What should we do with our vignettes?

[egillax]

Currently our vignettes are automatically built into html both for website and package. Additionally there are manually built pdf vignettes (built with this). This manual step is not ideal. It causes the pdf vignettes to easily be out of date and old vignettes are not removed. Recently I removed a bunch of old ones.

Further for example roxygen does not recommend anymore that built vignettes are stored in a package.

I think there are two ways forward that would simplify matters:

1. Switch completely over to html, people can the access our vignettes either on the website or with browseVignettes("PatientLevelPrediction") on their local computers.

2. Automatically build the pdf vignettes. This can be accomplished by switching the vignette engine to rmarkdown v2 (knitr::rmarkdown) instead of v1 (knitr::knitr).

I would be in favor of 1. Personally I never open the pdf's. Looking at which packages store pdf's in my R-lib folder, it's about ~80 out of ~440 packages.

But this needs discussion. @jreps @lhjohn @rossdwilliams @ChungsooKim

[ChungsooKim]

Hi @egillax, sorry for the late response. I also prefer option 1.

[egillax]

This has been mentioned at OHDSI europe PLP working group meeting as well as the online PLP working group meeting. People didn't seem to have strong opinions and seemed to be fine with moving to html for vignettes. Next release will be submitted to CRAN and we'll aim at not using pdf's anymore.