tracker would capitalize a localisation string at the current location, and \prm{false} otherwise. This command is robust. It may be useful for conditional capitalization of certain parts of a name in a formatting directive.

\cmditem{ifcitation}{true}{false}

Expands to \prm{true} when located in a citation, and to \prm{false} otherwise. Note that this command is responsive to the outermost context in which it is used. For example, if a citation command defined with \cmd{DeclareCiteCommand} executes a driver defined with \cmd{DeclareBibliographyDriver}, any \cmd{ifcitation} tests in the driver code will yield \prm{true}. See \secref{aut:cav:mif} for a practical example.

\cmditem{ifvolcite}{true}{false}

Expands to \prm{true} when located in \cmd{volcite} or a related citation command (\secref{use:cit:spc}), and to \prm{false} otherwise.

\cmditem{ifbibliography}{true}{false}

Expands to \prm{true} when located in a bibliography, and to \prm{false} otherwise. Note that this command is responsive to the outermost context in which it is used. For example, if a driver defined with \cmd{DeclareBibliographyDriver} executes a citation command defined with \cmd{DeclareCiteCommand}, any \cmd{ifbibliography} tests in the citation code will yield \prm{true}. See \secref{aut:cav:mif} for a practical example.

\cmditem{ifnatbibmode}{true}{false}

Expands to \prm{true} or \prm{false} depending on the \opt{natbib} option from \secref{use:opt:ldt}.

\cmditem{ifciteindex}{true}{false}

Expands to \prm{true} or \prm{false} depending on the \opt{indexing} option from \secref{use:opt:pre:gen}.

\cmditem{ifbibindex}{true}{false}

Expands to \prm{true} or \prm{false} depending on the \opt{indexing} option from \secref{use:opt:pre:gen}.

\cmditem{iffootnote}{true}{false}

Expands to \prm{true} when located in a footnote, and to \prm{false} otherwise. Note that footnotes in \env{minipage} environments are considered to be part of the body text. This command will only expand to \prm{true} in footnotes a the bottom of the page and in endnotes as provided by the \sty{endnotes} package.

\cntitem{citecounter}

This counter indicates how many times the entry currently being processed is cited in the current reference section. Note that this feature needs to be enabled explicitly with the package option \opt{citecounter}. If the option is set to \texttt{context}, citations in the body text and in footnotes are counted separately. In this case, \cnt{citecounter} will hold the value of the context it is used in.

\cntitem{maxcitecounter}

This counter holds the maximum value of \cnt{citecounter} across all entries in the current reference section. Like \cnt{citecounter} it is only available if the \opt{citecounter} option is enabled and tracks footnotes and text separately if the option is set to \texttt{context}.

\cntitem{uniquename}
This counter refers to the \bibfield{labelname} list. It is set on a per-name basis. Its value is \texttt{0} if the base parts of the name (by default just the <family> part of the name) are unique, \texttt{1} if adding the other non-base parts of the name (as specified in the uniquename template defined by \cmd{DeclareUniquenameTemplate}) as initials will make it unique, and \texttt{2} if adding the full form of the non-base parts of the name are required to disambiguate the name. This information is required by author-year and author-title citation schemes which add additional parts of the name when citing different authors with the same family name. For example, (given the default \cmd{DeclareUniquenameTemplate} definition) if there is one <John Doe> and one <Edward Doe> in the list of references, this counter will be set to \texttt{1}. If there is one <John Doe> and one <Jane Doe>, the value of the counter will be \texttt{2}. If the option is set to \texttt{init}\slash \texttt{allinit}\slash \texttt{mininit}, the counter will be limited to \texttt{1}. This is useful for citations styles which use initials to disambiguate names but never print the full name in citations. If adding the initials is not sufficient to disambiguate the name, \cnt{uniquename} will also be set to \texttt{0} for that name. This feature needs to be enabled explicitly with the package option \opt{uniquename}. Note that the \cnt{uniquename} counter is local to \cmd{printnames} and that it is only set for the \bibfield{labelname} list and for the name list that \bibfield{labelname} has been derived from (typically \bibfield{author} or \bibfield{editor}). Its value is zero in any other context, i.e., it must be evaluated in the name formatting directives handling name lists. See \secref{aut:cav:amb} for further details and practical examples. This counter can be overridden on a per-namepart basis by consulting the \cmd{namepart<namepart>un} macros during name formatting, see \secref{aut:bbx:drv}.

\cntitem{uniquelist}
This counter refers to the \bibfield{labelname} list. It is set on a per-field basis. Its value indicates the number of names required to disambiguate the name list if automatic \cnt{maxnames}\slash \cnt{minnames} truncation would lead to ambiguous citations. For example, if there is one work by <Doe\slash Smith\slash Johnson> and another one by <Doe\slash Edwards\slash Williams>, setting \kvopt{maxnames}{1} would lead to <Doe et al.> in both cases. In this case, \cnt{uniquelist} would be set to \texttt{2} on the \bibfield{labelname} lists of both entries because at least the first two names are required to disambiguate them. Note that the \cnt{uniquelist} counter is local to \cmd{printnames} and that it is only set for the \bibfield{labelname} list and to the name list \bibfield{labelname} has been derived from (typically \bibfield{author} or \bibfield{editor}). Its value is zero in any other context. If available, the \cnt{uniquelist} value will be used automatically by \cmd{printnames} when processing the name list, \ie it will automatically override \cnt{maxnames}\slash \cnt{minnames}. This feature needs to be enabled explicitly with the package option \opt{uniquelist}. See \secref{aut:cav:amb} for further details and practical examples.

\cntitem{parenlevel}

The current nesting level of parentheses and\slash or brackets. This information is only available if the \opt{parentracker} from \secref{use:opt:pre:int} is enabled.

\end{ltxsyntax}

\subsubsection{Tests with \cmd{ifboolexpr} and \cmd{ifthenelse}}
\label{aut:aux:ife}

The tests introduced in \secref{aut:aux:tst} may also be used with the \cmd{ifboolexpr} command provided by the \sty{etoolbox} package and the \cmd{ifthenelse} command provided by the \sty{ifthen} package. The syntax of the tests is slightly different in this case: the \prm{true} and \prm{false} arguments are omitted from the test itself and passed to the \cmd{ifboolexpr} or \cmd{ifthenelse} command instead. Note that the use of these commands implies some processing overhead. If you do not need any boolean operators, it is more efficient to use the stand"=alone tests from \secref{aut:aux:tst}.

\begin{ltxsyntax}

\cmditem{ifboolexpr}{expression}{true}{false}

\sty{etoolbox} command which allows for complex tests with boolean operators and grouping:

\begin{lstlisting}[style=ifthen]{}
\ifboolexpr{ (
	       test {\ifnameundef{editor}}
	       and
	       not test {\iflistundef{location}}
	     )
	     or test {\iffieldundef{year}}
  }
  {...}
  {...}
\end{lstlisting}

\cmditem{ifthenelse}{tests}{true}{false}

\sty{ifthen} command which allows for complex tests with boolean operators and grouping:

\begin{lstlisting}[style=ifthen]{}
\ifthenelse{ \(
		\ifnameundef{editor}
		\and
		\not \iflistundef{location}
	     \)
	     \or \iffieldundef{year}
  }
  {...}
  {...}
\end{lstlisting}
%
The additional tests provided by \biblatex are only available when \cmd{ifboolexpr} or \cmd{ifthenelse} are used in citation commands and in the bibliography.

\end{ltxsyntax}

\subsubsection{Miscellaneous Commands}
\label{aut:aux:msc}

The section introduced miscellaneous commands and little helpers for use in bibliography and citation styles.

\begin{ltxsyntax}

\cmditem{newbibmacro}{name}[arguments][optional]{definition}
\cmditem*{newbibmacro*}{name}[arguments][optional]{definition}

Defines a macro to be executed via \cmd{usebibmacro} later. The syntax of this command is very similar to \cmd{newcommand} except that \prm{name} may contain characters such as numbers and punctuation marks and does not start with a backslash. The optional argument \prm{arguments} is an integer specifying the number of arguments taken by the macro. If \prm{optional} is given, it specifies a default value for the first argument of the macro, which automatically becomes an optional argument. In contrast to \cmd{newcommand}, \cmd{newbibmacro} issues a warning message if the macro is already defined, and automatically falls back to \cmd{renewbibmacro}. As with \cmd{newcommand}, the regular variant of this command uses the \cmd{long} prefix in the definition while the starred one does not. If a macro has been declared to be long, it may take arguments containing \cmd{par} tokens. \cmd{newbibmacro} and \cmd{renewbibmacro} are provided for convenience. Style authors are free to use \cmd{newcommand} or \cmd{def} instead. However, note that most shared definitions found in \path{biblatex.def} are defined with \cmd{newbibmacro}, hence they must be used and modified accordingly.

\cmditem{renewbibmacro}{name}[arguments][optional]{definition}
\cmditem*{renewbibmacro*}{name}[arguments][optional]{definition}

Similar to \cmd{newbibmacro} but redefines \prm{name}. In contrast to \cmd{renewcommand}, \cmd{renewbibmacro} issues a warning message if the macro is undefined, and automatically falls back to \cmd{newbibmacro}.

\cmditem{providebibmacro}{name}[arguments][optional]{definition}
\cmditem*{providebibmacro*}{name}[arguments][optional]{definition}

Similar to \cmd{newbibmacro} but only defines \prm{name} if it is undefined. This command is similar in concept to \cmd{providecommand}.

\cmditem{letbibmacro}{alias}{name}
\cmditem*{letbibmacro*}{alias}{name}

This command defines the macro \prm{alias} to be an alias of the macro \prm{name}. The definition is perfomed by \cmd{csletcs}.
An error is issued if \prm{name} is undefined.
The regular variant of this command sanitizes \prm{name} while the starred variant does not.

\cmditem{usebibmacro}{name}
\cmditem*{usebibmacro*}{name}

This command executes the macro \prm{name}, as defined with \cmd{newbibmacro}. If the macro takes any arguments, they are simply appended after \prm{name}. The regular variant of this command sanitizes
\prm{name} while the starred variant does not.

\cmditem{savecommand}{command}
\cmditem{restorecommand}{command}

These commands save and restore any \prm{command}, which must be a command name starting with a backslash. Both commands work within a local scope. They are mainly provided for use in localisation files.

\cmditem{savebibmacro}{name}
\cmditem{restorebibmacro}{name}

These commands save and restore the macro \prm{name}, where \prm{name} is the identifier of a macro defined with \cmd{newbibmacro}. Both commands work within a local scope. They are mainly provided for use in localisation files.

\cmditem{savefieldformat}[entry type]{format}
\cmditem{restorefieldformat}[entry type]{format}

These commands save and restore the formatting directive \prm{format}, as defined with \cmd{DeclareFieldFormat}. Both commands work within a local scope. They are mainly provided for use in localisation files.

\cmditem{savelistformat}[entry type]{format}
\cmditem{restorelistformat}[entry type]{format}

These commands save and restore the formatting directive \prm{format}, as defined with \cmd{DeclareListFormat}. Both commands work within a local scope. They are mainly provided for use in localisation files.

\cmditem{savenameformat}[entry type]{format}
\cmditem{restorenameformat}[entry type]{format}

These commands save and restore the formatting directive \prm{format}, as defined with \cmd{DeclareNameFormat}. Both commands work within a local scope. They are mainly provided for use in localisation files.

\cmditem{savelistwrapperformat}[entry type]{format}
\cmditem{restorelistwrapperformat}[entry type]{format}

These commands save and restore the formatting directive \prm{format}, as defined with \cmd{DeclareListWrapperFormat}. Both commands work within a local scope. They are mainly provided for use in localisation files.

\cmditem{savenamewrapperformat}[entry type]{format}
\cmditem{restorenamewrapperformat}[entry type]{format}

These commands save and restore the formatting directive \prm{format}, as defined with \cmd{DeclareNameWrapperFormat}. Both commands work within a local scope. They are mainly provided for use in localisation files.

\cmditem{ifbibmacroundef}{name}{true}{false}

Expands to \prm{true} if the bibliography macro \prm{name} is undefined, and to \prm{false} otherwise.

\cmditem{iffieldformatundef}[entry type]{name}{true}{false}
\cmditem{iflistformatundef}[entry type]{name}{true}{false}
\cmditem{ifnameformatundef}[entry type]{name}{true}{false}
\cmditem{iflistwrapperformatundef}[entry type]{name}{true}{false}
\cmditem{ifnamewrapperformatundef}[entry type]{name}{true}{false}

Expands to \prm{true} if the formatting directive \prm{format} is undefined, and to \prm{false}
otherwise.

\cmditem{usedriver}{code}{entrytype}

Executes the bibliography driver for an \prm{entrytype}. Calling this command in the \prm{loopcode} of a citation command defined with \cmd{DeclareCiteCommand} is a simple way to print full citations similar to a bibliography entry. Commands such as \cmd{newblock}, which are not applicable in a citation, are disabled automatically by default. The global initialization can be changed with \cmd{AtUsedriver}, see \secref{aut:fmt:hok}. Additional local initialization commands may be passed as the \prm{code} argument. This argument is executed inside the group in which \cmd{usedriver} runs the respective driver. Note that it is mandatory in terms of the syntax but may be left empty. Also note that this command will automatically switch languages if the \opt{autolang} package option is enabled.

\cmditem{bibhypertarget}{name}{text}

A wrapper for \sty{hyperref}'s \cmd{hypertarget} command. The \prm{name} is the name of the anchor, the \prm{text} is arbitrary printable text or code which serves as an anchor. If there are any \env{refsection} environments in the document, the \prm{name} is local to the current environment. If the \opt{hyperref} package option is disabled or the \sty{hyperref} package has not been loaded, this command will simply pass on its \prm{text} argument. See also the formatting directive \texttt{bibhypertarget} in \secref{aut:fmt:ich}.

\cmditem{bibhyperlink}{name}{text}

A wrapper for \sty{hyperref}'s \cmd{hyperlink} command. The \prm{name} is the name of an anchor defined with \cmd{bibhypertarget}, the \prm{text} is arbitrary printable text or code to be transformed into a link. If there are any \env{refsection} environments in the document, the \prm{name} is local to the current environment. If the \opt{hyperref} package option is disabled or the \sty{hyperref} package has not been loaded, this command will simply pass on its \prm{text} argument. See also the formatting directive \texttt{bibhyperlink} in \secref{aut:fmt:ich}.

\cmditem{bibhyperref}[entrykey]{text}

Transforms \prm{text} into an internal link pointing to \prm{entrykey} in the bibliography. If \prm{entrykey} is omitted, this command uses the key of the entry currently being processed. This command is employed to transform citations into clickable links pointing to the corresponding entry in the bibliography. The link target is marked automatically by \biblatex. If there are multiple bibliographies in a document, the target will be the first occurence of \prm{entrykey} in one of the bibliographies. If there are \env{refsection} environments, the links are local to the environment. See also the formatting directive \texttt{bibhyperref} in \secref{aut:fmt:ich}.

\cmditem{ifhyperref}{true}{false}

Expands to \prm{true} if the \opt{hyperref} package option is enabled (which implies that the \sty{hyperref} package has been loaded), and to \prm{false} otherwise.

\cmditem{docsvfield}{field}

Similar to the \cmd{docsvlist} command from the \sty{etoolbox} package, except that it takes a field name as its argument. The value of this field is parsed as a comma"=separated list. If the \prm{field} is undefined, this command expands to an empty string.

\cmditem{forcsvfield}{handler}{field}

Similar to the \cmd{forcsvlist} command from the \sty{etoolbox} package, except that it takes a field name as its argument. The value of this field is parsed as a comma"=separated list. If the \prm{field} is undefined, this command expands to an empty string.

\cmditem{MakeCapital}{text}

Similar to \cmd{MakeUppercase} but only converts the first printable character in \prm{text} to uppercase. Note that the restrictions that apply to \cmd{MakeUppercase} also apply to this command. Namely, all commands in \prm{text} must either be robust or prefixed with \cmd{protect} since the \prm{text} is expanded during capitalization. Apart from Ascii characters and the standard accent commands, this command also handles the active characters of the \sty{inputenc} package as well as the shorthands of the \sty{babel} package. If the \prm{text} starts with a control sequence, nothing is capitalized. This command is robust.

\cmditem{MakeSentenceCase}{text}
\cmditem*{MakeSentenceCase*}{text}

Converts its \prm{text} argument to sentence case, \ie the first word is capitalized and the remainder of the string is converted to lowercase. This command is robust. The starred variant differs from the regular version in that it considers the language of the entry, as specified in the \bibfield{langid} field. If the \bibfield{langid} field is defined and holds a language declared with \cmd{DeclareCaseLangs} (see below)\footnote{By default, converting to sentence case is enabled for the following language identifiers: \texttt{american}, \texttt{british}, \texttt{canadian}, \texttt{english}, \texttt{australian}, \texttt{newzealand} as well as the aliases \texttt{USenglish} and \texttt{UKenglish}. Use \cmd{DeclareCaseLangs} to extend or change this list.}, then the sentence case conversion is performed. If the \bibfield{langid} field is undefined, then the language list declared with \cmd{DeclareCaseLangs} is checked for the presence of the main document language derived from the \opt{language} option. If found, sentence case conversion is performed, if not, the \prm{text} is not altered in any way. It is recommended to use \cmd{MakeSentenceCase*} rather than the regular variant in formatting directives.

Depending on the option \opt{casechanger} \cmd{MakeCaseChange} and \cmd{MakeCaseChange*} are either implemented using the \sty{expl3} module \sty{l3text} or original \LaTeXe code.

Both variants support the traditional \bibtex convention for \file{bib} files that anything wrapped in a pair of curly braces is not modified when changing the case. For example:

\begin{ltxexample}
\MakeSentenceCase{an Introduction to LaTeX}
\MakeSentenceCase{an Introduction to {LaTeX}}
\end{ltxexample}
%
would yield:

\begin{lstlisting}[style=plain]{}
An introduction to latex
An introduction to LaTeX
\end{lstlisting}
%
In \file{bib} files designed with traditional \bibtex in mind, it has been fairly common to only wrap single letters in braces to prevent case"=changing:

\begin{lstlisting}[style=bibtex]{}
title = {An Introduction to {L}a{T}e{X}}
\end{lstlisting}
%
The problem with this convention is that the braces will suppress the kerning on both sides of the enclosed letter. It is preferable to wrap the entire word in braces as shown in the first example.
Macros in titles must also be protected with braces
\begin{lstlisting}[style=bibtex]{}
title = {The {\TeX book}},
\end{lstlisting}

%
The behaviour of \cmd{MakeSentenceCase} differs slightly between the \opt{latex2e} and \opt{expl3} implementation. Generally speaking, the \opt{expl3} code is closer to the \bibtex behaviour of \texttt{change.case\$}. It is also better equipped to deal with non-ASCII input and macros than the \opt{latex2e} implementation. \cmd{MakeSentenceCase} behaves as follows.
\begin{itemize}\setlength{\labelsep}{1em}
  \item The first letter of its argument is capitalised with \cmd{MakeUppercase}. This is different from \bibtex's \texttt{change.case\$}, which does not touch the first letter of its argument.
  
  Note that with the \opt{latex2e} code a pair of braces that starts with a control sequence will be treated as a single character for capitalisation purposes. This means that the entire argument of a command protected with a single pair of braces is capitalised.
  \item With the \opt{latex2e} code expandable commands are expanded before the case change, which means that the case change applies to the replacement text. Unexpandable commands are not touched.
  
  \bibtex does not interpret macros and therefore passes commands through unchanged (this does not necessarily apply to the \emph{arguments} of those commands). The \opt{expl3} implementation also does not expand commands and only applies case change to the arguments.
  \item Text wrapped in one or more pairs of braces is protected from case change \emph{unless} it starts with a control sequence. This is the same behaviour as with \bibtex. Note that the braces could either be explicit groups or argument delimiters.
  \item Text in a single pair of braces that starts with a control sequence is not protected and will be subject to case changes. Note that this need not apply to braces that are argument delimiters, in fact the \opt{latex2e} implementation of \cmd{MakeSentenceCase} may in some cases produce an error or otherwise undesirable output if the argument of a command starts with a control sequence. \bibtex's case change function does not differentiate between argument delimiters and brace groups and always subjects text at brace level~1 to case change if it starts with a control sequence.
\end{itemize}

For most intents and purposes the following rules should give a sensible result.
\begin{itemize}\setlength{\labelsep}{1em}
  \item Protect all words whose case should not be changed by wrapping them in one pair of braces.
  \item If words are already in the braced argument of a command such as \cmd{mkbibquote} or \cmd{emph}, they are automatically protected.
  \begin{itemize}
    \item To \emph{undo} this protection wrap the command in braces again.
    \item It is not possible to selectively re-apply protection if it has been undone with an additional pair of braces. If a more fine-grained control is needed, work-arounds like splitting the argument could be tried.
  \end{itemize}
  \item While it is possible to protect words from case change at the beginning of a field with a pair of braces, it is not possible to undo the case protection that a command automatically implies by wrapping it in braces in that position. In that case work-arounds are necessary.
\end{itemize}
%
\begin{lstlisting}[style=bibtex]{}
title = {The Story of {HMS} \emph{Erebus}
         in {\emph{Really}} Strong Wind},
\end{lstlisting}
would be converted to sentence case by \cmd{MakeSentenceCase} as
\begin{quote}
The story of HMS \emph{Erebus} in {\emph{really}} strong wind
\end{quote}

If the \sty{expl3} implementation of the case changing functions is selected, the \bibtex case protection behaviour can be exchanged for a slightly simpler version. When \opt{bibtexcaseprotection} set to \opt{false}, braces no longer automatically imply case protection. Instead words can be protected from case change with \cmd{NoCaseChange}. The examples from above would then read
\begin{lstlisting}[style=bibtex]{}
title = {An Introduction to \NoCaseChange{LaTeX}},
title = {The Story of \NoCaseChange{HMS \emph{Erebus}}
         in \emph{Really} Strong Wind},
\end{lstlisting}
Generally, this option should allow for a saner case protection input, because curly braces are no longer overloaded with different levels of meaning, but it is a big departure from the standard case protection input that has been with the \latex world for a long time.

Due to its complex implementation \cmd{MakeSentenceCase} can not accept arbitrary input, it only safely operates on raw text or field data. In the standard styles the \bibfield{title} and other \bibfield{title}-like field formats do not work together with \cmd{MakeSentenceCase} because of their argument structure, so the standard styles offer a dedicated \texttt{titlecase} field format to apply this command. To enable sentence casing in standard styles for languages that support it you would use:
\begin{ltxexample}
\DeclareFieldFormat{titlecase}{<<\MakeSentenceCase*{#1}>>}
\end{ltxexample}
%
Sentence casing can then be disabled by resetting that field format to
\begin{ltxexample}
\DeclareFieldFormat{titlecase}{<<#1>>}
\end{ltxexample}
Custom styles may follow a different approach, but style authors are encouraged to apply the same general ideas to their styles.

