Document how to update infrastructure

[zkamvar]

Something that's come up and is quite relevant is that a lot of the instructions work perfectly for users who are brand new to R and who might not have it installed on their system.

Background

One of the participants actually had a good experience with the jekyll lesson template the first time they used it, but after Jekyll updated, they were left with a non-working version on their machine and they never quite got the hang of it afterwards. R is relatively good at avoiding these issues by sequestering libraries in version-specific paths by default, but there the minor version still updates every year in April and we need to prepare people for that situation that could be disruptive.

Profiles to be aware of

The issue comes down to situations where people already have R or pandoc installed on their machines. They fall into a few categories:

• People who use R regularly, but only update whenever forced to
• People who don't use R regularly and still have R version 3.0.3 installed
• People who use R regularly and refuse to update because their thesis work depends on the setup on their machine

Solutions/tools to consider

MacOS

The installation for R on MacOS is normally via the installer, but some people use brew to install R, which can lead to difficult situations sometimes. For maintaining several versions of R in parallel, there is RSwitch: https://rud.is/rswitch/

Windows

Updating R can be accomplished with the {installr} package: https://talgalili.github.io/installr/

R itself: https://web.archive.org/web/20200929093154/http://www.dahl-jacobsen.dk/data/2019/05/11/multiple-r-installations-on-windows/

with RStudio: https://community.rstudio.com/t/install-multiple-r-versions-os-windows-10-and-use-them-to-cover-specific-reasons-inside-rstudiodesktop/36271

Linux

RStudio has some resources for installing multiple versions: https://docs.rstudio.com/resources/install-r/