cran-task-views/MixedModels

list of packages to watch/exclude

Opened this issue · 5 comments

Consider adding to the repo, in some sensible format:

  • a list of archived or not-yet-on-CRAN packages to watch/check periodically for (re)inclusion
  • a list of packages that we decide are out of scope, with brief justification (so that this is documented and doesn't need to be rediscovered if someone asks why the package isn't included)

Good idea!

Just a quick comment for interesting packages to watch: If these are on GitHub and you think they are good quality already, you can include them with the r github("user/repo") tag. This shouldn't be overused (after all it's a task view for CRAN not GitHub) but it's certainly ok for some packages of good quality.

should we remove archived packages from the task view during the interval that they are archived?

It's a judgment call. If we know/think they're probably going to come back, it's easier to leave them. If we wanted we could set up some kind of rule of thumb like "if archived for >3 months drop from the TV", or "remove archived packages from the TV every 3 months"; the former rule is package-focused, the latter is TV-focused (i.e. we would review the TV for archived packages every 3 months and remove any packages that were currently archived).

It's a little bit of work but we could automate the check for "packages archived > 3 months" since the date of archiving can be scraped from the CRAN page ...

This is supported from the CRAN side now, albeit only semi-automatically:

  • When a package gets archived, it's usually fine to leave it in the task view. Only if the archival is due to persistent and/or substantial problems, I would remove it from the task view right away.
  • After about 60 days CRAN will create an issue for the task view to notify the task view maintainers about the problem, see, e.g., cran-task-views/Econometrics#12
  • After about 100 days CRAN will follow up and ask for remove from the task view.

At the moment the 60 days and 100 days are not yet enforced properly because we are still triggering the process by hand in order to catch potential problems and hiccups. We hope to switch to cronjobs in the not-so-distant future, though.