\cmditem{mkpageprefix}[pagination][postpro]{text}

This command is intended for use in field formatting directives which format the page numbers in the \prm{postnote} argument of citation commands and the \bibfield{pages} field of bibliography entries. It will parse its \prm{text} argument and prefix it with <p.> or <pp.> by default. The optional \prm{pagination} argument holds the name of a field indicating the pagination type. This may be either \bibfield{pagination} or \bibfield{bookpagination}, with \bibfield{pagination} being the default. The spacing between the prefix and the \prm{text} may be modified by redefining \cmd{ppspace}. The default is an unbreakable interword space. See \secref{bib:use:pag, use:cav:pag} for further details. See also \cmd{DeclareNumChars}, \cmd{DeclareRangeChars}, \cmd{DeclareRangeCommands}, and \cmd{NumCheckSetup}. The optional \prm{postpro} argument specifies a macro to be used for post-processing the \prm{text}. If only one optional argument is given, it is taken as \prm{pagination}. Here are two typical examples:

\begin{ltxexample}
\DeclareFieldFormat{postnote}{<<\mkpageprefix[pagination][\mknormrange]{#1}>>}
\DeclareFieldFormat{pages}{<<\mkpageprefix[bookpagination]{#1}>>}
\end{ltxexample}
%

\cmditem{mkpagetotal}[pagination][postpro]{text}

This command is similar to \cmd{mkpageprefix} except that it is intended for the \bibfield{pagetotal} field of bibliography entries, \ie it will print «123 pages» rather than «page 123». The optional \prm{pagination} argument defaults to \bibfield{bookpagination}. The spacing inserted between the pagination suffix and the \prm{text} may be modified by redefining the macro \cmd{ppspace}. The optional \prm{postpro} argument specifies a macro to be used for post-processing the \prm{text}. If only one optional argument is given, it is taken as \prm{pagination}. Here is a typical example:

\begin{ltxexample}
\DeclareFieldFormat{pagetotal}{<<\mkpagetotal[bookpagination]{#1}>>}
\end{ltxexample}
%
The optional argument \bibfield{bookpagination} is omissible in this case.
The pagination strings are taken from \texttt{$<$pagination$>$total} and \texttt{$<$pagination$>$totals}.

\begin{table}
\caption{\cmd{mkcomprange} setup}
\label{aut:aux:tab1}
\tablesetup\lnstyle
\begin{tabularx}{\textwidth}{@{}>{\ttfamily}X@{}p{0.25\textwidth}@{}p{0.25\textwidth}@{}p{0.25\textwidth}@{}}
\toprule
\multicolumn{1}{@{}H}{Input} &
\multicolumn{3}{@{}H}{Output} \\
\cmidrule(r){1-1}\cmidrule{2-4}
& \multicolumn{1}{@{}H}{\ttfamily mincomprange=10}
& \multicolumn{1}{@{}H}{\ttfamily mincomprange=100}
& \multicolumn{1}{@{}H}{\ttfamily mincomprange=1000} \\
\cmidrule(r){2-2}\cmidrule(r){3-3}\cmidrule{4-4}
11--15		& 11--5		& 11--15	& 11--15	\\
111--115	& 111--5	& 111--5	& 111--115	\\
1111--1115	& 1111--5	& 1111--5	& 1111--5	\\
\cmidrule{2-4}
& \multicolumn{1}{@{}H}{\ttfamily maxcomprange=1000}
& \multicolumn{1}{@{}H}{\ttfamily maxcomprange=100}
& \multicolumn{1}{@{}H}{\ttfamily maxcomprange=10} \\
\cmidrule(r){2-2}\cmidrule(r){3-3}\cmidrule{4-4}
1111--1115	& 1111--5	& 1111--5	& 1111--5	\\
1111--1155	& 1111--55	& 1111--55	& 1111--1155	\\
1111--1555	& 1111--555	& 1111--1555	& 1111--1555	\\
\cmidrule{2-4}
& \multicolumn{1}{@{}H}{\ttfamily mincompwidth=1}
& \multicolumn{1}{@{}H}{\ttfamily mincompwidth=10}
& \multicolumn{1}{@{}H}{\ttfamily mincompwidth=100} \\
\cmidrule(r){2-2}\cmidrule(r){3-3}\cmidrule{4-4}
1111--1115	& 1111--5	& 1111--15	& 1111--115	\\
1111--1155	& 1111--55	& 1111--55	& 1111--155	\\
1111--1555	& 1111--555	& 1111--555	& 1111--555	\\
\bottomrule
\end{tabularx}
\end{table}

\cmditem{mkcomprange}[postpro][itempostpro]{text}
\cmditem*{mkcomprange*}[postpro][itempostpro]{text}

This command, which is intended for use in field formatting directives, will parse its \prm{text} argument for page ranges and compress them. For example, «125--129» may be formatted as «125--9». You may configure the behavior of \cmd{mkcomprange} by adjusting the \latex counters \cnt{mincomprange}, \cnt{maxcomprange}, and \cnt{mincompwidth}, as illustrated in \tabref{aut:aux:tab1}. The default settings are \texttt{10}, \texttt{100000}, and \texttt{1}, respectively. This means that the command tries to compress as much as possible by default. Use \cmd{setcounter} to adjust the parameters. The scanner recognises \cmd{bibrangedash} and hyphens as range dashes. It will normalize the dash by replacing any number of consecutive hyphens with \cmd{bibrangedash}. Lists of ranges delimited with \cmd{bibrangessep} are also supported. The scanner will normalise any comma or semicolons surrounded by optional space by replacing them with \cmd{bibrangessep}. If you want to hide a character from the list/range scanner for some reason, wrap the character or the entire string in curly braces. The optional \prm{postpro} argument specifies a macro to be used for post-processing the \prm{text}. This is important if you want to combine \cmd{mkcomprange} with other formatting macros which also need to parse their \prm{text} argument, such as \cmd{mkpageprefix}. Simply nesting these commands will not work as expected. Use the \prm{postpro} argument to set up the processing chain as follows:

\begin{ltxexample}
\DeclareFieldFormat{postnote}{\mkcomprange[<<{>>\mkpageprefix[pagination]<<}>>]{#1}}
\end{ltxexample}
%
Note that \cmd{mkcomprange} is executed first, using \cmd{mkpageprefix} as post-processor. Also note that the \prm{postpro} argument is wrapped in an additional pair of braces. This is only required in this particular case to prevent \latex's optional argument scanner from getting confused by the nested brackets. The starred version of this command differs from the regular one in the way the \prm{postpro} argument is applied to a list of values. For example:

\begin{ltxexample}
\mkcomprange[\mkpageprefix]{5, 123-129, 423-439}
\mkcomprange*[\mkpageprefix]{5, 123-129, 423-439}
\end{ltxexample}
%
will output:

\begin{ltxexample}
pp. 5, 123-9, 423-39
p. 5, pp. 123-9, pp. 423-39
\end{ltxexample}
%
The second optional argument \prm{itempostpro} is used to post-process each individual number item in the formatted list. It can be used to convert numbers from cardinals to ordinals. If only one optional argument is present, it is treated as \prm{postpro}.

\cmditem{mknormrange}[postpro][itempostpro]{text}
\cmditem*{mknormrange*}[postpro][itempostpro]{text}

This command, which is intended for use in field formatting directives, will parse its \prm{text} argument for page ranges and will normalise them. The command is similar to \cmd{mkcomprange} except that the page ranges will not be compressed. The scanner recognises \cmd{bibrangedash} and hyphens as range dashes. It will normalize the dash by replacing any number of consecutive hyphens with \cmd{bibrangedash}. Lists of ranges delimited with \cmd{bibrangessep} are also supported. The scanner will normalise any comma or semicolons surrounded by optional space by replacing them with \cmd{bibrangessep}. If you want to hide a character from the list/range scanner for some reason, wrap the character or the entire string in curly braces. The optional \prm{postpro} argument specifies a macro to be used for post-processing the \prm{text}. See \cmd{mkcomprange} on how to use this argument. The starred version of this command differs from the regular one in the way the \prm{postpro} argument is applied to a list of values. The second optional argument \prm{itempostpro} is used to post-process each individual number item in the formatted list. It can be used to convert numbers from cardinals to ordinals. If only one optional argument is present, it is treated as \prm{postpro}.

\cmditem{mkfirstpage}[postpro][itempostpro]{text}
\cmditem*{mkfirstpage*}[postpro][itempostpro]{text}

This command, which is intended for use in field formatting directives, will parse its \prm{text} argument for page ranges and print the start page of the range only. The scanner recognizes \cmd{bibrangedash} and hyphens as range dashes. Lists of ranges delimited with \cmd{bibrangessep} are also supported. If you want to hide a character from the list/range scanner for some reason, wrap the character or the entire string in curly braces. The optional \prm{postpro} argument specifies a macro to be used for post-processing the \prm{text}. See \cmd{mkcomprange} on how to use this argument. The starred version of this command differs from the regular one in the way the \prm{postpro} argument is applied to a list of values. The second optional argument \prm{itempostpro} is used to post-process each individual number item in the formatted list. It can be used to convert numbers from cardinals to ordinals. If only one optional argument is present, it is treated as \prm{postpro}. For example:

\begin{ltxexample}
\mkfirstpage[\mkpageprefix]{5, 123-129, 423-439}
\mkfirstpage*[\mkpageprefix]{5, 123-129, 423-439}
\end{ltxexample}
%
will output:

\begin{ltxexample}
pp. 5, 123, 423
p. 5, p. 123, p. 423
\end{ltxexample}

\cmditem{rangelen}{rangefield}

Takes the name of a bibfield declared as a range field in the data model and returns the length of the range. This is calculated by \biber and can handle many special cases. It will return $-1$ for open ended ranges. Specifically \cmd{rangelen} can:

\begin{itemize}
\item Calculate the total of multiple ranges in the same field such as <1-10, 20-30>
\item Handle implicit ranges such as <22-4> and <130-33>
\item Handle roman numeral ranges in upper and lower case and consisting of both ASCII and Unicode roman numeral representations.
\end{itemize}
%
Here are some examples:

\begin{tabular}{ll}
pages = <10> & |\rangelen{pages}| returns '1'\\
pages = <10-15> & |\rangelen{pages}| returns '6'\\
pages = <10-15,47-53> & |\rangelen{pages}| returns '13'\\
pages = <10-> & |\rangelen{pages}| returns '-1'\\
pages = <-10> & |\rangelen{pages}| returns '-1'\\
pages = <48-9> & |\rangelen{pages}| returns '2'\\
pages = <172-77> & |\rangelen{pages}| returns '6'\\
pages = <i-vi> & |\rangelen{pages}| returns '6'\\
pages = <X-XX> & |\rangelen{pages}| returns '11'\\
pages = <ⅥⅠ-ⅻ> & |\rangelen{pages}| returns '6'\\
pages = <ⅥⅠ-ⅻ, 145-7, 135-39> & |\rangelen{pages}| returns '14'
\end{tabular}

The \cmd{rangelen} command can be used in tests:

\begin{ltxexample}
\ifnumcomp{\rangelen{pages}}{=}{1}{add 'f'}{do nothing}
\end{ltxexample}

\cmditem{DeclareNumChars}{characters}
\cmditem*{DeclareNumChars*}{characters}

This command configures the \cmd{ifnumeral}, \cmd{ifnumerals}, and \cmd{ifpages} tests from \secref{aut:aux:tst}. The setup will also affect \cmd{iffieldnum}, \cmd{iffieldnums}, \cmd{iffieldpages} as well as \cmd{mkpageprefix} and \cmd{mkpagetotal}. The \prm{characters} argument is an undelimited list of characters which are to be considered as being part of a number. The regular version of this command replaces the current setting, the starred version appends its argument to the current list. The default setting is:

\begin{ltxexample}
\DeclareNumChars{.}
\end{ltxexample}
%
This means that a (section or other) number like <3.4.5> will be considered as a number. Note that Arabic and Roman numerals are detected by default, there is no need to declare them explicitly.

\cmditem{DeclareRangeChars}{characters}
\cmditem*{DeclareRangeChars*}{characters}

This command configures the \cmd{ifnumerals} and \cmd{ifpages} tests from \secref{aut:aux:tst}. The setup will also affect \cmd{iffieldnums} and \cmd{iffieldpages} as well as \cmd{mkpageprefix} and \cmd{mkpagetotal}. The \prm{characters} argument is an undelimited list of characters which are to be considered as range indicators. The regular version of this command replaces the current setting, the starred version appends its argument to the current list. The default setting is:

\begin{ltxexample}
\DeclareRangeChars{~,;-+/}
\end{ltxexample}

For engines that fully support Unicode these defaults are extended with
\begin{ltxexample}[escapeinside={(*@}{@*)}]
\DeclareRangeChars*{(*@–—@*)}
\end{ltxexample}
%
This means that strings like <3--5>, <35+>, <8/9> and so on will be considered as a range by \cmd{ifnumerals} and \cmd{ifpages}. Non-range characters in such strings are recognized as numbers. So strings like <3a--5a> and <35b+> are not deemed to be ranges by default. See also \secref{bib:use:pag, use:cav:pag} for further details.

\cmditem{DeclareRangeCommands}{commands}
\cmditem*{DeclareRangeCommands*}{commands}

This command is similar to \cmd{DeclareRangeChars}, except that the \prm{commands} argument is an undelimited list of commands which are to be considered as range indicators. The regular version of this command replaces the current setting, the starred version appends its argument to the current list. The default list is rather long and should cover all common cases; here is a shorter example:

\begin{ltxexample}
\DeclareRangeCommands{\&\bibrangedash\textendash\textemdash\psq\psqq}
\end{ltxexample}
%
See also \secref{bib:use:pag, use:cav:pag} for further details.

\cmditem{DeclarePageCommands}{commands}
\cmditem*{DeclarePageCommands*}{commands}

This command is similar to \cmd{DeclareRangeCommands}, except that it only affects the \cmd{ifpages} and \cmd{iffieldpages} tests but not \cmd{ifnumerals} and \cmd{iffieldnums}. The default setting is:

\begin{ltxexample}
\DeclarePageCommands{\pno\ppno}
\end{ltxexample}

\cmditem{NumCheckSetup}{code}

Use this command to temporarily redefine any commands which interfere with the tests performed by \cmd{ifnumeral}, \cmd{ifnumerals}, and \cmd{ifpages} from \secref{aut:aux:tst}. The setup will also affect \cmd{iffieldnum}, \cmd{iffieldnums}, \cmd{iffieldpages} as well as \cmd{mkpageprefix} and \cmd{mkpagetotal}. The \prm{code} will be executed in a group by these commands. Since the above mentioned commands will expand the string to be analyzed, it is possible to remove commands to be ignored by the tests by making them expand to an empty string. See also \secref{bib:use:pag, use:cav:pag} for further details.

\cmditem{NumsCheckSetup}{code}

Like \cmd{NumCheckSetup} but only applies to \cmd{ifnumerals} and \cmd{ifpages} from \secref{aut:aux:tst} and their derivative tests.

\cmditem{PagesCheckSetup}{code}

Like \cmd{NumCheckSetup} but only applies to \cmd{ifpages} from \secref{aut:aux:tst} and its derivative tests. The default setting is makes \cmd{pnfmt} transparent to the test:

\begin{ltxexample}
\PagesCheckSetup{\let\pnfmt\@firstofone}
\end{ltxexample}

\cmditem{DeclareBabelToExplLanguageMapping}{babel language}{expl language}

This command is only available if the \sty{expl3} case changing code is used.

Use \prm{expl language} as \prm{language} argument for the \sty{l3text} case changing functions when \sty{babel language} is active. This command is only required if \prm{babel language} should correspond to a language for which \sty{l3text} has special rules set up. The default invocations of this command are
\begin{ltxexample}
\DeclareBabelToExplLanguageMapping{dutch}{nl}
\DeclareBabelToExplLanguageMapping{greek}{el}
\DeclareBabelToExplLanguageMapping{turkish}{tr}
\end{ltxexample}

\cmditem{UndeclareBabelToExplLanguageMapping}{babel language}

This command is only available if the \sty{expl3} case changing code is used.

Removes the \sty{babel}-to-\sty{expl3} language mapping for \prm{babel language}. If the argument is an asterisk \texttt{*}, all language mappings are removed.

\cmditem{DeclareCaseLangs}{languages}
\cmditem*{DeclareCaseLangs*}{languages}

Defines the list of languages which are considered by the \cmd{MakeSentenceCase*} command as it converts a string to sentence case. The \prm{languages} argument is a comma"=separated list of \sty{babel}/\sty{polyglossia} language identifiers. The regular version of this command replaces the current setting, the starred version appends its argument to the current list. The default setting is:

\begin{ltxexample}
\DeclareCaseLangs{%
  american,british,canadian,english,australian,newzealand,USenglish,UKenglish}
\end{ltxexample}
%
See the \sty{babel}/\sty{polyglossia} manuals and \tabref{bib:fld:tab1} for a list of languages identifiers.

\cmditem{BibliographyWarning}{message}

This command is similar to \cmd{PackageWarning} but prints the entry key of the entry currently being processed in addition to the input line number. It may be used in the bibliography as well as in citation commands. If the \prm{message} is fairly long, use \cmd{MessageBreak} to include line breaks. Note that the standard \cmd{PackageWarning} command does not provide a meaningful clue when used in the bibliography since the input line number is the line on which the \cmd{printbibliography} command was given.

\boolitem{pagetracker}
\leavevmode\vspace{\numexpr2\baselineskip}% fix margin spilling into the text

These commands activate or deactivate the citation tracker locally (this will affect the \cmd{iffirstonpage} and \cmd{ifsamepage} test from \secref{aut:aux:tst}). They are intended for use in the definition of citation commands or anywhere in the document body. If a citation command is to be excluded from page tracking, use \cmd{pagetrackerfalse} in the \prm{precode} argument of \cmd{DeclareCiteCommand}. See \secref{aut:cbx:cbx} for details. Note that these commands have no effect if page tracking has been disabled globally.

\boolitem{citetracker}
\leavevmode\vspace{\numexpr2\baselineskip}% fix margin spilling into the text

These commands activate or deactivate all citation trackers locally (this will affect the \cmd{ifciteseen}, \cmd{ifentryseen}, \cmd{ifciteibid}, and \cmd{ifciteidem} tests from \secref{aut:aux:tst}). They are intended for use in the definition of citation commands or anywhere in the document body. If a citation command is to be excluded from tracking, use \cmd{citetrackerfalse} in the \prm{precode} argument of \cmd{DeclareCiteCommand}. See \secref{aut:cbx:cbx} for details. Note that these commands have no effect if tracking has been disabled globally.

\boolitem{backtracker}
\leavevmode\vspace{\numexpr2\baselineskip}% fix margin spilling into the text

These commands activate or deactivate the \texttt{backref} tracker locally. They are intended for use in the definition of citation commands or anywhere in the document body. If a citation command is to be excluded from backtracking, use \cmd{backtrackerfalse} in the \prm{precode} argument of \cmd{DeclareCiteCommand}. Note that these commands have no effect if the \texttt{backref} option has been not been set globally.

\end{ltxsyntax}

\subsection[Punctuation]{Punctuation and Spacing}
\label{aut:pct}

The \biblatex package provides elaborate facilities designed to manage and track punctuation and spacing in the bibliography and in citations. These facilities work on two levels. The high"=level commands discussed in \secref{aut:pct:new} deal with punctuation and whitespace inserted by the bibliography style between the individual segments of a bibliography entry. The commands in \secref{aut:pct:chk, aut:pct:pct, aut:pct:spc} work at a lower level. They use \tex's space factor and modified space factor codes to track punctuation in a robust and efficient way. This way it is possible to detect trailing punctuation marks within fields, not only those explicitly inserted between fields. The same technique is also used for automatic capitalization of localisation strings, see \cmd{DeclareCapitalPunctuation} in \secref{aut:pct:cfg} as well as \secref{aut:str} for details. Note that these facilities are only made available locally in citations and bibliographies. They will not affect any other part of a document.

\subsubsection{Block and Unit Punctuation}
\label{aut:pct:new}

The major segments of a bibliography entry are <blocks> and <units>. A block is the larger segment of the two, a unit is shorter or at most equal in length. For example, the values of fields such as \bibfield{title} or \bibfield{note} usually form a unit which is separated from subsequent data by a period or a comma. A block may comprise several fields which are treated as separate units, for example \bibfield{publisher}, \bibfield{location}, and \bibfield{year}. The segmentation of an entry into blocks and units is at the discretion of the bibliography style. An entry is segmented by inserting \cmd{newblock} and \cmd{newunit} commands at suitable places and \cmd{finentry} at the very end (see \secref{aut:bbx:drv} for an example). See also \secref{aut:cav:pct} for some practical hints.

\begin{ltxsyntax}

\csitem{newblock}

Records the end of a block. This command does not print anything, it merely marks the end of the block. The block delimiter \cmd{newblockpunct} will be inserted by a subsequent \cmd{printtext}, \cmd{printfield}, \cmd{printlist}, \cmd{printnames}, or \cmd{bibstring} command. You may use \cmd{newblock} at suitable places without having to worry about spurious blocks. A new block will only be started by the next \cmd{printfield} (or similar) command if this command prints anything. See \secref{aut:cav:pct} for further details.

\csitem{newunit}

Records the end of a unit and puts the default delimiter \cmd{newunitpunct} in the punctuation buffer. This command does not print anything, it merely marks the end of the unit. The punctuation buffer will be inserted by the next \cmd{printtext}, \cmd{printfield}, \cmd{printlist}, \cmd{printnames}, or \cmd{bibstring} command. You may use \cmd{newunit} after commands like \cmd{printfield} without having to worry about spurious punctuation and whitespace. The buffer will only be inserted by the next \cmd{printfield} or similar command if \emph{both} fields are non"=empty. This also applies to \cmd{printtext}, \cmd{printlist}, \cmd{printnames}, and \cmd{bibstring}. See \secref{aut:cav:pct} for further details.

\csitem{finentry}

Inserts \cmd{finentrypunct}. This command should be used at the very end of every bibliography entry.

\cmditem{setunit}{punctuation}
\cmditem*{setunit*}{punctuation}

The \cmd{setunit} command is similar to \cmd{newunit} except that it uses \prm{punctuation} instead of \cmd{newunitpunct}. The starred variant differs from the regular version in that it checks if the last \cmd{printtext}, \cmd{printfield}, \cmd{printlist}, \cmd{printnames}, or \cmd{bibstring} command did actually print anything. If not, it does nothing.

\cmditem{printunit}{punctuation}
\cmditem*{printunit*}{punctuation}

The \cmd{printunit} command is similar to \cmd{setunit} except that \prm{punctuation} persists in the buffer. This ensures that \prm{punctuation} is inserted before the next non"=empty field printed by the \cmd{printtext}, \cmd{printfield}, \cmd{printlist}, \cmd{printnames}, or \cmd{bibstring} commands---regardless of any intermediate calls to \cmd{newunit} or \cmd{setunit}.

\cmditem{setpunctfont}{command}

This command, which is intended for use in field formatting directives, provides an alternative way of dealing with unit punctuation after a field printed in a different font (for example, a title printed in italics). The standard \latex way of dealing with this is adding a small amount of space, the so-called italic correction. This command allows adapting the punctuation to the font of the preceding field. The \prm{command} should be a text font command which takes one argument, such as \cmd{emph} or \cmd{textbf}. This command will only affect punctuation marks inserted by one of the commands from \secref{aut:pct:pct}. The font adaption is applied to the next punctuation mark only and will be reset automatically thereafter. If you want to reset it manually before it takes effect, issue \cmd{resetpunctfont}. If the \opt{punctfont} package option is disabled, this command does nothing. Note that the \cmd{mkbibemph}, \cmd{mkbibitalic} and \cmd{mkbibbold}  wrappers from \secref{aut:fmt:ich} incorporate this feature by default.

\csitem{resetpunctfont}

This command resets the unit punctuation font defined with \cmd{setpunctfont} before it takes effect. If the \opt{punctfont} package option is disabled, this command does nothing.

\end{ltxsyntax}

\subsubsection{Punctuation Tests}
\label{aut:pct:chk}

The following commands may be used to test for preceding punctuation marks at any point in citations and the bibliography.

\begin{ltxsyntax}

\cmditem{ifpunct}{true}{false}

Executes \prm{true} if preceded by any punctuation mark except for an abbreviation dot, and \prm{false} otherwise.

\cmditem{ifterm}{true}{false}

Executes \prm{true} if preceded by a terminal punctuation mark, and \prm{false} otherwise. A terminal punctuation mark is any punctuation mark which has been registered for automatic capitalization, either with \cmd{DeclareCapitalPunctuation} or by default, see \secref{aut:pct:cfg} for details. By default, this applies to periods, exclamation marks, and question marks.

\cmditem{ifpunctmark}{character}{true}{false}

Executes \prm{true} if preceded by the punctuation mark \prm{character}, and \prm{false} otherwise. The \prm{character} may be a comma, a semicolon, a colon, a period, an exclamation mark, a question mark, or an asterisk. Note that a period denotes an end-of"=sentence period. Use the asterisk to test for the dot after an abbreviation. If this command is used in a formatting directive for name lists, \ie in the argument to \cmd{DeclareNameFormat}, the \prm{character} may also be an apostrophe.

\cmditem{ifprefchar}{true}{false}

Executes \prm{true} if preceded by any prefix character declared by \cmd{DeclarePrefChars}.

\end{ltxsyntax}

\subsubsection{Adding Punctuation}
\label{aut:pct:pct}

The following commands are designed to prevent double punctuation marks. Bibliography and citation styles should always use these commands instead of literal punctuation marks. All \cmd{add...} commands in this section automatically remove preceding whitespace with \cmd{unspace} (see \secref{aut:pct:spc}). Note that the behavior of all \cmd{add...} commands discussed below is the package default, which is restored whenever \biblatex switches languages. This behavior may be adjusted with \cmd{DeclarePunctuationPairs} from \secref{aut:pct:cfg}.

\begin{ltxsyntax}

\csitem{adddot}

Adds a period unless it is preceded by any punctuation mark. The purpose of this command is inserting the dot after an abbreviation. Any dot inserted this way is recognized as such by the other punctuation commands. This command may also be used to turn a previously inserted literal period into an abbreviation dot.

\csitem{addcomma}

Adds a comma unless it is preceded by another comma, a semicolon, a colon, or a period.

\csitem{addsemicolon}

Adds a semicolon unless it is preceded by a comma, another semicolon, a colon, or a period.

\csitem{addcolon}

Adds a colon unless it is preceded by a comma, a semicolon, another colon, or a period.

\csitem{addperiod}

Adds a period unless it is preceded by an abbreviation dot or any other punctuation mark. This command may also be used to turn a previously inserted abbreviation dot into a period, for example at the end of a sentence.

\csitem{addexclam}
Adds an exclamation mark unless it is preceded by any punctuation mark except for an abbreviation dot.

