comment from Alex

[WandrilleD]

• There were issues finding the room, could have been better indicated in email, or better, on course website, or even physically (signposts?)
• I might be a good idea to mention other SIB courses that relate to the interests of participants
• There no longer seems to be a link to support for installing R/RStudio on the course page(?) Only hyperlinks to root R and RStudioWould it be nice to have a step-by-step installation guide? Perhaps a FAQ with support for participants w/ limited install rights / forced RServer usage (bc this happens pretty often)?
• It might have been nice to have a brief pause to check that everyone had things installed properly?

Question: how to change R language when system language is not the desired one? (Good point for an FAQ)
The red blobs icon is a bit overwhelming!
Earlier mention of .RData / .RHistory files?
Add "stretch goals" for those who finish quickly.
e.g. Projects management: create a separate project and switch between
When discussing this topic, point out an online reference: https://cran.r-project.org/doc/manuals/r-release/R-lang.html#Argument-matching
Note: Some RStudio installations automatically call readr code for importing datasets from the File panel.
Dead link: ecomodel / TidyData
table(): show you can give dimnames to it (prettier!)
Troubleshooting when writing files: note you sometimes need to refresh the file pane to see the file appear; if it doesn't, check the working directory!
Mention file.path(), esp. for those using windows
Recurring question: why make factors?
I think it's a bad idea to rely solely on factors as numerics for indexing colors, as levels can VERY easily get mixed up, esp. if you don’t mention reference levels (eg in the case of KO vs WT!). Colour palettes can be indexed using strings, which is more robust.
Point out why sometimes we don't use the quotes around names (library, subset, formula, ...)? Bc the co-existence of 2 types of grammar is confusing if not made explicit.
Noticed frequent issues with the concept of indexing, might be worth insisting on that
op <- par(no.readonly=TRUE) ; then later par(op) for resetting par
Insist on dev.off() after having an error while making a pdf/png/etc, I saw many not noticing that they still had devices open
Might be better practice to forego the niceties of the ~ operator in boxplots, points etc to force everyone to understand what is being done with the data