LtC Term Issue: Identifier

[wouteraddink]

Submitter

Wouter Addink @wouteraddink

Type of Issue

Improvement

Suggested Action

Add term

Justification

I would like to know for an resolvable identifier what data format it returns.

Term Type

Class

Term

Identifier

Term Identifier

No response

Description

I would like to know the format (mimetype) that a resolvable identifier returns. In openDS we use dcterms:format for this. For example for a Getty term I could include an ID that resolves to a HTML page for the Getty term and another ID that resolves to a JSON. If I implement a tool I would use this information to use the ID that returns JSON and parse it. So my suggestion is to add dcterms:format to the identifier class.

openDS has a few additional terms for identifiers that may also be interesting to evaluate for addition:
ods:isPartOfLabel, ods:isBarcodeOrNFCProperty, ods:isIDPersistent

Additional Comments

No response

[ben-norton]

@wouteraddink The format of a response is commonly (and should be) included in the header of a response as Content-Type. See the following endpoint for an example: https://geoapis.io/api/v1/services/spatial/point/union?lat=35.7795&lon=-78.6381
The usefulness of the content-type is programmatic, unless there's a use case I'm not familiar with. Also, it only applies to a subset of identifiers that are resolvable URIs. I understand the logic, but in practice, when would a format term be useful if the content-type is included in the header of the response?

[wouteraddink]

it would enable a machine to decide beforehand whether it is worthwhile to request a response or to pick a preferred endpoint if the identifier resolves to multiple. If the response is in a format the machine cannot handle a request would not be needed. This matters in an interconnected web of data with billions of data objects but not so much for siloed local implementations.

[jbstatgen]

This aspect we didn't discuss yesterday, and ruminating about the question over night, I agree with @wouteraddink .

Servers not needing to negotiating, just to find out that the target (format) is of no use or interest to them, will reduce the energy footprint of the internet (and traffic). It's an energy footprint that is far from sustainable along most dimensions, including mining, CO2, nuclear waste, biodiversity loss, water needs, human, multispecies & nature rights, supply chains, ...

All of this is a problem already now, however, if we are successful with linked data, then it will explode the internet.