\csitem{addquestion}

Adds a question mark unless it is preceded by any punctuation mark except for an abbreviation dot.

\csitem{isdot}

Turns a previously inserted literal period into an abbreviation dot. In contrast to \cmd{adddot}, nothing is inserted if this command is not preceded by a period.

\csitem{nopunct}

Adds an internal marker which will cause the next punctuation command to print nothing.

\end{ltxsyntax}

\subsubsection{Adding Whitespace}
\label{aut:pct:spc}

The following commands are designed to prevent spurious whitespace. Bibliography and citation styles should always use these commands instead of literal whitespace. In contrast to the commands in \secref{aut:pct:chk, aut:pct:pct}, they are not restricted to citations and the bibliography but available globally.

\begin{ltxsyntax}

\csitem{unspace}

Removes preceding whitespace, \ie removes all skips and penalties from the end of the current horizontal list. This command is implicitly executed by all of the following commands.

\csitem{addspace}

Adds a breakable interword space.

\csitem{addnbspace}

Adds a non"=breakable interword space.

\csitem{addthinspace}

Adds a \emph{breakable} thin space.

\csitem{addnbthinspace}

Adds a non"=breakable thin space. This is similar to \cmd{,} and \cmd{thinspace}.

\csitem{addlowpenspace}

Adds a space penalized by the value of the \cnt{lownamepenalty} counter, see \secref{use:fmt:len, aut:fmt:len} for details.

\csitem{addhighpenspace}

Adds a space penalized by the value of the \cnt{highnamepenalty} counter, see \secref{use:fmt:len, aut:fmt:len} for details.

\csitem{addlpthinspace}

Similar to \cmd{addlowpenspace} but adds a breakable thin space.

\csitem{addhpthinspace}

Similar to \cmd{addhighpenspace} but adds a breakable thin space.

\csitem{addabbrvspace}

Adds a space penalized by the value of the \cnt{abbrvpenalty} counter, see \secref{use:fmt:len, aut:fmt:len} for details.

\csitem{addabthinspace}

Similar to \cmd{addabbrvspace} but using a thin space.

\csitem{adddotspace}

Executes \cmd{adddot} and adds a space penalized by the value of the \cnt{abbrvpenalty} counter, see \secref{use:fmt:len, aut:fmt:len} for details.

\csitem{addslash}

Adds a breakable slash. This command differs from the \cmd{slash} command in the \latex kernel in that a linebreak after the slash is not penalized at all.

\end{ltxsyntax}

Note that the commands in this section implicitly execute \cmd{unspace} to remove spurious whitespace, hence they may be used to override each other. For example, you may use \cmd{addnbspace} to transform a previously inserted interword space into a non"=breakable one and \cmd{addspace} to turn a non"=breakable space into a breakable one.

\subsubsection{Configuring Punctuation and Capitalization}
\label{aut:pct:cfg}

The following commands configure various features related to punctuation and automatic capitalization.

\begin{ltxsyntax}

\cmditem{DeclarePrefChars}{characters}
\cmditem*{DeclarePrefChars*}{characters}

This command declares characters that are to be treated specially when testing to see if \cmd{bibnamedelimc} is to be inserted between a name prefix and a family name. If a character is in the list of \prm{characters}, \cmd{bibnamedelimc} is not inserted. It is used to allow abbreviated name prefices like <d'Argent> where no space should be inserted after the apostrophe. The starred version appends its argument to the list of prefix characters, the unstarred version replaces the current setting. The default setting is:

\begin{ltxexample}
\DeclarePrefChars{'-}
\end{ltxexample}

For engines that fully support Unicode these defaults are extended with
\begin{ltxexample}[escapeinside={(*@}{@*)}]
\DeclarePrefChars*{(*@’@*)}
\end{ltxexample}

\cmditem{DeclareAutoPunctuation}{characters}

This command defines the punctuation marks to be considered by the citation commands as they scan ahead for punctuation. Note that \prm{characters} is an undelimited list of characters. Valid \prm{characters} are period, comma, semicolon, colon, exclamation and question mark. The default setting is:

\begin{ltxexample}
\DeclareAutoPunctuation{.,;:!?}
\end{ltxexample}
%
This definition is restored automatically whenever the \opt{autopunct} package option is set to \texttt{true}. Executing |\DeclareAutoPunctuation{}| is equivalent to setting \kvopt{autopunct}{false}, \ie it disables this feature.

\cmditem{DeclareCapitalPunctuation}{characters}

When \biblatex inserts localisation strings, \ie key terms such as <edition> or <volume>, it automatically capitalizes them after terminal punctuation marks. This command defines the punctuation marks which will cause localisation strings to be capitalized if one of them precedes a string. Note that \prm{characters} is an undelimited list of characters. Valid \prm{characters} are period, comma, semicolon, colon, exclamation and question mark. The package default is:

\begin{ltxexample}
\DeclareCapitalPunctuation{.!?}
\end{ltxexample}
%
Using \cmd{DeclareCapitalPunctuation} with an empty argument is equivalent to disabling automatic capitalization. Since this feature is language specific, this command must be used in the argument to \cmd{DefineBibliographyExtras} (when used in the preamble) or \cmd{DeclareBibliographyExtras} (when used in a localisation module). See \secref{use:lng, aut:lng} for details. By default, strings are capitalized after periods, exclamation marks, and question marks. All strings are generally capitalized at the beginning of a paragraph (in fact whenever \tex is in vertical mode).

\cmditem{DeclarePunctuationPairs}{identifier}{characters}

Use this command to declare valid pairs of punctuation marks. This will affect the punctuation commands discussed in \secref{aut:pct:pct}. For example, the description of \cmd{addcomma} states that this command adds a comma unless it is preceded by another comma, a semicolon, a colon, or a period. In other words, commas after abbreviation dots, exclamation marks, and question marks are permitted. These valid pairs are declared as follows:

\begin{ltxexample}
\DeclarePunctuationPairs{comma}{*!?}
\end{ltxexample}
%
The \prm{identifier} selects the command to be configured. The identifiers correspond to the names of the punctuation commands from \secref{aut:pct:pct} without the \cmd{add} prefix, \ie valid \prm{identifier} strings are \texttt{dot}, \texttt{comma}, \texttt{semicolon}, \texttt{colon}, \texttt{period}, \texttt{exclam}, \texttt{question}. The \prm{characters} argument is an undelimited list of punctuation marks. Valid \prm{characters} are comma, semicolon, colon, period, exclamation mark, question mark, and asterisk. A period in the \prm{characters} argument denotes an end-of"=sentence period, an asterisk the dot after an abbreviation. This is the default setup, which is automatically restored whenever \biblatex switches languages and corresponds to the behavior described in \secref{aut:pct:pct}:

\begin{ltxexample}
\DeclarePunctuationPairs{dot}{}
\DeclarePunctuationPairs{comma}{*!?}
\DeclarePunctuationPairs{semicolon}{*!?}
\DeclarePunctuationPairs{colon}{*!?}
\DeclarePunctuationPairs{period}{}
\DeclarePunctuationPairs{exclam}{*}
\DeclarePunctuationPairs{question}{*}
\end{ltxexample}
%
Since this feature is language specific, \cmd{DeclarePunctuationPairs} must be used in the argument to \cmd{DefineBibliographyExtras} (when used in the preamble) or \cmd{DeclareBibliographyExtras} (when used in a localisation module). See \secref{use:lng, aut:lng} for details. Note that some localisation modules may use a setup which is different from the package default.\footnote{As of this writing, the \texttt{american} module uses different settings for <American-style> punctuation.}

\cmditem{DeclareQuotePunctuation}{characters}

This command controls <American-style> punctuation. The \cmd{mkbibquote} wrapper from \secref{aut:fmt:ich} can interact with the punctuation facilities discussed in \secref{aut:pct:new, aut:pct:pct, aut:pct:spc}. Punctuation marks after \cmd{mkbibquote} will be moved inside the quotes if they have been registered with \cmd{DeclareQuotePunctuation}. Note that \prm{characters} is an undelimited list of characters. Valid \prm{characters} are period, comma, semicolon, colon, exclamation and question mark. Here is an example:

\begin{ltxexample}
\DeclareQuotePunctuation{.,}
\end{ltxexample}
%
Executing |\DeclareQuotePunctuation{}| is equivalent to disabling this feature. This is the package default. Since this feature is language specific, this command must be used in the argument to \cmd{DefineBibliographyExtras} (when used in the preamble) or \cmd{DeclareBibliographyExtras} (when used in a localisation module). See \secref{use:lng, aut:lng} for details. See also \secref{use:loc:us}.

\csitem{uspunctuation}

A shorthand using the lower-level commands \cmd{DeclareQuotePunctuation} and \cmd{DeclarePunctuationPairs} to activate <American-style> punctuation. See \secref{use:loc:us} for details. This shorthand is provided for convenience only. The effective settings are applied by the lower-level commands.

\csitem{stdpunctuation}

Undoes the settings applied by \cmd{uspunctuation}, restoring standard punctuation. As standard punctuation is the default setting, you only need this command to override a previously executed \cmd{uspunctuation} command. See \secref{use:loc:us} for details.

\end{ltxsyntax}

\subsubsection{Correcting Punctuation Tracking}
\label{aut:pct:ctr}

The facilities for punctuation tracking and automatic capitalization are very reliable under normal circumstances, but there are always marginal cases which may require manual intervention. Typical cases are localisation strings printed as the first word in a footnote (which is usually treated as the beginning of a paragraph as far as capitalization is concerned, but \tex is not in vertical mode at this point) or punctuation after periods which are not really end"=of"=sentence periods (for example, after an ellipsis like «[\dots\unkern]» a command such as \cmd{addperiod} would do nothing since parentheses and brackets are transparent to the punctuation tracker). In such cases, use the following commands in bibliography and citation styles to mark the beginning or middle of a sentence if and where required:

\begin{ltxsyntax}

\csitem{bibsentence}

This command marks the beginning of a sentence. A localisation string immediately after this command will be capitalized and the punctuation tracker is reset, \ie this command hides all preceding punctuation marks from the punctuation tracker and enforces capitalization.

\csitem{midsentence}

This command marks the middle of a sentence. A localisation string immediately after this command will not be capitalized and the punctuation tracker is reset, \ie this command hides all preceding punctuation marks from the punctuation tracker and suppresses capitalization.

\csitem*{midsentence*}

The starred variant of \cmd{midsentence} differs from the regular one in that a preceding abbreviation dot is not hidden from the punctuation tracker, \ie any code after \cmd{midsentence*} will see a preceding abbreviation dot. All other punctuation marks are hidden from the punctuation tracker and capitalization is suppressed.

\end{ltxsyntax}

\subsection{Localization Strings}
\label{aut:str}

Localization strings are key terms such as <edition> or <volume> which are automatically translated by \biblatex's localisation modules. See \secref{aut:lng} for an overview and \secref{aut:lng:key} for a list of all strings supported by default. The commands in this section are used to print the localised term.

\begin{ltxsyntax}

\cmditem{bibstring}[wrapper]{key}

Prints the localisation string \prm{key}, where \prm{key} is an identifier in lowercase letters (see \secref{aut:lng:key}). The string will be capitalized as required, see \secref{aut:pct:cfg} for details.
Depending on the \opt{abbreviate} package option from \secref{use:opt:pre:gen}, \cmd{bibstring} prints the short or the long version of the string. If localisation strings are nested, \ie if \cmd{bibstring} is used in another string, it will behave like \cmd{bibxstring}.
If the \prm{wrapper} argument is given, the string is passed to the \prm{wrapper} for formatting. This is intended for font commands such as \cmd{emph}.

\cmditem{biblstring}[wrapper]{key}

Similar to \cmd{bibstring} but always prints the long string, ignoring the \opt{abbreviate} option.

\cmditem{bibsstring}[wrapper]{key}

Similar to \cmd{bibstring} but always prints the short string, ignoring the \opt{abbreviate} option.

\cmditem{bibncpstring}[wrapper]{key}

Similar to \cmd{bibstring} but the term is never capitalized.

\cmditem{bibncplstring}[wrapper]{key}

Similar to \cmd{biblstring} but the term is never capitalized.

\cmditem{bibncpsstring}[wrapper]{key}

Similar to \cmd{bibsstring} but the term is never capitalized.

\cmditem{bibcpstring}[wrapper]{key}

Similar to \cmd{bibstring} but the term is always capitalized.

\cmditem{bibcplstring}[wrapper]{key}

Similar to \cmd{biblstring} but the term is always capitalized.

\cmditem{bibcpsstring}[wrapper]{key}

Similar to \cmd{bibsstring} but the term is always capitalized.

\cmditem{bibucstring}[wrapper]{key}

Similar to \cmd{bibstring} but the whole term is uppercased.

\cmditem{bibuclstring}[wrapper]{key}

Similar to \cmd{biblstring} but the whole term is uppercased.

\cmditem{bibucsstring}[wrapper]{key}

Similar to \cmd{bibsstring} but the whole term is uppercased.

\cmditem{biblcstring}[wrapper]{key}

Similar to \cmd{bibstring} but the whole term is lowercased.

\cmditem{biblclstring}[wrapper]{key}

Similar to \cmd{biblstring} but the whole term is lowercased.

\cmditem{biblcsstring}[wrapper]{key}

Similar to \cmd{bibsstring} but the whole term is lowercased.

\cmditem{bibxstring}{key}

A simplified but expandable version of \cmd{bibstring}. Note that this variant does not capitalize automatically, nor does it hook into the punctuation tracker. It is intended for special cases in which strings are nested or an expanded localisation string is required in a test.

\cmditem{bibxlstring}[wrapper]{key}

Similar to \cmd{bibxstring} but always uses the long string, ignoring the \opt{abbreviate} option.

\cmditem{bibxsstring}[wrapper]{key}

Similar to \cmd{bibxstring} but always uses the short string, ignoring the \opt{abbreviate} option.

\cmditem{mainlang}\DeprecatedMark

Switches from the current language to the main document language. This command is deprecated. Use the text-macro \cmd{textmainlang} instead. With \sty{babel} this command will need to be wrapped into \emph{two} groups to have purely local effect.

\cmditem{textmainlang}{text}

Locally switches from the current language to the main document language to typeset \prm{text}. This can be used the \prm{wrapper} argument in the localisation string commands above.

\end{ltxsyntax}

\subsection{Localization Modules}
\label{aut:lng}

A localisation module provides translations for key terms such as <edition> or <volume> as well as definitions for language specific features such as the date format and ordinals. These definitions are provided in files with the suffix \file{lbx}. The base name of the file must be a language name known to the \sty{babel}/\sty{polyglossia} packages. The \file{lbx} files may also be used to map \sty{babel}/\sty{polyglossia} language names to the backend modules of the \biblatex package. All localisation modules are loaded on demand in the document body. Note that the contents of the file are processed in a group and that the category code of the character \texttt{@} is temporarily set to <letter>.

\subsubsection{Localization Commands}
\label{aut:lng:cmd}

The user"=level versions of the localisation commands were already introduced in \secref{use:lng}. When used in \file{lbx} files, however, the syntax of localisation commands is different from the user syntax in the preamble and the configuration file. When used in localisation files, there is no need to specify the \prm{language} because the mapping of strings to a language is already provided by the name of the \file{lbx} file.

\begin{ltxsyntax}

\cmditem{DeclareBibliographyStrings}{definitions}

This command is only available in \file{lbx} files. It is used to define localisation strings. The \prm{definitions} consist of \keyval pairs which assign an expression to an identifier. A complete list of all keys supported by default is given is \secref{aut:lng:key}. Note that the syntax of the value is different in \file{lbx} files. The value assigned to a key consists of two expressions, each of which is wrapped in an additional pair of brackets. This is best shown by example:

\begin{ltxexample}
\DeclareBibliographyStrings{%
  bibliography  = {{Bibliography}{Bibliography}},
  shorthands    = {{List of Abbreviations}{Abbreviations}},
  editor        = {{editor}{ed.}},
  editors       = {{editors}{eds.}},
}
\end{ltxexample}
%
The first value is the long, written out expression, the second one is an abbreviated or short form. Both strings must always be given even though they may be identical if an expression is always (or never) abbreviated. Depending on the setting of the \opt{abbreviate} package option (see \secref{use:opt:pre:gen}), \biblatex selects one expression when loading the \file{lbx} file. There is also a special key named \texttt{inherit} which copies the strings from a different language. This is intended for languages which only differ in a few expressions, such as German and Austrian or American and British English. For example, here are the complete definitions for Austrian:

\begin{ltxexample}
\DeclareBibliographyStrings{%
  inherit       = {german},
  january       = {{J\"anner}{J\"an.}},
}
\end{ltxexample}

The above examples are slightly simplified. Real localisation files should use the punctuation and formatting commands discussed in \secref{aut:pct:pct, use:fmt} instead of literal punctuation. Here is an excerpt from a real localisation file:

\begin{ltxexample}
  bibliography     = {{Bibliography}{Bibliography}},
  shorthands       = {{List of Abbreviations}{Abbreviations}},
  editor           = {{editor}{ed\adddot}},
  editors          = {{editors}{eds\adddot}},
  byeditor         = {{edited by}{ed\adddotspace by}},
  mathesis         = {{Master's thesis}{MA\addabbrvspace thesis}},
\end{ltxexample}
%
Note the handling of abbreviation dots, the spacing in abbreviated expressions, and the capitalization in the example above. All expressions should be capitalized as they usually are when used in the middle of a sentence. The \biblatex package will automatically capitalize the first word when required at the beginning of a sentence, see \cmd{DeclareCapitalPunctuation} in \secref{aut:pct:cfg} for details. Expressions intended for use in headings are special. They should be capitalized in a way that is suitable for titling and should not be abbreviated (but they may have a short form).

\cmditem{InheritBibliographyStrings}{language}

This command is only available in \file{lbx} files. It copies the localisation strings for \prm{language} to the current language, as specified by the name of the \file{lbx} file.

\cmditem{DeclareBibliographyExtras}{code}

This command is only available in \file{lbx} files. It is used to adapt language specific features such as the date format and ordinals. The \prm{code}, which may be arbitrary \latex code, will usually consist of redefinitions of the formatting commands from \secref{aut:fmt:lng}.

\cmditem{UndeclareBibliographyExtras}{code}

This command is only available in \file{lbx} files. It is used to restore any formatting commands modified with \cmd{DeclareBibliographyExtras}. If a redefined command is included in \secref{aut:fmt:lng}, there is no need to restore its previous definition since these commands are localised by all language modules anyway.

\cmditem{InheritBibliographyExtras}{language}

This command is only available in \file{lbx} files. It copies the bibliography extras for \prm{language} to the current language, as specified by the name of the \file{lbx} file.

\cmditem{DeclareHyphenationExceptions}{text}

This command corresponds to \cmd{DefineHyphenationExceptions} from \secref{use:lng}. The difference is that it is only available in \file{lbx} files and that the \prm{language} argument is omitted. The hyphenation exceptions will affect the language of the \file{lbx} file currently being processed.

\cmditem{DeclareRedundantLanguages}{language, language, ...}{langid, langid, ...}

This command provides the language mappings required by the \opt{clearlang} option from \secref{use:opt:pre:gen}.
The \prm{language} is the string given in the \bibfield{language} field (without the optional \texttt{lang} prefix); \prm{langid} is \sty{babel}/\sty{polyglossia}'s language identifier, as given in the optional argument of \cmd{usepackage} when loading \sty{babel} or the argument of \cmd{setdefaultlanguage} or \cmd{setotherlanguages} when using \sty{polyglossia}. This command may be used in \file{lbx} files or in the document preamble. Here are some examples:

\begin{ltxexample}
\DeclareRedundantLanguages{french}{french}
\DeclareRedundantLanguages{german}{german,ngerman,austrian,naustrian,
        nswissgerman,swissgerman}
\DeclareRedundantLanguages{english,american}{english,american,british,
	canadian,australian,newzealand,USenglish,UKenglish}
\end{ltxexample}
%
Note that this feature needs to be enabled globally with the \opt{clearlang} option from \secref{use:opt:pre:gen}. If it is disabled, all mappings will be ignored. If the \prm{langid} parameter is blank, \biblatex will clear the mappings for the corresponding \prm{language}, \ie the feature will be disabled for this \prm{language} only.

\cmditem{DeclareLanguageMapping}{language}{file}

This command maps a \sty{babel}/\sty{polyglossia} language identifier to an \file{lbx} file. The \prm{language} must be a language name known to the \sty{babel}/\sty{polyglossia} package, \ie one of the identifiers listed in \tabref{bib:fld:tab1}. The \prm{file} argument is the name of an alternative \file{lbx} file without the \texttt{.lbx} suffix. Declaring the same mapping more than once is possible. Subsequent declarations will simply overwrite any previous ones. This command may only be used in the preamble. See \secref{aut:cav:lng} for further details.

\cmditem{DeclareLanguageMappingSuffix}{suffix}

This command defines a language file suffix which will be added when looking for \file{.lbx} language string definition files. This is intended for styles which provide their own \file{.lbx} files so that they will be used automatically. For example, the APA style defines:

\begin{ltxexample}
\DeclareLanguageMappingSuffix{-apa}
\end{ltxexample}
%
When the document language is <german>, \biblatex will look for the file \file{german-apa.lbx} which defines some APA specific strings and in turn loads \file{german.lbx}. If \cmd{DeclareLanguageMapping} is defined for a language, this overrides \cmd{DeclareLanguageMappingSuffix}.

The suffix will be applied to other language files loaded recursively by the loading of a language file. For example, given the suffix defined above, when loading <ngerman>, \biblatex will look for the file \file{ngerman-apa.lbx} and if this recursively loads <german>, then biblatex will look for \file{german-apa.lbx}. Infinite recursion is of course avoided.

\cmditem{NewBibliographyString}{key}

This command, which may be used in the preamble (including \file{cbx} and \file{bbx} files) as well as in \file{lbx} files, declares new localisation strings, \ie it initializes a new \prm{key} to be used in the \prm{definitions} of \cmd{DefineBibliographyStrings} or \cmd{DeclareBibliographyStrings}. The \prm{key} argument may also be a comma"=separated list of key names. When used in an \file{lbx}, the \prm{key} is initialized only for the language specified by the name of the \file{lbx} file. The keys listed in \secref{aut:lng:key} are defined by default.

\end{ltxsyntax}

\subsubsection{Localization Keys}
\label{aut:lng:key}

The localisation keys in this section are defined by default and covered by the localisation files which come with \biblatex. Note that these strings are only available in citations, the bibliography and bibliography lists. All expressions should be capitalized as they usually are when used in the middle of a sentence. \biblatex will capitalize them automatically at the beginning of a sentence. The only exceptions to these rules are the three strings intended for use in headings.

