cheat sheet generator

[thibautjombart]

This follows from #58 (comment)

@batpigandme pointed out the Rstudio resource:
https://www.rstudio.com/resources/cheatsheets/how-to-contribute-a-cheatsheet/

@jsta pointed out his previous work with LaTeX (Beamer):
https://github.com/jsta/imagemagick_cheatsheet

as well as:
https://github.com/odeleongt/flexdashboard-poster
which uses rmarkdown.

[batpigandme]

I don't know if/how it would integrate at all with R, but Cheatography create seems to help people generate cheat sheets with a consistent style (one not terribly dissimilar to that of some of the RStudio ones (though obviously inferior).

[thibautjombart]

@batpigandme This looks nice and a bit simpler / easier to generalise. The Rstudio cheatsheets are awesome but I don't think we can easily generalise that, at least for most of them. Except for the R markdown one which is graphically simpler.

[thibautjombart]

For some packages representing workflows as graphs might provide a useful summary. I think @richfitz mentioned this has been tried a few times already?

[karthik]

@thibautjombart
One of rOpenSci's new summer interns (Diana Ly), who is starting on June 5th (and unfortunately wont be at unconf) has this as her internship project.

Her idea is that a package developer will create and maintain the cheatsheet in markdown, with yaml keeping track of version number and some cosmetic details. The package will generate pretty versions that can be printed and shared.

While it is beyond her abilities right now, we were also thinking of testing (rendering all the code in the cheatsheet after each package update) to make sure all the commands still work and that the whole process is not too manual.

I'm sure she would welcome collaborators and mentors.

[thibautjombart]

@karthik sounds great! Do you think we can break down the project into components small enough that we can produce something during the unconf that will be useful, easily re-usable in, and not getting in the way of her project?

I'm happy to help with whatever, but one thing that comes back every time somebody asks me 'What does this package do?' is me drawing a (rather ugly) graph of typical workflows. I have to say this applies to new packages of RECON which are meant to each have a narrow focus, which makes things simpler, but look at the 'What does it do' sections for incidence or epicontacts.

Here's an ugly example:

[stefaniebutland]

Repo: https://github.com/ropenscilabs/rcheatsheet

[thibautjombart]

Sweet! :)