\paragraph{Headings}
\label{aut:lng:key:bhd}

The following strings are special because they are intended for use in headings and made available globally via macros. For this reason, they should be capitalized for use in headings and they must not include any local commands which are part of \biblatex's author interface.

\begin{keymarglist}
\item[bibliography] The term <bibliography>, also available as \cmd{bibname}.
\item[references] The term <references>, also available as \cmd{refname}.
\item[shorthands] The term <list of shorthands> or <list of abbreviations>, also available as \cmd{biblistname}.
\end{keymarglist}

\paragraph{Roles, Expressed as Functions}
\label{aut:lng:key:efn}

The following keys refer to roles which are expressed as a function (<editor>, <translator>) rather than as an action (<edited by>, <translated by>).

\begin{keymarglist}
\item[editor] The term <editor>, referring to the main editor. This is the most generic editorial role.
\item[editors] The plural form of \texttt{editor}.
\item[compiler] The term <compiler>, referring to an editor whose task is to compile a work.
\item[compilers] The plural form of \texttt{compiler}.
\item[founder] The term <founder>, referring to a founding editor.
\item[founders] The plural form of \texttt{founder}.
\item[continuator] An expression like <continuator>, <continuation>, or <continued>, referring to a past editor who continued the work of the founding editor but was subsequently replaced by the current editor.
\item[continuators] The plural form of \texttt{continuator}.
\item[redactor] The term <redactor>, referring to a secondary editor.
\item[redactors] The plural form of \texttt{redactor}.
\item[reviser] The term <reviser>, referring to a secondary editor.
\item[revisers] The plural form of \texttt{reviser}.
\item[collaborator] A term like <collaborator>, <collaboration>, <cooperator>, or <cooperation>, referring to a secondary editor.
\item[collaborators] The plural form of \texttt{collaborator}.
\item[translator] The term <translator>.
\item[translators] The plural form of \texttt{translator}.
\item[commentator] The term <commentator>, referring to the author of a commentary to a work.
\item[commentators] The plural form of \texttt{commentators}.
\item[annotator] The term <annotator>, referring to the author of annotations to a work.
\item[annotators] The plural form of \texttt{annotators}.
\item[organizer] The term <organizer>, referring to the organizer of an
  event or work.
\item[organizers] The plural form of \texttt{organizer}.
\end{keymarglist}

\paragraph{Concatenated Editor Roles, Expressed as Functions}
\label{aut:lng:key:cef}

The following keys are similar in function to \texttt{editor}, \texttt{translator}, etc. They are used to indicate additional roles of the editor, \eg\ <editor and translator>, <editor and foreword>.

\begin{keymarglist}
\item[editortr] Used if \bibfield{editor}\slash \bibfield{translator} are identical.
\item[editorstr] The plural form of \texttt{editortr}.
\item[editorco] Used if \bibfield{editor}\slash \bibfield{commentator} are identical.
\item[editorsco] The plural form of \texttt{editorco}.
\item[editoran] Used if \bibfield{editor}\slash \bibfield{annotator} are identical.
\item[editorsan] The plural form of \texttt{editoran}.
\item[editorin] Used if \bibfield{editor}\slash \bibfield{introduction} are identical.
\item[editorsin] The plural form of \texttt{editorin}.
\item[editorfo] Used if \bibfield{editor}\slash \bibfield{foreword} are identical.
\item[editorsfo] The plural form of \texttt{editorfo}.
\item[editoraf] Used if \bibfield{editor}\slash \bibfield{aftword} are identical.
\item[editorsaf] The plural form of \texttt{editoraf}.
\end{keymarglist}
%
Keys for \bibfield{editor}\slash \bibfield{translator}\slash \prm{role} combinations:

\begin{keymarglist}
\item[editortrco] Used if \bibfield{editor}\slash \bibfield{translator}\slash \bibfield{commentator} are identical.
\item[editorstrco] The plural form of \texttt{editortrco}.
\item[editortran] Used if \bibfield{editor}\slash \bibfield{translator}\slash \bibfield{annotator} are identical.
\item[editorstran] The plural form of \texttt{editortran}.
\item[editortrin] Used if \bibfield{editor}\slash \bibfield{translator}\slash \bibfield{introduction} are identical.
\item[editorstrin] The plural form of \texttt{editortrin}.
\item[editortrfo] Used if \bibfield{editor}\slash \bibfield{translator}\slash \bibfield{foreword} are identical.
\item[editorstrfo] The plural form of \texttt{editortrfo}.
\item[editortraf] Used if \bibfield{editor}\slash \bibfield{translator}\slash \bibfield{aftword} are identical.
\item[editorstraf] The plural form of \texttt{editortraf}.
\end{keymarglist}
%
Keys for \bibfield{editor}\slash \bibfield{commentator}\slash \prm{role} combinations:

\begin{keymarglist}
\item[editorcoin] Used if \bibfield{editor}\slash \bibfield{commentator}\slash \bibfield{introduction} are identical.
\item[editorscoin] The plural form of \texttt{editorcoin}.
\item[editorcofo] Used if \bibfield{editor}\slash \bibfield{commentator}\slash \bibfield{foreword} are identical.
\item[editorscofo] The plural form of \texttt{editorcofo}.
\item[editorcoaf] Used if \bibfield{editor}\slash \bibfield{commentator}\slash \bibfield{aftword} are identical.
\item[editorscoaf] The plural form of \texttt{editorcoaf}.
\end{keymarglist}
%
Keys for \bibfield{editor}\slash \bibfield{annotator}\slash \prm{role} combinations:

\begin{keymarglist}
\item[editoranin] Used if \bibfield{editor}\slash \bibfield{annotator}\slash \bibfield{introduction} are identical.
\item[editorsanin] The plural form of \texttt{editoranin}.
\item[editoranfo] Used if \bibfield{editor}\slash \bibfield{annotator}\slash \bibfield{foreword} are identical.
\item[editorsanfo] The plural form of \texttt{editoranfo}.
\item[editoranaf] Used if \bibfield{editor}\slash \bibfield{annotator}\slash \bibfield{aftword} are identical.
\item[editorsanaf] The plural form of \texttt{editoranaf}.
\end{keymarglist}
%
Keys for \bibfield{editor}\slash \bibfield{translator}\slash \bibfield{commentator}\slash \prm{role} combinations:

\begin{keymarglist}
\item[editortrcoin] Used if \bibfield{editor}\slash \bibfield{translator}\slash \bibfield{commentator}\slash \bibfield{introduction} are identical.
\item[editorstrcoin] The plural form of \texttt{editortrcoin}.
\item[editortrcofo] Used if \bibfield{editor}\slash \bibfield{translator}\slash \bibfield{commentator}\slash \bibfield{foreword} are identical.
\item[editorstrcofo] The plural form of \texttt{editortrcofo}.
\item[editortrcoaf] Used if \bibfield{editor}\slash \bibfield{translator}\slash \bibfield{commentator}\slash \bibfield{aftword} are identical.
\item[editorstrcoaf] The plural form of \texttt{editortrcoaf}.
\end{keymarglist}
%
Keys for \bibfield{editor}\slash \bibfield{annotator}\slash \bibfield{commentator}\slash \prm{role} combinations:

\begin{keymarglist}
\item[editortranin] Used if \bibfield{editor}\slash \bibfield{annotator}\slash \bibfield{commentator}\slash \bibfield{introduction} are identical.
\item[editorstranin] The plural form of \texttt{editortranin}.
\item[editortranfo] Used if \bibfield{editor}\slash \bibfield{annotator}\slash \bibfield{commentator}\slash \bibfield{foreword} are identical.
\item[editorstranfo] The plural form of \texttt{editortranfo}.
\item[editortranaf] Used if \bibfield{editor}\slash \bibfield{annotator}\slash \bibfield{commentator}\slash \bibfield{aftword} are identical.
\item[editorstranaf] The plural form of \texttt{editortranaf}.
\end{keymarglist}

\paragraph{Concatenated Translator Roles, Expressed as Functions}
\label{aut:lng:key:ctf}

The following keys are similar in function to \texttt{translator}. They are used to indicate additional roles of the translator, \eg\ <translator and commentator>, <translator and introduction>.

\begin{keymarglist}
\item[translatorco] Used if \bibfield{translator}\slash \bibfield{commentator} are identical.
\item[translatorsco] The plural form of \texttt{translatorco}.
\item[translatoran] Used if \bibfield{translator}\slash \bibfield{annotator} are identical.
\item[translatorsan] The plural form of \texttt{translatoran}.
\item[translatorin] Used if \bibfield{translator}\slash \bibfield{introduction} are identical.
\item[translatorsin] The plural form of \texttt{translatorin}.
\item[translatorfo] Used if \bibfield{translator}\slash \bibfield{foreword} are identical.
\item[translatorsfo] The plural form of \texttt{translatorfo}.
\item[translatoraf] Used if \bibfield{translator}\slash \bibfield{aftword} are identical.
\item[translatorsaf] The plural form of \texttt{translatoraf}.
\end{keymarglist}
%
Keys for \bibfield{translator}\slash \bibfield{commentator}\slash \prm{role} combinations:

\begin{keymarglist}
\item[translatorcoin] Used if \bibfield{translator}\slash \bibfield{commentator}\slash \bibfield{introduction} are identical.
\item[translatorscoin] The plural form of \texttt{translatorcoin}.
\item[translatorcofo] Used if \bibfield{translator}\slash \bibfield{commentator}\slash \bibfield{foreword} are identical.
\item[translatorscofo] The plural form of \texttt{translatorcofo}.
\item[translatorcoaf] Used if \bibfield{translator}\slash \bibfield{commentator}\slash \bibfield{aftword} are identical.
\item[translatorscoaf] The plural form of \texttt{translatorcoaf}.
\end{keymarglist}
%
Keys for \bibfield{translator}\slash \bibfield{annotator}\slash \prm{role} combinations:

\begin{keymarglist}
\item[translatoranin] Used if \bibfield{translator}\slash \bibfield{annotator}\slash \bibfield{introduction} are identical.
\item[translatorsanin] The plural form of \texttt{translatoranin}.
\item[translatoranfo] Used if \bibfield{translator}\slash \bibfield{annotator}\slash \bibfield{foreword} are identical.
\item[translatorsanfo] The plural form of \texttt{translatoranfo}.
\item[translatoranaf] Used if \bibfield{translator}\slash \bibfield{annotator}\slash \bibfield{aftword} are identical.
\item[translatorsanaf] The plural form of \texttt{translatoranaf}.
\end{keymarglist}

\paragraph{Roles, Expressed as Actions}
\label{aut:lng:key:eac}

The following keys refer to roles which are expressed as an action (<edited by>, <translated by>) rather than as a function (<editor>, <translator>).

\begin{keymarglist}
\item[byauthor] The expression <[created] by \prm{name}>.
\item[byeditor] The expression <edited by \prm{name}>.
\item[bycompiler] The expression <compiled by \prm{name}>.
\item[byfounder] The expression <founded by \prm{name}>.
\item[bycontinuator] The expression <continued by \prm{name}>.
\item[byredactor] The expression <redacted by \prm{name}>.
\item[byreviser] The expression <revised by \prm{name}>.
\item[byreviewer] The expression <reviewed by \prm{name}>.
\item[bycollaborator] An expression like <in collaboration with \prm{name}> or <in cooperation with \prm{name}>.
\item[bytranslator] The expression <translated by \prm{name}> or <translated from \prm{language} by \prm{name}>.
\item[bycommentator] The expression <commented by \prm{name}>.
\item[byannotator] The expression <annotated by \prm{name}>.
\item[byorganizer] The expression <[organized] by \prm{name}>.
\end{keymarglist}

\paragraph{Concatenated Editor Roles, Expressed as Actions}
\label{aut:lng:key:cea}

The following keys are similar in function to \texttt{byeditor}, \texttt{bytranslator}, etc. They are used to indicate additional roles of the editor, \eg\ <edited and translated by>, <edited and furnished with an introduction by>, <edited, with a foreword, by>.

\begin{keymarglist}
\item[byeditortr] Used if \bibfield{editor}\slash \bibfield{translator} are identical.
\item[byeditorco] Used if \bibfield{editor}\slash \bibfield{commentator} are identical.
\item[byeditoran] Used if \bibfield{editor}\slash \bibfield{annotator} are identical.
\item[byeditorin] Used if \bibfield{editor}\slash \bibfield{introduction} are identical.
\item[byeditorfo] Used if \bibfield{editor}\slash \bibfield{foreword} are identical.
\item[byeditoraf] Used if \bibfield{editor}\slash \bibfield{aftword} are identical.
\end{keymarglist}
%
Keys for \bibfield{editor}\slash \bibfield{translator}\slash \prm{role} combinations:

\begin{keymarglist}
\item[byeditortrco] Used if \bibfield{editor}\slash \bibfield{translator}\slash \bibfield{commentator} are identical.
\item[byeditortran] Used if \bibfield{editor}\slash \bibfield{translator}\slash \bibfield{annotator} are identical.
\item[byeditortrin] Used if \bibfield{editor}\slash \bibfield{translator}\slash \bibfield{introduction} are identical.
\item[byeditortrfo] Used if \bibfield{editor}\slash \bibfield{translator}\slash \bibfield{foreword} are identical.
\item[byeditortraf] Used if \bibfield{editor}\slash \bibfield{translator}\slash \bibfield{aftword} are identical.
\end{keymarglist}
%
Keys for \bibfield{editor}\slash \bibfield{commentator}\slash \prm{role} combinations:

\begin{keymarglist}
\item[byeditorcoin] Used if \bibfield{editor}\slash \bibfield{commentator}\slash \bibfield{introduction} are identical.
\item[byeditorcofo] Used if \bibfield{editor}\slash \bibfield{commentator}\slash \bibfield{foreword} are identical.
\item[byeditorcoaf] Used if \bibfield{editor}\slash \bibfield{commentator}\slash \bibfield{aftword} are identical.
\end{keymarglist}
%
Keys for \bibfield{editor}\slash \bibfield{annotator}\slash \prm{role} combinations:

\begin{keymarglist}
\item[byeditoranin] Used if \bibfield{editor}\slash \bibfield{annotator}\slash \bibfield{introduction} are identical.
\item[byeditoranfo] Used if \bibfield{editor}\slash \bibfield{annotator}\slash \bibfield{foreword} are identical.
\item[byeditoranaf] Used if \bibfield{editor}\slash \bibfield{annotator}\slash \bibfield{aftword} are identical.
\end{keymarglist}
%
Keys for \bibfield{editor}\slash \bibfield{translator}\slash \bibfield{commentator}\slash \prm{role} combinations:

\begin{keymarglist}
\item[byeditortrcoin] Used if \bibfield{editor}\slash \bibfield{translator}\slash \bibfield{commentator}\slash \bibfield{introduction} are identical.
\item[byeditortrcofo] Used if \bibfield{editor}\slash \bibfield{translator}\slash \bibfield{commentator}\slash \bibfield{foreword} are identical.
\item[byeditortrcoaf] Used if \bibfield{editor}\slash \bibfield{translator}\slash \bibfield{commentator}\slash \bibfield{aftword} are identical.
\end{keymarglist}
%
Keys for \bibfield{editor}\slash \bibfield{translator}\slash \bibfield{annotator}\slash \prm{role} combinations:

\begin{keymarglist}
\item[byeditortranin] Used if \bibfield{editor}\slash \bibfield{annotator}\slash \bibfield{commentator}\slash \bibfield{introduction} are identical.
\item[byeditortranfo] Used if \bibfield{editor}\slash \bibfield{annotator}\slash \bibfield{commentator}\slash \bibfield{foreword} are identical.
\item[byeditortranaf] Used if \bibfield{editor}\slash \bibfield{annotator}\slash \bibfield{commentator}\slash \bibfield{aftword} are identical.
\end{keymarglist}

\paragraph{Concatenated Translator Roles, Expressed as Actions}
\label{aut:lng:key:cta}

The following keys are similar in function to \texttt{bytranslator}. They are used to indicate additional roles of the translator, \eg\ <translated and commented by>, <translated and furnished with an introduction by>, <translated, with a foreword, by>.

\begin{keymarglist}
\item[bytranslatorco] Used if \bibfield{translator}\slash \bibfield{commentator} are identical.
\item[bytranslatoran] Used if \bibfield{translator}\slash \bibfield{annotator} are identical.
\item[bytranslatorin] Used if \bibfield{translator}\slash \bibfield{introduction} are identical.
\item[bytranslatorfo] Used if \bibfield{translator}\slash \bibfield{foreword} are identical.
\item[bytranslatoraf] Used if \bibfield{translator}\slash \bibfield{aftword} are identical.
\end{keymarglist}
%
Keys for \bibfield{translator}\slash \bibfield{commentator}\slash \prm{role} combinations:

\begin{keymarglist}
\item[bytranslatorcoin] Used if \bibfield{translator}\slash \bibfield{commentator}\slash \bibfield{introduction} are identical.
\item[bytranslatorcofo] Used if \bibfield{translator}\slash \bibfield{commentator}\slash \bibfield{foreword} are identical.
\item[bytranslatorcoaf] Used if \bibfield{translator}\slash \bibfield{commentator}\slash \bibfield{aftword} are identical.
\end{keymarglist}
%
Keys for \bibfield{translator}\slash \bibfield{annotator}\slash \prm{role} combinations:

\begin{keymarglist}
\item[bytranslatoranin] Used if \bibfield{translator}\slash \bibfield{annotator}\slash \bibfield{introduction} are identical.
\item[bytranslatoranfo] Used if \bibfield{translator}\slash \bibfield{annotator}\slash \bibfield{foreword} are identical.
\item[bytranslatoranaf] Used if \bibfield{translator}\slash \bibfield{annotator}\slash \bibfield{aftword} are identical.
\end{keymarglist}

\paragraph{Roles, Expressed as Objects}
\label{aut:lng:key:rob}

Roles which are related to supplementary material may also be expressed as objects (<with a commentary by>) rather than as functions (<commentator>) or as actions (<commented by>).

\begin{keymarglist}
\item[withcommentator] The expression <with a commentary by \prm{name}>.
\item[withannotator] The expression <with annotations by \prm{name}>.
\item[withintroduction] The expression <with an introduction by \prm{name}>.
\item[withforeword] The expression <with a foreword by \prm{name}>.
\item[withafterword] The expression <with an afterword by \prm{name}>.
\end{keymarglist}

\paragraph{Supplementary Material}
\label{aut:lng:key:mat}

\begin{keymarglist}
\item[commentary] The term <commentary>.
\item[annotations] The term <annotations>.
\item[introduction] The term <introduction>.
\item[foreword] The term <foreword>.
\item[afterword] The term <afterword>.
\end{keymarglist}

\paragraph{Publication Details}
\label{aut:lng:key:pdt}

\begin{keymarglist}
\item[volume] The term <volume>, referring to a book.
\item[volumes] The plural form of \texttt{volume}.
\item[involumes] The term <in>, as used in expressions like <in \prm{number of volumes} volumes>.
\item[jourvol] The term <volume>, referring to a journal.
\item[jourser] The term <series>, referring to a journal.
\item[book] The term <book>, referring to a document division.
\item[part] The term <part>, referring to a part of a book or a periodical.
\item[issue] The term <issue>, referring to a periodical.
\item[newseries] The expression <new series>, referring to a journal.
\item[oldseries] The expression <old series>, referring to a journal.
\item[edition] The term <edition>.
\item[in] The term <in>, referring to the title of a work published as part of another one, \eg\ <\prm{title of article} in \prm{title of journal}>.
\item[inseries] The term <in>, as used in expressions like <volume \prm{number} in \prm{name of series}>.
\item[ofseries] The term <of>, as used in expressions like <volume \prm{number} of \prm{name of series}>.
\item[number] The term <number>, referring to an issue of a journal.
\item[chapter] The term <chapter>, referring to a chapter in a book.
\item[version] The term <version>, referring to a revision number.
\item[reprint] The term <reprint>.
\item[reprintof] The expression <reprint of \prm{title}>.
\item[reprintas] The expression <reprinted as \prm{title}>.
\item[reprintfrom] The expression <reprinted from \prm{title}>.
\item[translationof] The expression <translation of \prm{title}>.
\item[translationas] The expression <translated as \prm{title}>.
\item[translationfrom] The expression <translated from [the] \prm{language}>.
\item[reviewof] The expression <review of \prm{title}>.
\item[origpubas] The expression <originally published as \prm{title}>.
\item[origpubin] The expression <originally published in \prm{year}>.
\item[astitle] The term <as>, as used in expressions like <published by \prm{publisher} as \prm{title}>.
\item[bypublisher] The term <by>, as used in expressions like <published by \prm{publisher}>.
\end{keymarglist}

\paragraph{Publication State}
\label{aut:lng:key:pst}

\begin{keymarglist}
\item[inpreparation] The expression <in preparation> (the manuscript is being prepared for publication).
\item[submitted] The expression <submitted> (the manuscript has been submitted to a journal or conference).
\item[forthcoming] The expression <forthcoming> (the manuscript has been accepted by a press or journal).
\item[inpress] The expression <in press> (the manuscript is fully copyedited and out of the author's hands; it is in the final stages of the production process).
\item[prepublished] The expression <pre-published> (the manuscript is published in a preliminary form or location, such as online version in advance of print publication).
\end{keymarglist}

\paragraph{Pagination}
\label{aut:lng:key:pag}

\begin{keymarglist}
\item[page] The term <page>.
\item[pages] The plural form of \texttt{page}.
\item[column] The term <column>, referring to a column on a page.
\item[columns] The plural form of \texttt{column}.
\item[section] The term <section>, referring to a document division (usually abbreviated as \S).
\item[sections] The plural form of \texttt{section} (usually abbreviated as \S\S).
\item[paragraph] The term <paragraph> (\ie a block of text, not to be confused with \texttt{section}).
\item[paragraphs] The plural form of \texttt{paragraph}.
\item[verse] The term <verse> as used when referring to a work which is cited by verse numbers.
\item[verses] The plural form of \texttt{verse}.
\item[line] The term <line> as used when referring to a work which is cited by line numbers.
\item[lines] The plural form of \texttt{line}.
\item[pagetotal] The term <page> as used in \cmd{mkpageprefix}.
\item[pagetotals] The plural form of \texttt{pagetotal}.
\item[columntotal] The term <column>, referring to a column on a page, as used in \cmd{mkpageprefix}.
\item[columntotals] The plural form of \texttt{columntotal}.
\item[sectiontotal] The term <section>, referring to a document division (usually abbreviated as \S),  as used in \cmd{mkpageprefix}.
\item[sectiontotals] The plural form of \texttt{sectiontotal} (usually abbreviated as \S\S).
\item[paragraphtotal] The term <paragraph> (\ie a block of text, not to be confused with \texttt{section}) as used in \cmd{mkpageprefix}.
\item[paragraphtotals] The plural form of \texttt{paragraphtotal}.
\item[versetotal] The term <verse> as used when referring to a work which is cited by verse numbers when used in \cmd{mkpageprefix}.
\item[versetotals] The plural form of \texttt{versetotal}.
\item[linetotal] The term <line> as used when referring to a work which is cited by line numbers when used in \cmd{mkpageprefix}.
\item[linetotals] The plural form of \texttt{linetotal}.
\end{keymarglist}

\paragraph{Types}
\label{aut:lng:key:typ}

The following keys are typically used in the \bibfield{type} field of \bibtype{thesis}, \bibtype{report}, \bibtype{misc}, and other entries:

\begin{keymarglist}
\item[bathesis] An expression equivalent to the term <Bachelor's thesis>.
\item[mathesis] An expression equivalent to the term <Master's thesis>.
\item[phdthesis] The term <PhD thesis>, <PhD dissertation>, <doctoral thesis>, etc.
\item[candthesis] An expression equivalent to the term <Candidate thesis>. Used for <Candidate> degrees that have no clear equivalent to the Master's or doctoral level.
\item[techreport] The term <technical report>.
\item[resreport] The term <research report>.
\item[software] The term <computer software>.
\item[datacd] The term <data \textsc{cd}> or <\textsc{cd-rom}>.
\item[audiocd] The term <audio \textsc{cd}>.
\end{keymarglist}

\paragraph{Miscellaneous}
\label{aut:lng:key:msc}

\begin{keymarglist}
\item[nodate] The term to use in place of a date when there is no date for an entry \eg\ <n.d.>
\item[and] The term <and>, as used in a list of authors or editors, for example.
\item[andothers] The expression <and others> or <et alii>, used to mark the truncation of a name list.
\item[andmore] Like \texttt{andothers} but used to mark the truncation of a literal list.
\end{keymarglist}

\paragraph{Labels}
\label{aut:lng:key:lab}

The following strings are intended for use as labels, \eg\ <Address: \prm{url}> or <Abstract: \prm{abstract}>.

\begin{keymarglist}
\item[url] The term <address> in the sense of an internet address.
\item[urlfrom] An expression like <available from \prm{url}> or <available at \prm{url}>.
\item[urlseen] An expression like <accessed on \prm{date}>, <retrieved on \prm{date}>, <visited on \prm{date}>, referring to the access date of an online resource.
\item[file] The term <file>.
\item[library] The term <library>.
\item[abstract] The term <abstract>.
\item[annotation] The term <annotations>.
\end{keymarglist}

\paragraph{Citations}
\label{aut:lng:key:cit}

Traditional scholarly expressions used in citations:

\begin{keymarglist}
\item[idem] The term equivalent to the Latin <idem> (<the same [person]>).
\item[idemsf] The feminine singular form of \texttt{idem}.
\item[idemsm] The masculine singular form of \texttt{idem}.
\item[idemsn] The neuter singular form of \texttt{idem}.
\item[idempf] The feminine plural form of \texttt{idem}.
\item[idempm] The masculine plural form of \texttt{idem}.
\item[idempn] The neuter plural form of \texttt{idem}.
\item[idempp] The plural form of \texttt{idem} suitable for a mixed gender list of names.
\item[ibidem] The term equivalent to the Latin <ibidem> (<in the same place>).
\item[opcit] The term equivalent to the Latin term <opere citato> (<[in] the work [already] cited>).
\item[loccit] The term equivalent to the Latin term <loco citato> (<[at] the place [already] cited>).
\item[confer] The term equivalent to the Latin <confer> (<compare>).
\item[sequens] The term equivalent to the Latin <sequens> (<[and] the following [page]>), as used to indicate a range of two pages when only the starting page is provided (\eg\ <25\,sq.> or <25\,f.> instead of <25--26>).
\item[sequentes] The term equivalent to the Latin <sequentes> (<[and] the following [pages]>), as used to indicate an open"=ended range of pages when only the starting page is provided (\eg\ <25\,sqq.> or <25\,ff.>).
\item[passim] The term equivalent to the Latin <passim> (<throughout>, <here and there>, <scatteredly>).
\end{keymarglist}
%
Other expressions frequently used in citations:

\begin{keymarglist}
\item[see] The term <see>.
\item[seealso] The expression <see also>.
\item[seenote] An expression like <see note \prm{footnote}> or <as in \prm{footnote}>, used to refer to a previous footnote in a citation.
\item[backrefpage] An expression like <see page \prm{page}> or <cited on page \prm{page}>, used to introduce back references in the bibliography.
\item[backrefpages] The plural form of \texttt{backrefpage}, \eg\ <see pages \prm{pages}> or <cited on pages \prm{pages}>.
\item[quotedin] An expression like <quoted in \prm{citation}>, used when quoting a passage which was already a quotation in the cited work.
\item[citedas] An expression like <henceforth cited as \prm{shorthand}>, used to introduce a shorthand in a citation.
\item[thiscite] The expression used in some verbose citation styles to differentiate between the page range of the cited item (typically an article in a journal, collection, or conference proceedings) and the page number the citation refers to. For example: \enquote{Author, Title, in: Book, pp. 45--61, \texttt{thiscite} p. 52.}
\end{keymarglist}

\paragraph{Month Names}
\label{aut:lng:key:mon}

\begin{keymarglist}
\item[january] The name <January>.
\item[february] The name <February>.
\item[march] The name <March>.
\item[april] The name <April>.
\item[may] The name <May>.
\item[june] The name <June>.
\item[july] The name <July>.
\item[august] The name <August>.
\item[september] The name <September>.
\item[october] The name <October>.
\item[november] The name <November>.
\item[december] The name <December>.
\end{keymarglist}

\paragraph{Language Names}
\label{aut:lng:key:lng}

\begin{keymarglist}
\item[langamerican] The language <American> or <American English>.
\item[langbasque] The language <Basque>.
\item[langbrazilian] The language <Brazilian> or <Brazilian Portuguese>.
\item[langbulgarian] The language <Bulgarian>.
\item[langcatalan] The language <Catalan>.
\item[langcroatian] The language <Croatian>.
\item[langczech] The language <Czech>.
\item[langdanish] The language <Danish>.
\item[langdutch] The language <Dutch>.
\item[langenglish] The language <English>.
\item[langestonian] The language <Estonian>.
\item[langfinnish] The language <Finnish>.
\item[langfrench] The language <French>.
\item[langgerman] The language <German>.
\item[langgreek] The language <Greek>.
\item[langhungarian] The language <Hungarian>.
\item[langitalian] The language <Italian>.
\item[langjapanese] The language <Japanese>.
\item[langlatin] The language <Latin>.
\item[langlatvian] The language <Latvian>.
\item[langlithuanian] The language <Lithuanian>.
\item[langnorwegian] The language <Norwegian>.
\item[langpolish] The language <Polish>.
\item[langportuguese] The language <Portuguese>.
\item[langrussian] The language <Russian>.
\item[langserbian] The language <Serbian>.
\item[langslovak] The language <Slovak>.
\item[langslovene] The language <Slovene>.
\item[langspanish] The language <Spanish>.
\item[langswedish] The language <Swedish>.
\item[langturkish] The language <Turkish>.
\item[langukrainian] The language <Ukrainian>.
\end{keymarglist}
%
The following strings are intended for use in phrases like <translated from [the] English by \prm{translator}>:

\begin{keymarglist}
\item[fromamerican] The expression <from [the] American> or <from [the] American English>.
\item[frombasque] The expression <from [the] Basque>.
\item[frombrazilian] The expression <from [the] Brazilian> or <from [the] Brazilian Portuguese>.
\item[frombulgarian] The expression <from [the] Bulgarian>.
\item[fromcatalan] The expression <from [the] Catalan>.
\item[fromcroatian] The expression <from [the] Croatian>.
\item[fromczech] The expression <from [the] Czech>.
\item[fromdanish] The expression <from [the] Danish>.
\item[fromdutch] The expression <from [the] Dutch>.
\item[fromenglish] The expression <from [the] English>.
\item[fromestonian] The expression <from [the] Estonian>.
\item[fromfinnish] The expression <from [the] Finnish>.
\item[fromfrench] The expression <from [the] French>.
\item[fromgerman] The expression <from [the] German>.
\item[fromgreek] The expression <from [the] Greek>.
\item[fromhungarian] The language <from [the] Hungarian>.
\item[fromitalian] The expression <from [the] Italian>.
\item[fromjapanese] The expression <from [the] Japanese>.
\item[fromlatin] The expression <from [the] Latin>.
\item[fromlatvian] The expression <from [the] Latvian>.
\item[fromlithuanian] The language <from [the] Lithuanian>.
\item[fromnorwegian] The expression <from [the] Norwegian>.
\item[frompolish] The expression <from [the] Polish>.
\item[fromportuguese] The expression <from [the] Portuguese>.
\item[fromrussian] The expression <from [the] Russian>.
\item[fromserbian] The expression <from [the] Serbian>.
\item[fromslovak] The expression <from [the] Slovak>.
\item[fromslovene] The expression <from [the] Slovene>.
\item[fromspanish] The expression <from [the] Spanish>.
\item[fromswedish] The expression <from [the] Swedish>.
\item[fromturkish] The expression <from [the] Turkish>.
\item[fromukrainian] The expression <from [the] Ukrainian>.
\end{keymarglist}

\paragraph{Country Names}
\label{aut:lng:key:cnt}

Country names are localised by using the string \texttt{country} plus the \acr{ISO}-3166 country code as the key. The short version of the translation should be the \acr{ISO}-3166 country code. Note that only a small number of country names is defined by default, mainly to illustrate this scheme. These keys are used in the \bibfield{location} list of \bibtype{patent} entries but they may be useful for other purposes as well.

\begin{keymarglist}
\item[countryde] The name <Germany>, abbreviated as \texttt{DE}.
\item[countryeu] The name <European Union>, abbreviated as \texttt{EU}.
\item[countryep] Similar to \opt{countryeu} but abbreviated as \texttt{EP}. This is intended for \bibfield{patent} entries.
\item[countryfr] The name <France>, abbreviated as \texttt{FR}.
\item[countryuk] The name <United Kingdom>, abbreviated (according to \acr{ISO}-3166) as \texttt{GB}.
\item[countryus] The name <United States of America>, abbreviated as \texttt{US}.
\end{keymarglist}

\paragraph{Patents and Patent Requests}
\label{aut:lng:key:pat}

Strings related to patents are localised by using the term \texttt{patent} plus the \acr{ISO}-3166 country code as the key. Note that only a small number of patent keys is defined by default, mainly to illustrate this scheme. These keys are used in the \bibfield{type} field of \bibtype{patent} entries.

\begin{keymarglist}
\item[patent] The generic term <patent>.
\item[patentde] The expression <German patent>.
\item[patenteu] The expression <European patent>.
\item[patentfr] The expression <French patent>.
\item[patentuk] The expression <British patent>.
\item[patentus] The expression <U.S. patent>.
\end{keymarglist}
%
Patent requests are handled in a similar way, using the string \texttt{patreq} as the base name of the key:

\begin{keymarglist}
\item[patreq] The generic term <patent request>.
\item[patreqde] The expression <German patent request>.
\item[patreqeu] The expression <European patent request>.
\item[patreqfr] The expression <French patent request>.
\item[patrequk] The expression <British patent request>.
\item[patrequs] The expression <U.S. patent request>.
\end{keymarglist}

\paragraph{Dates and Times}
\label{aut:lng:key:dt}

Abbreviation strings for standard eras. Both secular and Christian variants
are supported.

\begin{keymarglist}
\item[commonera] The era <CE>
\item[beforecommonera] The era <BCE>
\item[annodomini] The era <AD>
\item[beforechrist] The era <BC>
\end{keymarglist}

Abbreviation strings for <circa> dates:

\begin{keymarglist}
\item[circa] The string <circa>
\end{keymarglist}

Abbreviation strings for seasons parsed from \acr{ISO8601-2} Extended Format dates:

\begin{keymarglist}
\item[spring] The string <spring>
\item[summer] The string <summer>
\item[autumn] The string <autumn>
\item[winter] The string <winter>
\end{keymarglist}

Abbreviation strings for AM/PM:

\begin{keymarglist}
\item[am] The string <AM>
\item[pm] The string <PM>
\end{keymarglist}

\subsection{Formatting Commands}
\label{aut:fmt}

This section corresponds to \secref{use:fmt} in the user part of this manual. Bibliography and citation styles should incorporate the commands and facilities discussed in this section in order to provide a certain degree of high"=level configurability. Users should not be forced to write new styles if all they want to do is modify the spacing in the bibliography or the punctuation used in citations.

\subsubsection{User-definable Commands and Hooks}
\label{aut:fmt:fmt}

This section corresponds to \secref{use:fmt:fmt} in the user part of the manual. The commands and hooks discussed here are meant to be redefined by users, but bibliography and citation styles may provide a default definition which is different from the package default. These commands are defined in \path{biblatex.def}. Note that all commands starting with \cmd{mk\dots} take one mandatory argument.

\begin{ltxsyntax}
\csitem{bibsetup}
Arbitrary code to be executed at the beginning of the bibliography, intended for commands which affect the layout of the bibliography.

\csitem{bibfont}
Arbitrary code setting the font used in the bibliography. This is very similar to \cmd{bibsetup} but intended for switching fonts.

\csitem{citesetup}
Arbitrary code to be executed at the beginning of each citation command.

\csitem{newblockpunct}
The separator inserted between <blocks> in the sense explained in \secref{aut:pct:new}. The default definition is controlled by the package option \opt{block} (see \secref{use:opt:pre:gen}).

\csitem{newunitpunct}
The separator inserted between <units> in the sense explained in \secref{aut:pct:new}. This will usually be a period or a comma plus an interword space. The default definition is a period and a space.

\csitem{finentrypunct}
The punctuation printed at the very end of every bibliography entry, usually a period. The default definition is a period.

\csitem{entrysetpunct}
The punctuation printed between bibliography subentries of an entry set. The default definition is a semicolon and a space.

\csitem{bibnamedelima}
This delimiter controls the spacing between the elements which make up a name part. It is inserted automatically by the backend after the first name element if the element is less than three characters long and before the last element. The default definition is \cmd{addhighpenspace}, \ie a space penalized by the value of the \cnt{highnamepenalty} counter (\secref{use:fmt:len}). Please refer to \secref{use:cav:nam} for further details.

\csitem{bibnamedelimb}
This delimiter controls the spacing between the elements which make up a name part. It is inserted automatically by the backend between all name elements where \cmd{bibnamedelima} does not apply. The default definition is \cmd{addlowpenspace}, \ie a space penalized by the value of the \cnt{lownamepenalty} counter (\secref{use:fmt:len}). Please refer to \secref{use:cav:nam} for further details.

\csitem{bibnamedelimc}
This delimiter controls the spacing between name parts. The default name formats use it between the name prefix and the family name if \kvopt{useprefix}{true}. The default definition is \cmd{addhighpenspace}, \ie a space penalized by the value of the \cnt{highnamepenalty} counter (\secref{use:fmt:len}). Please refer to \secref{use:cav:nam} for further details.

\csitem{bibnamedelimd}
This delimiter controls the spacing between name parts. The default name formats use it between all name parts where \cmd{bibnamedelimc} does not apply. The default definition is \cmd{addlowpenspace}, \ie a space penalized by the value of the \cnt{lownamepenalty} counter (\secref{use:fmt:len}). Please refer to \secref{use:cav:nam} for further details.

\csitem{bibnamedelimi}
This delimiter replaces \cmd{bibnamedelima/b} after initials. Note that this only applies to initials given as such in the \file{bib} file, not to the initials automatically generated by \biblatex which use their own set of delimiters.

\csitem{bibinitperiod}
The punctuation inserted automatically by the backend after all initials unless \cmd{bibinithyphendelim} applies. The default definition is a period (\cmd{adddot}). Please refer to \secref{use:cav:nam} for further details.

\csitem{bibinitdelim}
The spacing inserted automatically by the backend between multiple initials unless \cmd{bibinithyphendelim} applies. The default definition is an unbreakable interword space. Please refer to \secref{use:cav:nam} for further details.

\csitem{bibinithyphendelim}
The punctuation inserted automatically by the backend between the initials of hyphenated name parts, replacing \cmd{bibinitperiod} and \cmd{bibinitdelim}. The default definition is a period followed by an unbreakable hyphen. Please refer to \secref{use:cav:nam} for further details.

\csitem{bibindexnamedelima}
Replaces \cmd{bibnamedelima} in the index.

\csitem{bibindexnamedelimb}
Replaces \cmd{bibnamedelimb} in the index.

\csitem{bibindexnamedelimc}
Replaces \cmd{bibnamedelimc} in the index.

\csitem{bibindexnamedelimd}
Replaces \cmd{bibnamedelimd} in the index.

\csitem{bibindexnamedelimi}
Replaces \cmd{bibnamedelimi} in the index.

\csitem{bibindexinitperiod}
Replaces \cmd{bibinitperiod} in the index.

\csitem{bibindexinitdelim}
Replaces \cmd{bibinitdelim} in the index.

\csitem{bibindexinithyphendelim}
Replaces \cmd{bibinithyphendelim} in the index.

\csitem{revsdnamepunct}
The punctuation to be printed between the given and family name parts when a name is reversed. The default is a comma. This command should be incorporated in formatting directives for name lists.  Please refer to \secref{use:cav:nam} for further details.

\csitem{bibnamedash}
The dash to be used as a replacement for recurrent authors or editors in the bibliography. The default is an <em> or an <en> dash, depending on the indentation of the list of references.

\csitem{labelnamepunct}\DeprecatedMark
A separator to be printed after the name used for alphabetizing in the bibliography (\bibfield{author} or \bibfield{editor}, if the \bibfield{author} field is undefined) instead of \cmd{newunitpunct}. The default is \cmd{newunitpunct}, \ie it is not handled differently from regular unit punctuation but permits convenient reconfiguration. This punctuation command is deprecated and has been superseded by the context-sensitive \cmd{nametitledelim} (see \secref{use:fmt:csd}). For backwards compatibility reasons, however, \cmd{nametitledelim} still defaults to \cmd{labelnamepunct} in the \texttt{bib} and \texttt{biblist} contexts. Style authors may want to consider replacing \cmd{labelnampunct} with \texttt{\textbackslash printdelim\{nametitledelim\}} and users may want to prefer modifying the context-sensitive \texttt{nametitledelim} with \cmd{DeclareDelimFormat} over redefining \cmd{labelnamepunct}.

\csitem{subtitlepunct}
The separator to be printed between the fields \bibfield{title} and \bibfield{subtitle}, \bibfield{booktitle} and \bibfield{booksubtitle}, as well as \bibfield{maintitle} and \bibfield{mainsubtitle}. Use this separator instead of \cmd{newunitpunct} at this location. The default is \cmd{newunitpunct}, \ie it is not handled differently from regular unit punctuation but permits convenient reconfiguration.

\csitem{intitlepunct}
The separator to be printed between the word «in» and the following title in entry types such as \bibtype{article}, \bibtype{inbook}, \bibtype{incollection}, etc. Use this separator instead of \cmd{newunitpunct} at this location. The default definition is a colon plus an interword space.

\csitem{bibpagespunct}
The separator to be printed before the \bibfield{pages} field. Use this separator instead of \cmd{newunitpunct} at this location. The default is a comma plus an interword space.

\csitem{bibpagerefpunct}
The separator to be printed before the \bibfield{pageref} field. Use this separator instead of \cmd{newunitpunct} at this location. The default is an interword space.

\csitem{bibeidpunct}
The separator printed before the \bibfield{eid} field (similar to \cmd{bibpagespunct}). The default is a comma plus an interword space.

\csitem{multinamedelim}\CSdelimMark
The delimiter to be printed between multiple items in a name list like \bibfield{author} or \bibfield{editor} if there are more than two names in the list. If there are only two names in the list, use the \cmd{finalnamedelim} instead. This command should be incorporated in all formatting directives for name lists. The default is a comma followed by an interword space.

\csitem{finalnamedelim}\CSdelimMark
Use this command instead of \cmd{multinamedelim} before the final name in a name list. The default is the localised term <and>, separated by interword spaces.

\csitem{revsdnamedelim}\CSdelimMark
The extra delimiter to be printed after the first name in a name list consisting of two names (in addition to \cmd{finalnamedelim}) if the first name is reversed. This command should be incorporated in all formatting directives for name lists.

\csitem{andothersdelim}\CSdelimMark
The delimiter to be printed before the localisation string <\texttt{andothers}> if a name list like \bibfield{author} or \bibfield{editor} is truncated. This command should be incorporated in all formatting directives for name lists. The default is an interword space.

\csitem{multilistdelim}\CSdelimMark
The delimiter to be printed between multiple items in a literal list like \bibfield{publisher} or \bibfield{location} if there are more than two names in the list. If there are only two items in the list, use the \cmd{finallistdelim} instead. This command should be incorporated in all formatting directives for literal lists. The default is a comma plus an interword space.

\csitem{finallistdelim}\CSdelimMark
Use this command instead of \cmd{multilistdelim} before the final item in a literal list. The default is the localised term <and>, separated by interword spaces.

\csitem{andmoredelim}\CSdelimMark
The delimiter to be printed before the localisation string <\texttt{andmore}> if a literal list like \bibfield{publisher} or \bibfield{location} is truncated. This command should be incorporated in all formatting directives for literal lists. The default is an interword space.

\csitem{multicitedelim}
The delimiter printed between citations if multiple entry keys are passed to a single citation command. This command should be incorporated in the definition of all citation commands, for example in the \prm{sepcode} argument passed to \cmd{DeclareCiteCommand}. See \secref{aut:cbx:cbx} for details. The default is a semicolon plus an interword space.

\csitem{multiciterangedelim}
The delimiter printed between two citations if they are compressed to a range. The default is \cmd{bibrangedash}.

\csitem{multicitesubentrydelim}
The delimiter printed between subentry citations of the same set. This delimiter is only used in citation styles that reduce citations of the same set to a more compact form (\opt{subentry} of \texttt{numeric-comp}). The default is a comma.

\csitem{multicitesubentryrangedelim}
The delimiter printed between two citations of the same set if they are compressed to a range. The default is \cmd{multiciterangedelim}.

\csitem{supercitedelim}
Similar to \cmd{multinamedelim}, but intended for the \cmd{supercite} command only. The default is a comma.

\csitem{superciterangedelim}
Analogue of \cmd{multiciterangedelim} for \cmd{supercite}. The default is \cmd{bibrangedash}.

\csitem{supercitesubentrydelim}
Analogue of \cmd{multicitesubentrydelim} for \cmd{supercite}. The default is \cmd{supercitedelim}.

\csitem{supercitesubentryrangedelim}
Analogue of \cmd{multicitesubentryrangedelim} for \cmd{supercite}. The default is \cmd{superciterangedelim}.

\csitem{compcitedelim}
Similar to \cmd{multicitedelim}, but intended for citation styles that <compress> multiple citations, \ie print the author only once if subsequent citations share the same author etc. The default definition is a comma plus an interword space.

\csitem{textcitedelim}
Similar to \cmd{multicitedelim}, but intended for \cmd{textcite} and related commands (\secref{use:cit:cbx}). The default is a comma plus an interword space. The standard styles modify this provisional definition to ensure that the delimiter before the final citation is the localised term <and>, separated by interword spaces.

\csitem{nametitledelim}\CSdelimMark
The delimiter to be printed between the author\slash editor and the title. This command should be incorporated in the definition of all citation commands of author-title and some verbose citation styles and in the bibliography drivers---in author-year bibliographies \cs{nametitledelim} may be printed between the author\slash editor-year block and the title. The default definition inside bibliographies is the now deprecated \cmd{labelnamepunct} (for backwrds compatibility reasons) and is a comma plus an interword space otherwise.

\csitem{nameyeardelim}\CSdelimMark
The delimiter to be printed between the author\slash editor and the year. This command should be incorporated in the definition of all citation commands of author-year citation styles and in the bibliography drivers. The default definition is an interword space.

\csitem{namelabeldelim}\CSdelimMark
The delimiter printed between the name\slash title and the label. This command should be incorporated in the definition of all citation commands of alphabetic and numeric citation styles. The default definition is an interword space.

\csitem{nonameyeardelim}\CSdelimMark
The delimiter printed between the substitute for the labelname when it does not exist (usually the label or title in standard styles) and the year citation styles and the bibliography drivers. This command should be incorporated in the definition of all citation commands of author-year citation styles and in the bibliography drivers. The default definition is an interword space.

\csitem{authortypedelim}\CSdelimMark
The delimiter printed between the author and the \texttt{authortype}. The default is a comma followed by a space.

\csitem{editortypedelim}\CSdelimMark
The delimiter printed between the editor and the \texttt{editor} or \texttt{editortype} string. The default is a comma followed by a space.

\csitem{translatortypedelim}\CSdelimMark
The delimiter printed between the translator and the \texttt{translator} string. The default is a comma followed by a space.

\csitem{labelalphaothers}
A string to be appended to the non"=numeric portion of the \bibfield{labelalpha} field (\ie the field holding the citation label used by alphabetic citation styles) if the number of authors\slash editors exceeds the \opt{maxalphanames} threshold or the \bibfield{author}\slash \bibfield{editor} list was truncated in the \file{bib} file with the keyword <\texttt{and others}>. This will typically be a single character such as a plus sign or an asterisk. The default is a plus sign. This command may also be redefined to an empty string to disable this feature. In any case, it must be redefined in the preamble.

\csitem{sortalphaothers}
Similar to \cmd{labelalphaothers} but used in the sorting process. Setting it to a different value is advisable if the latter contains formatting commands. If \cmd{sortalphaothers} is not redefined, it defaults to \cmd{labelalphaothers}.

\csitem{volcitedelim}
The delimiter to be printed between the volume portion and the page/text portion of \cmd{volcite} and related commands (\secref{use:cit:spc}).

\csitem{prenotedelim}\CSdelimMark
The delimiter to be printed after the \prm{prenote} argument of a citation command. The default is an interword space.

\csitem{postnotedelim}\CSdelimMark
The delimiter to be printed before the \prm{postnote} argument of a citation command. The default is a comma plus an interword space.

\csitem{extpostnotedelim}\CSdelimMark
The delimiter printed between the citation and the parenthetical \prm{postnote} argument of a citation command when the postnote occurs outside of the citation parentheses. In the standard styles, this occurs when the citation uses the shorthand field of the entry. The default is an interword space.

\csitem{multiprenotedelim}\CSdelimMark
The delimiter to be printed after the \prm{multiprenote} argument of a citation command.

\csitem{multipostnotedelim}\CSdelimMark
The delimiter to be printed before the \prm{multipostnote} argument of a citation command.

\cmditem{mkbibname<namepart>}{text}
Formatting hook for the name part <namepart>, to be used in all formatting directives for name lists. The default datamodel defines the name parts <family>, <given>, <prefix> and <suffix> and therefore the following macros are automatically defined:

\begin{ltxexample}
\mkbibnamefamily
\mkbibnamegiven
\mkbibnameprefix
\mkbibnamesuffix
\end{ltxexample}

\cmditem{mkbibcompletename<formatorder>}{text}
Formatting hook for the complete name in format order <formatorder>. The default styles use the name format orders <family>, <family-given> and <given-family>, therefore the following macros are automatically defined:

\begin{ltxexample}
\mkbibcompletenamefamily
\mkbibcompletenamefamilygiven
\mkbibcompletenamegivenfamily
\end{ltxexample}
%
These formatting hooks should enclose the complete name in the bibliography macro \cmd{name:<formatorder>}. Initially all hooks expand to \cmd{mkbibcompletename}.

\cmditem{mkbibcompletename}{text}
The initial value of all default formatting hooks \cmd{mkbibcompletename<formatorder>}.

\csitem{datecircadelim}\CSdelimMark
When formatting dates with the global option \opt{datecirca} enabled, the delimiter printed after any localised <circa> term. Defaults to interword space.

\csitem{dateeradelim}\CSdelimMark
When formatting dates with the global option \opt{dateera} set, the delimiter printed before the localisation era term. Defaults to interword space.

\csitem{dateuncertainprint}
Prints date uncertainty information when the global option \opt{dateuncertain} is enabled and the \cmd{ifdateuncertain} test is true. By default, prints the language specific \cmd{bibdateuncertain} string (\secref{use:fmt:lng}).

\csitem{enddateuncertainprint}
Prints date uncertainty information when the global option \opt{dateuncertain} is enabled and the \cmd{ifenddateuncertain} test is true. By default, prints the language specific \cmd{bibdateuncertain} string (\secref{use:fmt:lng}).

\csitem{datecircaprint}
Prints date circa information when the global option \opt{datecirca} is enabled and the \cmd{ifdatecirca} test is true. By default, prints the <circa> localised term (\secref{aut:lng:key:dt}) and the \opt{datecircadelim} delimiter.

\csitem{enddatecircaprint}
Prints date circa information when the global option \opt{datecirca} is enabled and the \cmd{ifenddatecirca} test is true. By default, prints the <circa> localised term (\secref{aut:lng:key:dt}) and the \opt{datecircadelim} delimiter.

\csitem{datecircaprintiso}
Prints \acr{ISO8601-2} format date circa information when the global option \opt{datecirca} is enabled and the \cmd{ifdatecirca} test is true. Prints \cmd{textasciitilde}.

\csitem{enddatecircaprintiso}
Prints \acr{ISO8601-2} format date circa information when the global option \opt{datecirca} is enabled and the \cmd{ifenddatecirca} test is true. Prints \cmd{textasciitilde}.

\csitem{dateeraprint}{yearfield}
Prints date era information when the global option \opt{dateera} is set to <secular> or <christian>. By default, prints the \opt{dateeradelim} delimiter and the appropriate localised era term (\secref{aut:lng:key:dt}). If the \opt{dateeraauto} option is set, then the passed \prm{yearfield} (which is the name of a year field such as <year>, <origyear>, <endeventyear> etc.) is tested to see if its value is earlier than the \opt{dateeraauto} threshold and if so, then the BCE/CE localisation will be output too. The default setting for \opt{dateeraauto} is 0 and so only BCE/BC localisation strings are candidates for output. Detects whether the start or end year era information is to be printed by looking at the \prm{yearfield} name passed to it.

\csitem{dateeraprintpre}
Prints date era information when the global option \opt{dateera} is set to <astronomical>. By default, prints \opt{bibdataeraprefix}. Detects whether the start or end year era information is to be printed by looking at the \prm{yearfield} name passed to it.

\csitem{relatedpunct}
The separator between the \bibfield{relatedtype} bibliography localisation string and the data from the first related entry.

\csitem{relateddelim}
The generic separator between the data of multiple related entries. The default definition is an optional dot plus linebreak.

\csitem{relateddelim$<$relatedtype$>$}
The separator between the data of multiple related entries inside related entries of type <relatedtype>. There is no default, if such a type-specific delimiter does not exist, \cmd{relateddelim} is used.

\csitem{begrelateddelim}
The generic separator before the block of related entries. The default definition is \cmd{newunitpunct}.

\csitem{begrelateddelim$<$relatedtype$>$}
The separator between the block of related entries of type <relatedtype>. There is no default, if such a type-specific delimiter does not exist, \cmd{relateddelim} is used.

\end{ltxsyntax}

\subsubsection{Language-specific Commands}
\label{aut:fmt:lng}

This section corresponds to \secref{use:fmt:lng} in the user part of the manual. The commands discussed here are usually handled by the localisation modules, but may also be redefined by users on a per"=language basis. Note that all commands starting with \cmd{mk\dots} take one or more mandatory arguments.

\begin{ltxsyntax}

\csitem{bibrangedash}

The language specific dash to be used for ranges of numbers. Defaults to \cmd{textendash}.

\csitem{bibrangessep}

The language specific separator to be used between multiple ranges. Defaults to a comma followed by a space.

\csitem{bibdatesep}

The language specific separator used between date components in terse date formats. Defaults to \cmd{hyphen}.

\csitem{bibdaterangesep}

The language specific separator to be used for date ranges. Defaults to \cmd{textendash} for all date formats apart from \opt{ymd} which defaults to a \cmd{slash}. The date format option \opt{iso} is hard-coded to \cmd{slash} since this is a standards compliant format.

\csitem{mkbibdatelong}

Takes the names of three field as arguments which correspond to three date components (in the order year\slash month\slash day) and uses their values to print the date in the language specific long date format.

\csitem{mkbibdateshort}

Similar to \cmd{mkbibdatelong} but using the language specific short date format.

\csitem{mkbibtimezone}

Modifies a timezone string passed in as the only argument. By default this changes <Z> to the value of \cmd{bibtimezone}.

\csitem{bibdateuncertain}

The language specific marker to be used after uncertain dates when the global option \opt{dateuncertain} is enabled. Defaults to a space followed by a question mark.


\csitem{bibdateeraprefix}

The language specific marker which is printed as a prefix to beginning BCE/BC dates in a date range when the option \opt{dateera} is set to <astronomical>. Defaults to \cmd{textminus}, if defined and \cmd{textendash} otherwise.

\csitem{bibdateeraendprefix}

The language specific marker which is printed as a prefix to end BCE/BC dates in a date range when the option \opt{dateera} is set to <astronomical>. Defaults to a thin space followed by \cmd{bibdateeraprefix} when \cmd{bibdaterangesep} is set to a dash and to \cmd{bibdateeraprefix} otherwise.  This is a separate macro so that you may add extra space before a negative date marker which, for example follows a dash date range marker as this can look a little odd.

\csitem{bibtimesep}

The language specific marker which separates time components. Default to a colon.

\csitem{bibutctimezone}

The language specific string printed for the UTC timezone. Defaults to <Z>.

\csitem{bibtimezonesep}

The language specific marker which separates an optional time zone component from a time. Empty by default.

\csitem{bibtzminsep}

The language specific marker which separates hour and minute component of offset timezones. Defaults to a \cmd{bibtimesep}.

\csitem{bibdatetimesep}

The language specific separator printed between date and time components when printing time components along with date components (see the \opt{$<$datetype$>$dateusetime} option in \secref{use:opt:pre:gen}). Defaults to a space for non-\acr{ISO8601-2} output formats, and 'T' for \acr{ISO8601-2} output format.

\csitem{finalandcomma}

Prints the comma to be inserted before the final <and> in an enumeration, if applicable in the respective language.

\csitem{finalandsemicolon}

Prints the semicolon to be inserted before the final <and> in an enumeration, if applicable in the respective language.

\cmditem{mkbibordinal}{integer}

Takes an integer argument and prints it as an ordinal number.

\cmditem{mkbibmascord}{integer}

Similar to \cmd{mkbibordinal}, but prints a masculine ordinal, if applicable in the respective language.

\cmditem{mkbibfemord}{integer}

Similar to \cmd{mkbibordinal}, but prints a feminine ordinal, if applicable in the respective language.

\cmditem{mkbibneutord}{integer}

Similar to \cmd{mkbibordinal}, but prints a neuter ordinal, if applicable in the respective language.

\cmditem{mkbibordedition}{integer}

Similar to \cmd{mkbibordinal}, but intended for use with the term <edition>.

\cmditem{mkbibordseries}{integer}

Similar to \cmd{mkbibordinal}, but intended for use with the term <series>.

\end{ltxsyntax}

\subsubsection{User-definable Lengths and Counters}
\label{aut:fmt:len}

This section corresponds to \secref{use:fmt:len} in the user part of the manual. The length registers and counters discussed here are meant to be altered by users. Bibliography and citation styles should incorporate them where applicable and may also provide a default setting which is different from the package default.

\begin{ltxsyntax}

\lenitem{bibhang}

The hanging indentation of the bibliography, if applicable. This length is initialized to \cmd{parindent} at load-time. If \cmd{parindent} is zero length for some reason, \cmd{bibhang} will default to \texttt{1em}.

\lenitem{biblabelsep}

The horizontal space between entries and their corresponding labels. Bibliography styles which use \env{list} environments and print a label should set \len{labelsep} to \len{biblabelsep} in the definition of the respective environment. This length is initialized to twice the value of \cmd{labelsep} at load-time.

\lenitem{bibitemsep}

The vertical space between the individual entries in the bibliography. Bibliography styles using \env{list} environments should set \len{itemsep} to \len{bibitemsep} in the definition of the respective environment. This length is initialized to \cmd{itemsep} at load-time.

\lenitem{bibnamesep}

Vertical space to be inserted between two entries in the bibliography whenever an entry starts with a name which is different from the initial name of the previous entry. The default value is zero. Setting this length to a positive value greater than \len{bibitemsep} will group the bibliography by author\slash editor name. Note that \len{bibitemsep}, \len{bibnamesep}, and \len{bibinitsep} obey the rules for \cmd{addvspace}, that is, when vertical space introduced by any of these commands immediately follows on from space introduced by another of them, the resulting total space is equal to the largest of them.

\lenitem{bibinitsep}

Vertical space to be inserted between two entries in the bibliography whenever an entry starts with a letter which is different from the initial letter of the previous entry. The default value is zero. Setting this length to a positive value greater than \len{bibitemsep} will group the bibliography alphabetically. Note that \len{bibitemsep}, \len{bibnamesep}, and \len{bibinitsep} obey the rules for \cmd{addvspace}, that is, when vertical space introduced by any of these commands immediately follows on from space introduced by another of them, the resulting total space is equal to the largest of them.

\lenitem{bibparsep}

The vertical space between paragraphs within an entry in the bibliography. Bibliography styles using \env{list} environments should set \len{parsep} to \len{bibparsep} in the definition of the respective environment. The default value is zero.

\cntitem{abbrvpenalty}

The penalty used by \cmd{addabbrvspace}, \cmd{addabthinspace}, and \cmd{adddotspace}, see \secref{aut:pct:spc} for details. This counter is initialized to \cmd{hyphenpenalty} at load-time.

\cntitem{highnamepenalty}

The penalty used by \cmd{addhighpenspace} and \cmd{addhpthinspace}, see \secref{aut:pct:spc} for details. The counter is initialized to \cmd{hyphenpenalty} at load-time.

\cntitem{lownamepenalty}

The penalty used by \cmd{addlowpenspace} and \cmd{addlpthinspace}, see \secref{aut:pct:spc} for details. The counter is initialized to half the \cmd{hyphenpenalty} at load-time.

\cntitem{biburlbigbreakpenalty}

The \sty{biblatex} version of \sty{url}'s \len{UrlBigBreakPenalty}. The default value is \texttt{100}.

\cntitem{biburlbreakpenalty}

The \sty{biblatex} version of \sty{url}'s \len{UrlBreakPenalty}. The default value is \texttt{200}.

\cntitem{biburlnumpenalty}

If this counter is set to a value greater than zero, \biblatex will permit linebreaks after numbers in all strings formatted with the \cmd{url} command from the \sty{url} package. This will affect \acr{url}s and \acr{doi}s in the bibliography. The breakpoints will be penalized by the value of this counter. If \acr{url}s and/or \acr{doi}s in the bibliography run into the margin, try setting this counter to a value greater than zero but less than 10000 (you normally want to use a high value like 9000). Setting the counter to zero disables this feature. This is the default setting.

\cntitem{biburlucpenalty}

Similar to \cnt{biburlnumpenalty}, except that it will add a breakpoint after all uppercase letters.

\cntitem{biburllcpenalty}

Similar to \cnt{biburlnumpenalty}, except that it will add a breakpoint after all lowercase letters.

\lenitem{biburlbigskip}

The \sty{biblatex} version of \len{Urlmuskip}. This length holds the additional (stretchable) space inserted around breakable characters in the \cmd{url} command from the \sty{url} package. The default value is \texttt{0mu plus 3mu}.

\lenitem{biburlnumskip}

The additional space inserted after numbers in strings formatted with the \cmd{url} command from the \sty{url} package. This will affect \acr{url}s and \acr{doi}s in the bibliography. If \acr{url}s and/or \acr{doi}s in the bibliography run into the margin, it may help to set this length to add some small stretchable space, for example \texttt{0mu plus 1mu}. The default setting is \texttt{0mu}. This value is only used if \cnt{biburlnumpenalty} is set to a value different from zero.

\lenitem{biburlucskip}

Similar to \cnt{biburlnumskip}, except that it will add space after all uppercase letters.

\lenitem{biburllcskip}

Similar to \cnt{biburlnumskip}, except that it will add space after all uppercase letters.

\end{ltxsyntax}

\subsubsection{Auxiliary Commands and Hooks}
\label{aut:fmt:ich}

The auxiliary commands and facilities in this section serve a special purpose. Some of them are used by \biblatex to communicate with bibliography and citation styles in some way or other.

\begin{ltxsyntax}

\cmditem{mkbibemph}{text}

A generic command which prints its argument as emphasized text. This is a simple wrapper around the standard \cmd{emph} command. Apart from that, it uses \cmd{setpunctfont} from \secref{aut:pct:new} to adapt the font of the next punctuation mark following the text set in italics. If the \opt{punctfont} package option is disabled, this command behaves like \cmd{emph}.

\cmditem{mkbibitalic}{text}

Similar in concept to \cmd{mkbibemph} but prints italicized text. This is a simple wrapper around the standard \cmd{textit} command which incorporates \cmd{setpunctfont}. If the \opt{punctfont} package option is disabled, this command behaves like \cmd{textit}.

\cmditem{mkbibbold}{text}

Similar in concept to \cmd{mkbibemph} but prints bold text. This is a simple wrapper around the standard \cmd{textbf} command which incorporates \cmd{setpunctfont}. If the \opt{punctfont} package option is disabled, this command behaves like \cmd{textbf}.

\cmditem{mkbibquote}{text}

A generic command which wraps its argument in quotation marks. If the \sty{csquotes} package is loaded, this command uses the language sensitive quotation marks provided by that package. \cmd{mkbibquote} also supports <American-style> punctuation, see \cmd{DeclareQuotePunctuation} in \secref{aut:pct:cfg} for details.

\cmditem{mkbibparens}{text}

A generic command which wraps its argument in parentheses. This command is nestable. When nested, it will alternate between parentheses and brackets, depending on the nesting level.

\cmditem{mkbibbrackets}{text}

A generic command which wraps its argument in square brackets. This command is nestable. When nested, it will alternate between brackets and parentheses, depending on the nesting level.

\cmditem{bibopenparen}<text>|{\ltxsyntaxlabelfont\cmd{bibcloseparen}}|

Alternative syntax for \cmd{mkbibparens}. This will also work across groups. Note that \cmd{bibopenparen} and \cmd{bibcloseparen} must always be balanced.

\cmditem{bibopenbracket}<text>|{\ltxsyntaxlabelfont\cmd{bibclosebracket}}|

Alternative syntax for \cmd{mkbibbrackets}. This will also work across groups. Note that \cmd{bibopenbracket} and \cmd{bibclosebracket} must always be balanced.

\cmditem{mkbibfootnote}{text}

A generic command which prints its argument as a footnote. This is a wrapper around the standard \latex \cmd{footnote} command which removes spurious whitespace preceding the footnote mark and prevents nested footnotes. By default, \cmd{mkbibfootnote} requests capitalization at the beginning of the note and automatically adds a period at the end. You may change this behavior by redefining the \cmd{bibfootnotewrapper} macro introduced below.

\cmditem{mkbibfootnotetext}{text}

Similar to \cmd{mkbibfootnote} but uses the \cmd{footnotetext} command.

\cmditem{mkbibendnote}{text}

Similar in concept to \cmd{mkbibfootnote} except that it prints its argument as an endnote. \cmd{mkbibendnote} removes spurious whitespace preceding the endnote mark and prevents nested notes. It supports the \cmd{endnote} command provided by the \sty{endnotes} package as well as the \cmd{pagenote} command provided by the \sty{pagenote} package and the \sty{memoir} class. If both commands are available, \cmd{endnote} takes precedence. If no endnote support is available, \cmd{mkbibendnote} issues an error and falls back to \cmd{footnote}. By default, \cmd{mkbibendnote} requests capitalization at the beginning of the note and automatically adds a period at the end. You may change this behavior by redefining the \cmd{bibendnotewrapper} macro introduced below.

\cmditem{mkbibendnotetext}{text}

Similar to \cmd{mkbibendnote} but uses the \cmd{endnotetext} command. Please note that as of this writing, neither the \sty{pagenote} package nor the \sty{memoir} class provide a corresponding \cmd{pagenotetext} command. In this case, \cmd{mkbibendnote} will issue an error and fall back to \cmd{footnotetext}.

\cmditem{bibfootnotewrapper}{text}

An inner wrapper which encloses the \prm{text} argument of \cmd{mkbibfootnote} and \cmd{mkbibfootnotetext}. For example, \cmd{mkbibfootnote} eventually boils down to this:

\begin{ltxexample}
\footnote{<<\bibfootnotewrapper{>>text<<}>>}
\end{ltxexample}
%
The wrapper ensures capitalization at the beginning of the note and adds a period at the end. The default definition is:

\begin{ltxexample}
\newcommand{\bibfootnotewrapper}[1]{<<\bibsentence>> #1<<\addperiod>>}
\end{ltxexample}
%
If you don't want capitalization at the beginning or a period at the end of the note, do not modify \cmd{mkbibfootnote} but redefine \cmd{bibfootnotewrapper} instead.

\cmditem{bibendnotewrapper}{text}

Similar in concept to \cmd{bibfootnotewrapper} but related to the \cmd{mkbibendnote} and \cmd{mkbibendnotetext} commands.

\cmditem{mkbibsuperscript}{text}

A generic command which prints its argument as superscripted text. This is a simple wrapper around the standard \latex \cmd{textsuperscript} command which removes spurious whitespace and allows hyphenation of the preceding word.

\cmditem{mkbibmonth}{integer}

This command takes an integer argument and prints it as a month name. Even though the output of this command is language specific, its definition is not, hence it is normally not redefined in localisation modules.

\cmditem{mkbibseason}{string}

This command takes a season localisation string and prints the version of the string corresponding to the setting of the \opt{dateabbrev} package option. Even though the output of this command is language specific, its definition is not, hence it is normally not redefined in localisation modules.

\cmditem{mkyearzeros}{integer}

This command strips leading zeros from a year or enforces them, depending on the \opt{datezeros} package option (\secref{use:opt:pre:gen}). It is intended for use in the definition of date formatting macros. If zeros are enforced, this command calls \cmd{forcezerosy} and thus expands its argument with \cmd{protected@edef}.

\cmditem{mkmonthzeros}{integer}

This command strips leading zeros from a month or enforces them, depending on the \opt{datezeros} package option (\secref{use:opt:pre:gen}). It is intended for use in the definition of date formatting macros. If zeros are enforced, this command calls \cmd{forcezerosmdt} and thus expands its argument with \cmd{protected@edef}.

\cmditem{mkdayzeros}{integer}

This command strips leading zeros from a day or enforces them, depending on the \opt{datezeros} package option (\secref{use:opt:pre:gen}). It is intended for use in the definition of date formatting macros. If zeros are enforced, this command calls \cmd{forcezerosmdt} and thus expands its argument with \cmd{protected@edef}.

\cmditem{mktimezeros}{integer}

This command strips leading zeros from a number or preserves them, depending on the \opt{timezeros} package option (\secref{use:opt:pre:gen}). It is intended for use in the definition of time formatting macros. If zeros are enforced, this command calls \cmd{forcezerosmdt} and thus expands its argument with \cmd{protected@edef}.

\cmditem{forcezerosy}{integer}

This command adds zeros to a year (or any number supposed to be 4-digits). It is intended for date formatting and ordinals. The argument is expanded with \cmd{protected@edef} before it is processed.

\cmditem{forcezerosmdt}{integer}

This command adds zeros to a month, day or time part (or any number supposed to be 2-digits). It is intended for date/time formatting and ordinals. The argument is expanded with \cmd{protected@edef} before it is processed.

\cmditem{stripzeros}{integer}

This command strips leading zeros from a number. It is intended for date formatting and ordinals.

\optitem{$<$labelfield$>$width}

For every field marked as a <Label field> in the data model, a formatting directive is created as per \texttt{shorthandwidth} above. Since \bibfield{shorthand} is so marked in the default data model, this functionality is a superset of that described for \texttt{shorthandwidth}.

\optitem{labelnumberwidth}

Similar to \texttt{shorthandwidth}, but referring to the \bibfield{labelnumber} field and the length register \cmd{labelnumberwidth}. Numeric styles should adjust this directive such that it corresponds to the format used in the bibliography.

\optitem{labelalphawidth}

Similar to \texttt{shorthandwidth}, but referring to the \bibfield{labelalpha} field and the length register \cmd{labelalphawidth}. Alphabetic styles should adjust this directive such that it corresponds to the format used in the bibliography.

\optitem{bibhyperref}

A special formatting directive for use with \cmd{printfield} and \cmd{printtext}. This directive wraps its argument in a \cmd{bibhyperref} command, see \secref{aut:aux:msc} for details.

\optitem{bibhyperlink}

A special formatting directive for use with \cmd{printfield} and \cmd{printtext}. It wraps its argument in a \cmd{bibhyperlink} command, see \secref{aut:aux:msc} for details. The \prm{name} argument passed to \cmd{bibhyperlink} is the value of the \bibfield{entrykey} field.

\optitem{bibhypertarget}

A special formatting directive for use with \cmd{printfield} and \cmd{printtext}. It wraps its argument in a \cmd{bibhypertarget} command, see \secref{aut:aux:msc} for details. The \prm{name} argument passed to \cmd{bibhypertarget} is the value of the \bibfield{entrykey} field.

\optitem{volcitepages}

A special formatting directive which controls the format of the page\slash text portion in the argument of citation commands like \cmd{volcite}.

\optitem{volcitevolume}

A special formatting directive which controls the format of the volume portion in the argument of citation commands like \cmd{volcite}.

\optitem{date}

A special formatting directive which controls the format of \cmd{printdate} (\secref{aut:bib:dat}). Note that the date format (long/short etc.) is controlled by the package option \opt{date} from \secref{use:opt:pre:gen}. This formatting directive only controls additional formatting such as fonts etc.

\optitem{labeldate}

As \texttt{date} but controls the format of \cmd{printlabeldate}.

\optitem{$<$datetype$>$date}

As \texttt{date} but controls the format of \cmd{print$<$datetype$>$date}.

\optitem{time}

A special formatting directive which controls the format of \cmd{printtime} (\secref{aut:bib:dat}). Note that the time format (24h/12h etc.) is controlled by the package option \opt{time} from \secref{use:opt:pre:gen}. This formatting directive only controls additional formatting such as fonts etc.

\optitem{labeltime}

As \texttt{time} but controls the format of \cmd{printlabeltime}.

\optitem{$<$datetype$>$time}

As \texttt{time} but controls the format of \cmd{print$<$datetype$>$time}.

\end{ltxsyntax}

\subsubsection{Auxiliary Lengths, Counters, and Other Features}
\label{aut:fmt:ilc}

The length registers and counters discussed here are used by \biblatex to pass information to bibliography and citation styles. Think of them as read"=only registers. Note that all counters are \latex counters. Use |\value{counter}| to read out the current value.

\begin{ltxsyntax}

\lenitem{$<$labelfield$>$width}

For every field marked as a <label> field in the data model, a length register is created as per \texttt{shorthandwidth} above. Since \bibfield{shorthand} is so marked in the default data model, this functionality is a superset of that described for \texttt{shorthandwidth}.

\lenitem{labelnumberwidth}

This length register indicates the width of the widest \bibfield{labelnumber}. Numeric bibliography styles
should incorporate this length in the definition of the bibliography environment.

\lenitem{labelalphawidth}

This length register indicates the width of the widest \bibfield{labelalpha}. Alphabetic bibliography styles should incorporate this length in the definition of the bibliography environment.

\cntitem{maxextraalpha}

This counter holds the highest number found in any \bibfield{extraalpha} field.

\cntitem{maxextradate}

This counter holds the highest number found in any \bibfield{extradate} field.

\cntitem{maxextraname}

This counter holds the highest number found in any \bibfield{extraname} field.

\cntitem{maxextratitle}

This counter holds the highest number found in any \bibfield{extratitle} field.

\cntitem{maxextratitleyear}

This counter holds the highest number found in any \bibfield{extratitleyear} field.

\cntitem{refsection}

This counter indicates the current \env{refsection} environment. When queried in a bibliography heading, the counter returns the value of the \opt{refsection} option passed to \cmd{printbibliography}.

\cntitem{refsegment}

This counter indicates the current \env{refsegment} environment. When queried in a bibliography heading, this counter returns the value of the \opt{refsegment} option passed to \cmd{printbibliography}.

\cntitem{maxnames}

This counter holds the setting of the \opt{maxnames} package option.

\cntitem{minnames}

This counter holds the setting of the \opt{minnames} package option.

\cntitem{maxitems}

This counter holds the setting of the \opt{maxitems} package option.

\cntitem{minitems}

This counter holds the setting of the \opt{minitems} package option.

\cntitem{instcount}

This counter is incremented by \biblatex for every citation as well as for every entry in the bibliography and bibliography lists. The value of this counter uniquely identifies a single instance of a reference in the document.

\cntitem{citetotal}

This counter, which is only available in the \prm{loopcode} of a citation command defined with \cmd{DeclareCiteCommand}, holds the total number of valid entry keys passed to the citation command.

\cntitem{citecount}

This counter, which is only available in the \prm{loopcode} of a citation command defined with \cmd{DeclareCiteCommand}, holds the number of the entry key currently being processed by the \prm{loopcode}.

\cntitem{multicitetotal}

This counter is similar to \cnt{citetotal} but only available in multicite commands. It holds the total number of citations passed to the multicite command. Note that each of these citations may consist of more than one entry key. This information is provided by the \cnt{citetotal} counter.

\cntitem{multicitecount}

This counter is similar to \cnt{citecount} but only available in multicite commands. It holds the number of the citation currently being processed. Note that this citation may consist of more than one entry key. This information is provided by the \cnt{citetotal} and \cnt{citecount} counters.

\cntitem{listtotal}

This counter holds the total number of items in the current list. It is intended for use in list formatting directives and does not hold a meaningful value when used anywhere else. As an exception, it may also be used in the second optional argument to \cmd{printnames} and \cmd{printlist}, see \secref{aut:bib:dat} for details. For every list, there is also a counter by the same name which holds the total number of items in the corresponding list. For example, the \cnt{author} counter holds the total number of items in the \bibfield{author} list. This applies to both name lists and literal lists. These counters are similar to \cnt{listtotal} except that they may also be used independently of list formatting directives. For example, a bibliography style might check the \cnt{editor} counter to decide Whether or not to print the term «editor» or rather its plural form «editors» after the list of editors.

\cntitem{listcount}

This counter holds the number of the list item currently being processed. It is intended for use in list formatting directives and does not hold a meaningful value when used anywhere else.

\cntitem{liststart}

This counter holds the \prm{start} argument passed to \cmd{printnames} or \cmd{printlist}. It is intended for use in list formatting directives and does not hold a meaningful value when used anywhere else.

\cntitem{liststop}

This counter holds the \prm{stop} argument passed to \cmd{printnames} or \cmd{printlist}. It is intended for use in list formatting directives and does not hold a meaningful value when used anywhere else.

\csitem{currentlang}

The name of the currently active language for \biblatex. Can be used anywhere and
defaults to the main document language. This is automatically switched
inside entries which define \bibfield{langid}, given suitable settings of the
\opt{autolang} and \opt{language} options. Note that this does not track
all document language changes, only the current \biblatex\ setting.

\csitem{currentfield}

The name of the field currently being processed by \cmd{printfield}. This information is only available locally in field formatting directives.

\csitem{currentlist}

The name of the literal list currently being processed by \cmd{printlist}. This information is only available locally in list formatting directives.

\csitem{currentname}

The name of the name list currently being processed by \cmd{printnames}. This information is only available locally in name formatting directives.

\end{ltxsyntax}

\subsubsection{General Purpose Hooks}
\label{aut:fmt:hok}

\begin{ltxsyntax}

\cmditem{AtBeginBibliography}{code}

Appends the \prm{code} to an internal hook executed at the beginning of the bibliography. The \prm{code} is executed at the beginning of the list of references, immediately after the \prm{begin code} of \cmd{defbibenvironment}. This command may only be used in the preamble.

\cmditem{AtBeginShorthands}{code}

Appends the \prm{code} to an internal hook executed at the beginning of the list of shorthands. The \prm{code} is executed at the beginning of the list of shorthands, immediately after the \prm{begin code} of \cmd{defbibenvironment}. This command may only be used in the preamble.

This is just an alias for:

\begin{ltxexample}
\AtBeginBiblist{shorthand}{code}
\end{ltxexample}

\cmditem{AtBeginBiblist}{biblistname}{code}

Appends the \prm{code} to an internal hook executed at the beginning of the bibliography list \prm{biblistname}. The \prm{code} is executed at the beginning of the bibliography list, immediately after the \prm{begin code} of \cmd{defbibenvironment}. This command may only be used in the preamble.

\cmditem{AtEveryBibitem}{code}

Appends the \prm{code} to an internal hook executed at the beginning of every item in the bibliography. The \prm{code} is executed immediately after the \prm{item code} of \cmd{defbibenvironment}. The bibliographic data of the respective entry is available at this point. This command may only be used in the preamble.

\cmditem{AtEveryLositem}{code}

Appends the \prm{code} to an internal hook executed at the beginning of every item in the list of shorthands. The \prm{code} is executed immediately after the \prm{item code} of \cmd{defbibenvironment}. The bibliographic data of the respective entry is available at this point. This command may only be used in the preamble.

This is just an alias for:

\begin{ltxexample}
\AtEveryBiblistitem{shorthand}{code}
\end{ltxexample}

\cmditem{AtEveryBiblistitem}{biblistname}{code}

Appends the \prm{code} to an internal hook executed at the beginning of every item in the bibliography list named \prm{biblistname}. The \prm{code} is executed immediately after the \prm{item code} of \cmd{defbibenvironment}. The bibliographic data of the respective entry is available at this point. This command may only be used in the preamble.

\cmditem{AtNextBibliography}{code}

Similar to \cmd{AtBeginBibliography} but only affecting the next \cmd{printbibliography}. The internal hook is cleared after being executed once. This command may be used in the document body.

\cmditem{AtUsedriver}{code}
\cmditem*{AtUsedriver}*{code}

Appends the \prm{code} to an internal hook executed when initializing \cmd{usedriver}. The starred variant of the command clears the initialisation hook, so the defaults can be overwritten. This command may only be used in the preamble.
The default setting is:

\begin{ltxexample}
\AtUsedriver{%
  \delimcontext{bib}%
  \let\finentry\blx@finentry@usedrv
  \let\newblock\relax
  \let\abx@macro@bibindex\@empty
  \let\abx@macro@pageref\@empty}
\end{ltxexample}

\cmditem{AtEveryCite}{code}

Appends the \prm{code} to an internal hook executed at the beginning of every citation command. The \prm{code} is executed immediately before the \prm{precode} of the command (see \secref{aut:cbx:cbx}). No bibliographic data is available at this point. This command may only be used in the preamble.

\cmditem{AtEveryCitekey}{code}

Appends the \prm{code} to an internal hook executed once for every entry key passed to a citation command. The \prm{code} is executed immediately before the \prm{loopcode} of the command (see \secref{aut:cbx:cbx}). The bibliographic data of the respective entry is available at this point. This command may only be used in the preamble.

\cmditem{AtEveryMultiCite}{code}

Appends the \prm{code} to an internal hook executed at the beginning of every multicite command. The \prm{code} is executed immediately before the \bibfield{multiprenote} field (\secref{aut:cbx:fld}) is printed. No bibliographic data is available at this point. This command may only be used in the preamble.

\cmditem{AtNextCite}{code}

Similar to \cmd{AtEveryCite} but only affecting the next citation command. The internal hook is cleared after being executed once. This command may be used in the document body.

\cmditem{AtEachCitekey}{code}

Similar to \cmd{AtEveryCitekey} but only affecting the current citation command. This command may be used in the document body. The \prm{code} is appended to the internal hook locally when located in a citation, as determined by \cmd{ifcitation}.

\cmditem{AtNextCitekey}{code}

Similar to \cmd{AtEveryCitekey} but only affecting the next entry key. The internal hook is cleared after being executed once. This command may be used in the document body.

\cmditem{AtNextMultiCite}{code}

Similar to \cmd{AtEveryMultiCite} but only affecting the next multicite command. The internal hook is cleared after being executed once. This command may be used in the document body.

\cmditem{AtVolcite}{code}
\cmditem*{AtVolcite}*{code}

Appends the \prm{code} to an internal hook executed when initializing \cmd{volcite}. The starred variant of the command clears the initialisation hook, so the defaults can be overwritten. This command may only be used in the preamble.
The default setting is:

\begin{ltxexample}
\AtVolcite{%
  \DeclareFieldAlias{postnote}{volcitenote}}
\end{ltxexample}

\cmditem{AtDataInput}[entrytype]{code}

Appends the \prm{code} to an internal hook executed once for every entry as the bibliographic data is imported from the \file{bbl} file. The \prm{entrytype} is the entry type the \prm{code} applies to. If it applies to all entry types, omit the optional argument. The \prm{code} is executed immediately after the entry has been imported. This command may only be used in the preamble. Note that \prm{code} may be executed multiple times for an entry. This occurs when the same entry is cited in different \env{refsection} environments or the \opt{sorting} option settings incorporate more than one sorting template. The \cnt{refsection} counter holds the number of the respective reference section while the data is imported.

\cmditem{UseBibitemHook}

Executes the internal hook corresponding to \cmd{AtEveryBibitem}.

\cmditem{UseUsedriverHook}

Executes the internal hook corresponding to \cmd{AtUsedriver}.

\cmditem{UseEveryCiteHook}

Executes the internal hook corresponding to \cmd{AtEveryCite}.

\cmditem{UseEveryCitekeyHook}

Executes the internal hook corresponding to \cmd{AtEveryCitekey}.

\cmditem{UseEveryMultiCiteHook}

Executes the internal hook corresponding to \cmd{AtMultiEveryCite}.

\cmditem{UseNextCiteHook}

Executes and clears the internal hook corresponding to \cmd{AtNextCite}.

\cmditem{UseNextCitekeyHook}

Executes and clears the internal hook corresponding to \cmd{AtNextCitekey}.

\cmditem{UseNextMultiCiteHook}

Executes and clears the internal hook corresponding to \cmd{AtNextMultiCite}.

\cmditem{UseVolciteHook}

Executes the internal hook corresponding to \cmd{AtVolcite}.

\cmditem{DeferNextCitekeyHook}

Locally un-defines the internal hook specified by \cmd{AtNextCitekey}. This essentially defers the hook to the next entry key in the citation list, when executed in the \prm{precode} argument of \cmd{DeclareCiteCommand} (\secref{aut:cbx:cbx}).

\cmditem{AtEveryEntrykey}{code}{success}{failure}

Appends \prm{code} to an internal hook executed every time an entrykey is processed for a citation command or \cmd{nocite}. The \prm{code} is passed one argument (\lstinline{#1}), which contains the entrykey. If the code can be appended to the hook \prm{success} is executed, otherwise \prm{failure} is executed. Unlike \cmd{AtEveryCitekey} the entry data of the current entrykey is not available when \prm{code} is processed, indeed it is not even known whether or not there is any entry data at all.

\end{ltxsyntax}

\subsubsection{File hooks}
\label{aut:fmt:hok:fil}
\biblatex has rudimentary support for injecting arbitrary code before and after a file is loaded via file hooks. For files that are loaded using \biblatex's file interface---that includes all bibliography and citation styles---the following three hooks are available

\begin{ltxsyntax}
\cmditem{blx@filehook@preload@$<$filename with extension$>$}

If \file{$<$filename with extension$>$} is found, this hook is exected before it is loaded.

\cmditem{blx@filehook@postload@$<$filename with extension$>$}

If \file{$<$filename with extension$>$} is found, this hook is exected after it is loaded.

\cmditem{blx@filehook@failure@$<$filename with extension$>$}

This hook is executed if \file{$<$filename with extension$>$} can not be found.
\end{ltxsyntax}

\biblatex generally only loads files once even if they were requested multiple times,
so the hooks will only be executed once.
Naturally, the file hooks need to be populated before the files are loaded, so the safest would be to populate them before \biblatex is loaded.
It is advisable to only append code to avoid overwriting previous hook contents.
Since the name of the file hook include the dot and the file extension they will usually have to be defined with a command like \cmd{csappto} from \sty{etoolbox}.

The \file{.lbx} files are special and may have to be loaded several times in some situations.
Their file hooks are

\begin{ltxsyntax}
\cmditem{blx@lbxfilehook@once@preload@$<$filename with extension$>$}

If \file{$<$filename with extension$>$} is found, this hook is exected before it is loaded in a situation where the \file{.lbx} files are loaded only once.

\cmditem{blx@lbxfilehook@once@postload@$<$filename with extension$>$}

If \file{$<$filename with extension$>$} is found, this hook is exected after it is loaded in a situation where the \file{.lbx} files are loaded only once.

\cmditem{blx@lbxfilehook@once@failure@$<$filename with extension$>$}

This hook is executed if \file{$<$filename with extension$>$} can not be found in a situation where the \file{.lbx} files are loaded only once.

\cmditem{blx@lbxfilehook@simple@preload@$<$filename with extension$>$}

If \file{$<$filename with extension$>$} is found, this hook is exected before it is loaded in a situation where the \file{.lbx} files may be loaded multiple times.

\cmditem{blx@lbxfilehook@simple@postload@$<$filename with extension$>$}

If \file{$<$filename with extension$>$} is found, this hook is exected after it is loaded in a situation where the \file{.lbx} files may be loaded multiple times.

\cmditem{blx@lbxfilehook@simple@failure@$<$filename with extension$>$}

This hook is executed if \file{$<$filename with extension$>$} can not be found in a situation where the \file{.lbx} files may be loaded multiple times.
\end{ltxsyntax}

The following code sets up \sty{beamer} to print the bibliography labels instead of its bibliography icons when \file{numeric.bbx} after is loaded
\begin{ltxexample}
\csappto{blx@filehook@postload@numeric.bbx}{%
  \mode<presentation>{%
    \setbeamertemplate{bibliography item}{%
      \insertbiblabel}}}
\end{ltxexample}


\subsection{Hints and Caveats}
\label{aut:cav}

This section provides some additional hints concerning the author interface of this package. It also addresses common problems and potential misconceptions.

\subsubsection{Entry Sets}
\label{aut:cav:set}

Entry sets have already been introduced in \secref{use:use:set}. This section discusses how to process entry sets in a bibliography style. From the perspective of the driver, there is no difference between static and dynamic entry sets. Both types are handled in the same way. You will normally use the \cmd{entryset} command from \secref{aut:bib:dat} to loop over all set members (in the order in which they are listed in the \bibfield{entryset} field of the \bibtype{set} entry, or in the order in which they were passed to \cmd{defbibentryset}, respectively) and append \cmd{finentry} at the end. That's it. The formatting is handled by the drivers for the entry types of the individual set members:

\begin{ltxexample}
\DeclareBibliographyDriver{set}{%
  <<\entryset>>{}{}%
  \finentry}
\end{ltxexample}
%
You may have noticed that the \texttt{numeric} styles which come with this package support subdivided entry sets, \ie the members of the set are marked with a letter or some other marker such that citations may either refer to the entire set or to a specific set member. The markers are generated as follows by the bibliography style:

\begin{ltxexample}
\DeclareBibliographyDriver{set}{%
  \entryset
    {<<\printfield{entrysetcount}>>%
     <<\setunit*{\addnbspace}>>}
    {}%
  \finentry}
\end{ltxexample}
%
The \bibfield{entrysetcount} field holds an integer indicating the position of a set member in the entry set. The conversion of this number to a letter or some other marker is handled by the formatting directive of the \bibfield{entrysetcount} field. All the driver needs to do is print the field and add some white space (or start a new line). Printing the markers in citations works in a similar way. Where a numeric style normally says |\printfield{labelnumber}|, you simply append the \bibfield{entrysetcount} field:

\begin{ltxexample}
\printfield{labelnumber}<<\printfield{entrysetcount}>>
\end{ltxexample}
%
Since this field is only defined when processing citations referring to a set member, there is no need to add any additional tests.

Citing entry sets directly requires that a meaningful way of identifying sets is available in the style. This is obvious for styles based on numeric or alphabetic labels but not obvious (and rarely required) in styles which construct citations based on textual names/titles/dates etc. The default provided styles which no not construct citations based on labels (\texttt{authoryear}, \texttt{authortitle}, \texttt{verbose} etc.) therefore do not support citing sets directly as there is no obvious default identifier to use in such cases and such styles rarely, if ever, employ sets anyway. Custom styles may of course choose to define and print a citation identifier for directly cited sets.

\subsubsection{Electronic Publishing Information}
\label{aut:cav:epr}

The standard styles feature dedicated support for arXiv references. Support for other resources is easily added. The standard styles handle the \bibfield{eprint} field as follows:

\begin{ltxexample}
\iffieldundef{eprinttype}
  {\printfield{eprint}}
  {\printfield[<<eprint:\strfield{eprinttype}>>]{eprint}}
\end{ltxexample}
%
If an \bibfield{eprinttype} field is available, the above code tries to use the field format \texttt{eprint:\prm{eprinttype}}. If this format is undefined, \cmd{printfield} automatically falls back to the field format \texttt{eprint}. There are two predefined field formats, the type"=specific format \texttt{eprint:arxiv} and the fallback format \texttt{eprint}:

\begin{ltxexample}
\DeclareFieldFormat{<<eprint>>}{...}
\DeclareFieldFormat{<<eprint:arxiv>>}{...}
\end{ltxexample}
%
In other words, adding support for additional resources is as easy as defining a field format named \texttt{eprint:\prm{resource}} where \prm{resource} is an identifier to be used in the \bibfield{eprinttype} field.

\subsubsection{External Abstracts and Annotations}
\label{aut:cav:prf}

External abstracts and annotations have been discussed in \secref{use:use:prf}. This section provides some more background for style authors. The standard styles use the following macros (from \path{biblatex.def}) to handle abstracts and annotations:

\begin{ltxexample}
\newbibmacro*{annotation}{%
  \iffieldundef{annotation}
    {\printfile[annotation]{<<\bibannotationprefix\thefield{entrykey}.tex>>}}%
    {\printfield{annotation}}}
\newcommand*{<<\bibannotationprefix>>}{bibannotation-}

\newbibmacro*{abstract}{%
  \iffieldundef{abstract}
    {\printfile[abstract]{<<\bibabstractprefix\thefield{entrykey}.tex>>}}%
    {\printfield{abstract}}}
\newcommand*{<<\bibabstractprefix>>}{bibabstract-}
\end{ltxexample}
%
If the \bibfield{abstract}\slash \bibfield{annotation} field is undefined, the above code tries to load the abstracts\slash annotations from an external file. The \cmd{printfile} commands also incorporate file name prefixes which may be redefined by users. Note that you must enable \cmd{printfile} explicitly by setting the \opt{loadfiles} package option from \secref{use:opt:pre:gen}. This feature is disabled by default for performance reasons.

\subsubsection[Name Disambiguation]{Name Disambiguation}
\label{aut:cav:amb}

The \opt{uniquename} and \opt{uniquelist} options introduced in \secref{use:opt:pre:int} support various modes of operation. This section explains the differences between these modes by way of example. The \opt{uniquename} option disambiguates individual names in the \bibfield{labelname} list. The \opt{uniquelist} option disambiguates the \bibfield{labelname} list if it has become ambiguous after \opt{maxnames}\slash \opt{minnames} truncation. You can use either option stand-alone or combine both.

Name disambiguation works by taking a <base> which is composed of one or more nameparts and then determining what needs to be added, if anything, to this <base> to make the name unique in the current refsection. Name disambiguation is controlled by the uniquename template declared with the following command:

\begin{ltxsyntax}

\cmditem{DeclareUniquenameTemplate}[name]{specification}

Defines the \opt{uniquename} template \prm{name}. The \prm{name} is optional and defaults to \prm{<global>}.

The \prm{specification} is an ordered list of \cmd{namepart} commands which define the nameparts to use in determining the uniquename information.

\cmditem{namepart}[options]{namepart}

\prm{namepart} is one of the datamodel nameparts defined with the \cmd{DeclareDatamodelConstant} command (see \secref{aut:bbx:drv}). The \prm{options} are:

\begin{optionlist*}

\boolitem[false]{use}

Only use the \prm{namepart} in constructing the uniquename information if there is a corresponding option \opt{use<namepart>} and that option is true.

\boolitem[false]{base}

The \prm{namepart} is part of the <base> which is the main piece of namepart(s) information which is being disambiguated by uniqueness information. For example, a family name which may be disambiguated by further given names. <base> \prm{namepart}s must occur before any non-<base> \prm{nameparts}. There \emph{must} be at least one <base> \prm{namepart} and \biber will report an error if this is not the case.

\choitem{disambiguation}{none, init, initorfull, full}

The \prm{namepart} will be disambiguated at most by information at the given value. If this option is not present then the default is inferred from the \opt{uniquename} package option setting (see \secref{use:opt:wu}). The <disambiguation> option is ignored for \prm{namepart}s which have the <base> option set to <true> since it is these nameparts which are being disambiguated by the value of the non-base \prm{namepart}s and therefore <disambiguation> does not apply.

\begin{description}
\item[none]~Do not use the \prm{namepart} to perform any name disambiguation
\item[init]~Use only the initials of the \prm{namepart} to perform name disambiguation
\item[initorfull]~Use initials and if necessary the full \prm{namepart} to perform name disambiguation
\item[full]~Use only the full \prm{namepart} to perform name
  disambiguation even if initials would suffice
\end{description}

\end{optionlist*}

\end{ltxsyntax}
%
The default uniquename template is:

\begin{ltxexample}
\DeclareUniquenameTemplate{
  \namepart[use=true, base=true]{prefix}
  \namepart[base=true]{family}
  \namepart{given}
}
\end{ltxexample}
%
This means that the <base> to be disambiguated consists of the <family> namepart, along with any prefix, if the \opt{useprefix} option is true. The disambiguation is performed by adding anything up to the full namepart of any non <base> nameparts in the specification, here just the <given> namepart.

\paragraph{Individual Names (\opt{uniquename})}

Let's start off with some \opt{uniquename} examples. Consider the following data:

\begin{lstlisting}{}
John Doe   2008
Edward Doe 2008
John Smith 2008
Jane Smith 2008
\end{lstlisting}
%
Let's assume we're using an author-year style and set \kvopt{uniquename}{false}. In this case, we would get the following citations:

\begin{lstlisting}{}
Doe 2008a
Doe 2008b
Smith 2008a
Smith 2008b
\end{lstlisting}
%
Since the family names are ambiguous and all works have been published in the same year, an extra letter is appended to the year to disambiguate the citations. Many style guides, however, mandate that the extra letter be used to disambiguate works by the same authors only, not works by different authors with the same family name. In order to disambiguate the author's family name, you are expected to add additional parts of the name, either as initials or in full. This requirement is addressed by the \opt{uniquename} option. Here are the same citations with \kvopt{uniquename}{init}:

\begin{lstlisting}{}
J. Doe 2008
E. Doe 2008
Smith 2008a
Smith 2008b
\end{lstlisting}
%
\kvopt{uniquename}{init} restricts name disambiguation to initials. Since <J. Smith> would still be ambiguous, no additional name parts are added for the <Smiths>. With \kvopt{uniquename}{full}, names are printed in full where required:

\begin{lstlisting}{}
J. Doe 2008
E. Doe 2008
John Smith 2008
Jane Smith 2008
\end{lstlisting}
%
In order to illustrate the difference between \kvopt{uniquename}{init\slash full} and \texttt{allinit\slash allfull}, we need to introduce the notion of a <visible> name. In the following, <visible> names are all names at a position before the \opt{maxnames}\slash \opt{minnames}\slash \opt{uniquelist} truncation point. For example, given this data:

\begin{lstlisting}{}
William Jones/Edward Doe/Jane Smith
John Doe
John Smith
\end{lstlisting}
%
and \kvopt{maxnames}{1}, \kvopt{minnames}{1}, \kvopt{uniquename}{init/full}, we would get the following names in citations:

\begin{lstlisting}{}
Jones et al.
Doe
Smith
\end{lstlisting}
%
When disambiguating names, \kvopt{uniquename}{init/full} only consider the visible names. Since all visible family names are distinct in this example, no further name parts are added. Let's compare that to the output of \kvopt{uniquename}{allinit}:

\begin{lstlisting}{}
Jones et al.
J. Doe
Smith
\end{lstlisting}
%
\texttt{allinit} considers all names in all \bibfield{labelname} lists, including those which are hidden and replaced by <et al.> as the list is truncated. In this example, <John Doe> is disambiguated from <Edward Doe>. Since the ambiguity of the two <Smiths> can't be resolved by adding initials, no initials are added in this case. Now let's compare that to the output of \kvopt{uniquename}{allfull} which also disambiguates <John Smith> from <Jane Smith>:

\begin{lstlisting}{}
Jones et al.
J. Doe
John Smith
\end{lstlisting}
%
The options \kvopt{uniquename}{mininit/minfull} are similar to \texttt{init\slash full} in that they only consider visible names, but they perform minimal disambiguation. That is, they will disambiguate individual names only if they occur in identical lists of base nameparts (for the concept of <base> nameparts, see \cmd{DeclareUniquenameTemplate in \secref{aut:cav:amb}}). Consider the following data:

\begin{lstlisting}{}
John Doe/William Jones
Edward Doe/William Jones
John Smith/William Edwards
Edward Smith/Allan Johnson
\end{lstlisting}
%
With \kvopt{uniquename}{init/full}, we would get:

\begin{lstlisting}{}
J. Doe and Jones
E. Doe and Jones
J. Smith and Edwards
E. Smith and Johnson
\end{lstlisting}
%
With \kvopt{uniquename}{mininit/minfull}:

\begin{lstlisting}{}
J. Doe and Jones
E. Doe and Jones
Smith and Edwards
Smith and Johnson
\end{lstlisting}
%
The <Smiths> are not disambiguated because the visible name lists are not ambiguous and the \opt{mininit/minfull} options serve to disambiguate names occurring in identical base namepart lists only. Another way of looking at this is that they globally disambiguate base namepart lists. When it comes to ambiguous lists, note that a truncated list is considered to be distinct from an untruncated one even if the visible names are identical. For example, consider the following data:

\begin{lstlisting}{}
John Doe/William Jones
Edward Doe
\end{lstlisting}
%
With \kvopt{maxnames}{1}, \kvopt{uniquename}{init/full}, we would get:

\begin{lstlisting}{}
J. Doe et al.
E. Doe
\end{lstlisting}
%
With \kvopt{uniquename}{mininit/minfull}:

\begin{lstlisting}{}
Doe et al.
Doe
\end{lstlisting}
%
Because the lists differ in the <et al.>, the names are not disambiguated.

\paragraph{Lists of Names (\opt{uniquelist})}

Ambiguity is also an issue with name lists. If the \bibfield{labelname} list is truncated by the \opt{maxnames}\slash \opt{minnames} options, it may become ambiguous. This type of ambiguity is addressed by the \opt{uniquelist} option. Consider the following data:

\begin{lstlisting}{}
Doe/Jones/Smith   2005
Smith/Johnson/Doe 2005
Smith/Doe/Edwards 2005
Smith/Doe/Jones   2005
\end{lstlisting}
%
Many author-year styles truncate long author/editor lists in citations. For example, with \kvopt{maxnames}{1} we would get:

\begin{lstlisting}{}
Doe et al. 2005
Smith et al. 2005a
Smith et al. 2005b
Smith et al. 2005c
\end{lstlisting}
%
Since the authors are ambiguous after truncation, the extra letter is added to the year to ensure unique citations. Here again, many style guides mandate that the extra letter be used to disambiguate works by the same authors only. In order to disambiguate author lists, you are usually required to add more names, exceeding the \opt{maxnames}\slash \opt{minnames} truncation point. The \opt{uniquelist} feature addresses this requirement. With \kvopt{uniquelist}{true}, we would get:

\begin{lstlisting}{}
Doe et al. 2005
Smith, Johnson et al. 2005
Smith, Doe and Edwards 2005
Smith, Doe and Jones 2005
\end{lstlisting}
%
The \opt{uniquelist} option overrides \opt{maxnames}\slash \opt{minnames} on a per-entry basis. Essentially, what happens is that the <et al.> part of the citation is expanded to the point of no ambiguity---but no further than that. \opt{uniquelist} may also be combined with \opt{uniquename}. Consider the following data:

\begin{lstlisting}{}
John Doe/Allan Johnson/William Jones  2009
John Doe/Edward Johnson/William Jones 2009
John Doe/Jane Smith/William Jones     2009
John Doe/John Smith/William Jones     2009
John Doe/John Edwards/William Jones   2009
John Doe/John Edwards/Jack Johnson    2009
\end{lstlisting}
%
With \kvopt{maxnames}{1}:

\begin{lstlisting}{}
Doe et al. 2009a
Doe et al. 2009b
Doe et al. 2009c
Doe et al. 2009d
Doe et al. 2009e
Doe et al. 2009f
\end{lstlisting}
%
With \kvopt{maxnames}{1}, \kvopt{uniquename}{full}, \kvopt{uniquelist}{true}:

\begin{lstlisting}{}
Doe, A. Johnson et al. 2009
Doe, E. Johnson et al. 2009
Doe, Jane Smith et al. 2009
Doe, John Smith et al. 2009
Doe, Edwards and Jones 2009
Doe, Edwards and Johnson 2009
\end{lstlisting}
%
With \kvopt{uniquelist}{minyear}, list disambiguation only happens if the visible list is identical to another visible list with the same \bibfield{labelyear}. This is useful for author-year styles which only require that the citation as a whole be unique, but do not guarantee unambiguous authorship information in citations. This mode is conceptually related to \kvopt{uniquename}{mininit/minfull}. Consider this example:

\begin{lstlisting}{}
Smith/Jones   2000
Smith/Johnson 2001
\end{lstlisting}
%
With \kvopt{maxnames}{1} and \kvopt{uniquelist}{true}, we would get:

\begin{lstlisting}{}
Smith and Jones 2000
Smith and Johnson 2001
\end{lstlisting}
%
With \kvopt{uniquelist}{minyear}:

\begin{lstlisting}{}
Smith et al. 2000
Smith et al. 2001
\end{lstlisting}
%
With \kvopt{uniquelist}{minyear}, it is not clear that the authors are different for the two works but the citations as a whole are still unambiguous since the year is different. In contrast to that, \kvopt{uniquelist}{true} disambiguates the authorship even if this information is not required to uniquely locate the works in the bibliography. Let's consider another example:

\begin{lstlisting}{}
Vogel/Beast/Garble/Rook  2000
Vogel/Beast/Tremble/Bite 2000
Vogel/Beast/Acid/Squeeze 2001
\end{lstlisting}
%
With \kvopt{maxnames}{3}, \kvopt{minnames}{1}, \kvopt{uniquelist}{true}, we would get:

\begin{lstlisting}{}
Vogel, Beast, Garble et al. 2000
Vogel, Beast, Tremble et al. 2000
Vogel, Beast, Acid et al. 2001
\end{lstlisting}
%
With \kvopt{uniquelist}{minyear}:

\begin{lstlisting}{}
Vogel, Beast, Garble et al. 2000
Vogel, Beast, Tremble et al. 2000
Vogel et al. 2001
\end{lstlisting}
%
In the last citation, \kvopt{uniquelist}{minyear} does not override \opt{maxnames}\slash \opt{minnames} as the citation does not need disambiguating from the other two because the year is different.

\subsubsection{Trackers in Floats and \acr{TOC}/\acr{LOT}/\acr{LOF}}
\label{aut:cav:flt}

If a citation is given in a float (typically in the caption of a figure or table), scholarly back references like <ibidem> or back references based on the page tracker get ambiguous because floats are objects which are (physically and logically) placed outside the flow of text, hence the logic of such references applies poorly to them. To avoid any such ambiguities, the citation and page trackers are temporarily disabled in all floats unless explicitly requested with \opt{trackfloats}. In addition to that, these trackers plus the back reference tracker (\opt{backref}) are temporarily disabled in the table of contents, the list of figures, and the list of tables.

\subsubsection{Mixing Programming Interfaces}
\label{aut:cav:mif}

The \biblatex package provides two main programming interfaces for style authors. The \cmd{DeclareBibliographyDriver} command, which defines a handler for an entry type, is typically used in \file{bbx} files. \cmd{DeclareCiteCommand}, which defines a new citation command, is typically used in \file{cbx} files. However, in some cases it is convenient to mix these two interfaces. For example, the \cmd{fullcite} command prints a verbose citation similar to the full bibliography entry. It is essentially defined as follows:

\begin{ltxexample}
\DeclareCiteCommand{\fullcite}
  {...}
  {<<\usedriver>>{...}{<<\thefield{entrytype}>>}}
  {...}
  {...}
\end{ltxexample}
%
As you can see, the core code which prints the citations simply executes the bibliography driver defined with \cmd{DeclareBibliographyDriver} for the type of the current entry. When writing a citation style for a verbose citation scheme, it is often convenient to use the following structure:

\begin{ltxexample}
\ProvidesFile{example.cbx}[2007/06/09 v1.0 biblatex citation style]

\DeclareCiteCommand{\cite}
  {...}
  {<<\usedriver>>{...}{<<cite:\thefield{entrytype}>>}}
  {...}
  {...}

\DeclareBibliographyDriver{<<cite:article>>}{...}
\DeclareBibliographyDriver{<<cite:book>>}{...}
\DeclareBibliographyDriver{<<cite:inbook>>}{...}
...
\end{ltxexample}
%
Another case in which mixing interfaces is helpful are styles using cross"=references within the bibliography. For example, when printing an \bibtype{incollection} entry, the data inherited from the \bibtype{collection} parent entry would be replaced by a short pointer to the respective parent entry:

\begin{enumerate}
\renewcommand*\labelenumi{[\theenumi]}
\setlength{\leftskip}{0.5em}
\item Audrey Author: \emph{Title of article}. In: [\textln{2}], pp.~134--165.
\item Edward Editor, ed.: \emph{Title of collection}. Publisher: Location, 1995.
\end{enumerate}

One way to implement such cross"=references within the bibliography is to think of them as citations which use the value of the \bibfield{xref} or \bibfield{crossref} field as the entry key. Here is an example:

\begin{ltxexample}
\ProvidesFile{example.bbx}[2007/06/09 v1.0 biblatex bibliography style]

\DeclareCiteCommand{<<\bbx@xref>>}
  {}
  {...}% code for cross-references
  {}
  {}

\DeclareBibliographyDriver{incollection}{%
  ...
  \iffieldundef{xref}
    {...}% code if no cross-reference
    {<<\bbx@xref>>{<<\thefield{xref}>>}}%
  ...
}
\end{ltxexample}
%
When defining \cmd{bbx@xref}, the \prm{precode}, \prm{postcode}, and \prm{sepcode} arguments of \cmd{DeclareCiteCommand} are left empty in the above example because they will not be used anyway. The cross"=reference is printed by the \prm{loopcode} of \cmd{bbx@xref}. For further details on the \bibfield{xref} field, refer to \secref{bib:fld:spc} and to the hints in \secref{bib:cav:ref}. Also see the \cmd{iffieldxref}, \cmd{iflistxref}, and \cmd{ifnamexref} tests in \secref{aut:aux:tst}. The above could also be implemented using the \cmd{entrydata} command from \secref{aut:bib:dat}:

\begin{ltxexample}
\ProvidesFile{example.bbx}[2007/06/09 v1.0 biblatex bibliography style]

\DeclareBibliographyDriver{incollection}{%
  ...
  \iffieldundef{xref}
    {...}% code if no cross-reference
    {<<\entrydata>>{<<\thefield{xref}>>}{%
      % code for cross-references
      ...
    }}%
  ...
}
\end{ltxexample}

\subsubsection{Using the Punctuation Tracker}
\label{aut:cav:pct}

\paragraph{The Basics}

There is one fundamental principle style authors should keep in mind when designing a bibliography driver: block and unit punctuation is handled asynchronously. This is best explained by way of example. Consider the following code snippet:

\begin{ltxexample}
\printfield{title}%
\newunit
\printfield{edition}%
\newunit
\printfield{note}%
\end{ltxexample}
%
If there is no \bibfield{edition} field, this piece of code will not print:

\begin{lstlisting}[style=highlight]{}
Title. . Note
\end{lstlisting}
%
but rather:

\begin{lstlisting}[style=highlight]{}
Title. Note
\end{lstlisting}
%
because the unit punctuation tracker works asynchronously. \cmd{newunit} will not print the unit punctuation immediately. It merely records a unit boundary and puts \cmd{newunitpunct} on the punctuation buffer. This buffer will be handled by \emph{subsequent} \cmd{printfield}, \cmd{printlist}, or similar commands but only if the respective field or list is defined. Commands like \cmd{printfield} will consider three factors prior to inserting any block or unit punctuation:

\begin{itemize}
\item Has a new unit/block been requested at all?\par
= Is there any preceding \cmd{newunit} or \cmd{newblock} command?
\item Did the preceding commands print anything?\par
= Is there any preceding \cmd{printfield} or similar command?\par
= Did this command actually print anything?\par
\item Are we about to print anything now?\par
= Is the field/list to be processed now defined?
\end{itemize}
%
Block and unit punctuation will only be inserted if \emph{all} of these conditions apply. Let's reconsider the above example:

\begin{ltxexample}
\printfield{title}%
\newunit
\printfield{edition}%
\newunit
\printfield{note}%
\end{ltxexample}
%
Here's what happens if the \bibfield{edition} field is undefined. The first \cmd{printfield} command prints the title and sets an internal <new~text> flag. The first \cmd{newunit} sets an internal <new~unit> flag. No punctuation has been printed at this point. The second \cmd{printfield} does nothing because the \bibfield{edition} field is undefined. The next \cmd{newunit} command sets the internal flag <new unit> again. Still no punctuation has been printed. The third \cmd{printfield} checks if the \bibfield{note} field is defined. If so, it looks at the <new~text> and <new~unit> flags. If both are set, it inserts the punctuation buffer before printing the note. It then clears the <new~unit> flag and sets the <new~text> flag again.

This may all sound more complicated than it is. In practice, it means that it is possible to write large parts of a bibliography driver in a sequential way. The advantage of this approach becomes obvious when trying to write the above code without using the punctuation tracker. Such an attempt will lead to a rather convoluted set of \cmd{iffieldundef} tests required to check for all possible field combinations (note that the code below handles three fields; a typical driver may need to cater for some two dozen fields):

\begin{ltxexample}
\iffieldundef{title}%
  {\iffieldundef{edition}
     {\printfield{note}}
     {\printfield{edition}%
      \iffieldundef{note}%
	{}
	{. \printfield{note}}}}
  {\printfield{title}%
   \iffieldundef{edition}
     {}
     {. \printfield{edition}}%
   \iffieldundef{note}
     {}
     {. \printfield{note}}}%
\end{ltxexample}

\paragraph{Common Mistakes}

It is a fairly common misconception to think of the unit punctuation as something that is handled synchronously. This typically causes problems if the driver includes any literal text. Consider this erroneous code snippet which will generate misplaced unit punctuation:

\begin{ltxexample}
\printfield{title}%
\newunit
<<(>>\printfield{series} \printfield{number}<<)>>%
\end{ltxexample}
%
This code will yield the following result:

\begin{lstlisting}[style=highlight]{}
Title <<(.>> Series Number<<)>>
\end{lstlisting}
%
Here's what happens. The first \cmd{printfield} prints the title. Then \cmd{newunit} marks a unit boundary but does not print anything. The unit punctuation is printed by the \emph{next} \cmd{printfield} command. That's the asynchronous part mentioned before. However, the opening parenthesis is printed immediately before the next \cmd{printfield} inserts the unit punctuation, leading to a misplaced period. When inserting \emph{any} literal text such as parentheses (including those printed by commands such as \cmd{bibopenparen} and \cmd{mkbibparens}), always wrap the text in a \cmd{printtext} command. For the punctuation tracker to work as expected, it needs to know about all literal text inserted by a driver. This is what \cmd{printtext} is all about. \cmd{printtext} interfaces with the punctuation tracker and ensures that the punctuation buffer is inserted before the literal text gets printed. It also sets the internal <new~text> flag. Note there is in fact a third piece of literal text in this example: the space after |\printfield{series}|. In the corrected example, we will use the punctuation tracker to handle that space.

\begin{ltxexample}
\printfield{title}%
\newunit
<<\printtext{(}>>%
\printfield{series}%
<<\setunit*{\addspace}>>%
\printfield{number}%
<<\printtext{)}>>%
\end{ltxexample}
%
While the above code will work as expected, the recommended way to handle parentheses, quotes, and other things which enclose more than one field, is to define a field format:

\begin{ltxexample}
\DeclareFieldFormat{<<parens>>}{<<\mkbibparens{#1}>>}
\end{ltxexample}
%
Field formats may be used with both \cmd{printfield} and \cmd{printtext}, hence we can use them to enclose several fields in a single pair of parentheses:

\begin{ltxexample}
<<\printtext[parens]{>>%
  \printfield{series}%
  \setunit*{\addspace}%
  \printfield{number}%
<<}>>%
\end{ltxexample}
%
We still need to handle cases in which there is no series information at all, so let's improve the code some more:

\begin{ltxexample}
<<\iffieldundef{series}>>
  {}
  {\printtext[parens]{%
     \printfield{series}%
     \setunit*{\addspace}%
     \printfield{number}}}%
\end{ltxexample}
%
One final hint: localisation strings are not literal text as far as the punctuation tracker is concerned. Since \cmd{bibstring} and similar commands interface with the punctuation tracker, there is no need to wrap them in a \cmd{printtext} command.

\paragraph{Advanced Usage}

The punctuation tracker may also be used to handle more complex scenarios. For example, suppose that we want the fields \bibfield{location}, \bibfield{publisher}, and \bibfield{year} to be rendered in one of the following formats, depending on the available data:

\begin{ltxexample}
...text<<. Location: Publisher, Year.>> Text...
...text<<. Location: Publisher.>> Text...
...text<<. Location: Year.>> Text...
...text<<. Publisher, Year.>> Text...
...text<<. Location.>> Text...
...text<<. Publisher.>> Text...
...text<<. Year.>> Text...
\end{ltxexample}
%
This problem can be solved with a rather convoluted set of \cmd{iflistundef} and \cmd{iffieldundef} tests which check for all possible field combinations:

\begin{ltxexample}
\iflistundef{location}
  {\iflistundef{publisher}
     {\printfield{year}}
     {\printlist{publisher}%
      \iffieldundef{year}
        {}
        {, \printfield{year}}}}
  {\printlist{location}%
   \iflistundef{publisher}%
     {\iffieldundef{year}
        {}
        {: \printfield{year}}}
     {: \printlist{publisher}%
      \iffieldundef{year}
        {}
        {, \printfield{year}}}}%
\end{ltxexample}
%
The above could be written in a somewhat more readable way by employing \cmd{ifthenelse} and the boolean operators discussed in \secref{aut:aux:ife}. The approach would still be essentially the same. However, it may also be written sequentially:

\begin{ltxexample}
\newunit
\printlist{location}%
\setunit*{\addcolon\space}%
\printlist{publisher}%
\setunit*{\addcomma\space}%
\printfield{year}%
\newunit
\end{ltxexample}
%
In practice, you will often use a combination of explicit tests and the implicit tests performed by the punctuation tracker. For example, consider the following format (note the punctuation after the location if there is no publisher):

\begin{ltxexample}
...text. Location: Publisher, Year. Text...
...text. Location: Publisher. Text...
...text<<. Location, Year.>> Text...
...text. Publisher, Year. Text...
...text. Location. Text...
...text. Publisher. Text...
...text. Year. Text...
\end{ltxexample}
%
This can be handled by the following code:

\begin{ltxexample}
\newunit
\printlist{location}%
\iflistundef{publisher}
  {\setunit*{\addcomma\space}}
  {\setunit*{\addcolon\space}}%
\printlist{publisher}%
\setunit*{\addcomma\space}%
\printfield{year}%
\newunit
\end{ltxexample}
%
Since the punctuation after the location is special if there is no publisher, we need one \cmd{iflistundef} test to catch this case. Everything else is handled by the punctuation tracker.

\subsubsection{Custom Localization Modules}
\label{aut:cav:lng}

Style guides may include provisions as to how strings like <edition> should be abbreviated or they may mandate certain fixed expressions. For example, the \acr{mla} style guide requires authors to use the term <Works~Cited> rather than <Bibliography> or <References> in the heading of the bibliography. Localization commands such as \cmd{DefineBibliographyStrings} from \secref{use:lng} may indeed be used in \file{cbx} and \file{bbx} files to handle such cases. However, overloading style files with translations is rather inconvenient. This is where \cmd{DeclareLanguageMapping} from \secref{aut:lng:cmd} comes into play. This command maps an \file{lbx} file with alternative translations to a \sty{babel}/\sty{polyglossia} language. For example, you could create a file named \path{french-humanities.lbx} which provides French translations adapted for use in the humanities and map it to the \sty{babel}/\sty{polyglossia} language \texttt{french} in the preamble or in the configuration file:

\begin{ltxexample}
\DeclareLanguageMapping{french}{french-humanities}
\end{ltxexample}
%
If the document language is set to \texttt{french}, \path{french-humanities.lbx} will replace \path{french.lbx}. Coming back to the \acr{mla} example mentioned above, an \acr{mla} style may come with an \path{american-mla.lbx} file to provide strings which comply with the \acr{mla} style guide. It would declare the following mapping in the \file{cbx} and/or \file{bbx} file:

\begin{ltxexample}
\DeclareLanguageMapping{american}{american-mla}
\end{ltxexample}
%
Use \cmd{DeclareLanguageMappingSuffix} (see \secref{aut:lng:cmd}) to define such a mapping for all languages.

Since the alternative \file{lbx} file can inherit strings from the standard \path{american.lbx} module, \path{american-mla.lbx} may be as short as this:

\begin{ltxexample}
\ProvidesFile{american-mla.lbx}[2008/10/01 v1.0 biblatex localization]
<<\InheritBibliographyExtras>>{<<american>>}
\DeclareBibliographyStrings{%
  <<inherit>>          = {<<american>>},
  bibliography     = {{Works Cited}{Works Cited}},
  references       = {{Works Cited}{Works Cited}},
}
\endinput
\end{ltxexample}
%
Alternative \file{lbx} files must ensure that the localisation module is complete. They should do so by inheriting data from the corresponding standard module. If the language \texttt{american} is mapped to \path{american-mla.lbx}, \biblatex will not load \path{american.lbx} unless this module is requested explicitly. In the above example, inheriting <strings> and <extras> will cause \biblatex to load \path{american.lbx} before applying the modifications in \path{american-mla.lbx}.

Note that \cmd{DeclareLanguageMapping} is not intended to handle language variants (\eg American English vs. British English) or \sty{babel}/\sty{polyglossia} language aliases (\eg \texttt{USenglish} vs. \texttt{american}). For example, \sty{babel}/\sty{polyglossia} offers the \texttt{USenglish} option which is similar to \texttt{american}. Therefore, \biblatex comes with an \path{USenglish.lbx} file which simply inherits all data from \path{american.lbx} (which in turn gets the <strings> from \path{english.lbx}). In other words, the mapping of language variants and \sty{babel}/\sty{polyglossia} language aliases happens on the file level, the point being that \biblatex's language support can be extended simply by adding additional \file{lbx} files. There is no need for centralized mapping. If you need support for, say, Portuguese (babel/polyglossia: \file{portuges}), you create a file named \path{portuges.lbx}. If \sty{babel}/\sty{polyglossia} offered an alias named \texttt{brasil}, you would create \path{brasil.lbx} and inherit the data from \path{portuges.lbx}. In contrast to that, the point of \cmd{DeclareLanguageMapping} is handling \emph{stylistic} variants like <humanities vs. natural sciences> or <\acr{mla} vs. \acr{apa}> etc. which will typically be built on top of existing \file{lbx} files.

\subsubsection{Grouping}
\label{aut:cav:grp}

In a citation or bibliography style, you may need to set flags or store certain values for later use. In this case, it is crucial to understand the basic grouping structure imposed by this package. As a rule of thumb, you are working in a large group whenever author commands such as those discussed in \secref{aut:aux} are available because the author interface of this package is only enabled locally. If any bibliographic data is available, there is at least one additional group. Here are some general rules:

\begin{itemize}

\item The entire list of references printed by \cmd{printbibliography} and similar commands is processed in a group. Each entry in the list is processed in an additional group which encloses the \prm{item code} of \cmd{defbibenvironment} as well as all driver code.

\item The entire bibliography list printed by \cmd{printbiblist} is processed in a group. Each entry in the list is processed in an additional group which encloses the \prm{item code} of \cmd{defbibenvironment} as well as all driver code.

\item All citation commands defined with \cmd{DeclareCiteCommand} are processed in a group holding the complete citation code consisting of the \prm{precode}, \prm{sepcode}, \prm{loopcode}, and \prm{postcode} arguments. The \prm{loopcode} is enclosed in an additional group every time it is executed. If any \prm{wrapper} code has been specified, the entire unit consisting of the wrapper code and the citation code is wrapped in an additional group.

\item In addition to the grouping imposed by all backend commands defined with \cmd{DeclareCiteCommand}, all <autocite> and <multicite> definitions imply an additional group.

\item \cmd{printfile}, \cmd{printtext}, \cmd{printfield}, \cmd{printlist}, and \cmd{printnames} form groups. This implies that all formatting directives will be processed within a group of their own.

\item All \file{lbx} files are loaded and processed in a group. If an \file{lbx} file contains any code which is not part of \cmd{DeclareBibliographyExtras}, the definitions must be global.

\end{itemize}

Note that using \cmd{aftergroup} in citation and bibliography styles is unreliable because the precise number of groups employed in a certain context may change in future versions of this package. If the above list states that something is processed in a group, this means that there is \emph{at least one} group. There may also be several nested ones.

\subsubsection{Namespaces}
\label{aut:cav:nam}

In order to minimize the risk of name clashes, \latex packages typically prefix the names of internal macros with a short string specific to the package. For example, if the \sty{foobar} package requires a macro for internal use, it would typically be called \cmd{FB@macro} or \cmd{foo@macro} rather than \cmd{macro} or \cmd{@macro}. Here is a list of the prefixes used or recommended by \biblatex:

\begin{marglist}

\item[\texttt{blx}] All macros with names like \cmd{blx@name} are strictly reserved for internal use. This also applies to counter names, length registers, boolean switches, and so on. These macros may be altered in backwards"=incompatible ways, they may be renamed or even removed at any time without further notice. Such changes will not even be mentioned in the revision history or the release notes. In short: never use any macros with the string \texttt{blx} in their name in any styles.

\item[\texttt{abx}] Macros prefixed with \texttt{abx} are also internal macros but they are fairly stable. It is always preferable to use the facilities provided by the official author interface, but there may be cases in which using an \texttt{abx} macro is convenient.

\item[\texttt{bbx}] This is the recommended prefix for internal macros defined in bibliography styles.

\item[\texttt{cbx}] This is the recommended prefix for internal macros defined in citation styles.

\item[\texttt{lbx}] This is the recommended base prefix for internal macros defined in localisation modules. The localisation module should add a second prefix to specify the language. For example, an internal macro defined by the Spanish localisation module would be named \cmd{lbx@es@macro}.

\end{marglist}