The apache R-universe is managed here, but it seems unclear at this time whether it is for release builds or edge builds. https://github.com/r-universe-org/apache.r-universe.dev

If it is for edge builds, R-multiverse may be more suitable for installing release builds. Or it may be worth updating the apache universe to show the same version as the CRAN release.

Hmmm. Fair point. Let's wait for the arrow devs to weigh in. It may be easiest to keep the current managed repo and just remove the "development" language.

Thanks for making this PR @grantmcdermott! Just to confirm, we do use it for release builds and not dev builds. This addition is still welcome though, just with that bit updated.

Thanks for the feedback all. I've actioned the requested changes (modulo the uppercase suggestion). Let me know if you want changes to the wording.

Some other points/thoughts:

• I didn't mention anything about R-universe potentially guaranteeing a richer feature set because of CRAN restrictions (e.g., like we've experienced on MacOS with certain releases in the past w.r.t. S3 access etc.) I feel that this is a potentially important point that users should know about. But you all have better visibility to this stuff on the dev side.

• Looking over the README, I feel like there's quite a bit of unnecessary repetition, made worse by inconsistent referencing the the upstream C++ docs and the R package website. For example...

  To learn more about the Apache Arrow project, see the parent documentation of the Arrow Project. The Arrow project provides functionality for a wide range of data analysis tasks to store, process and move data fast. See the read/write article to learn about reading and writing data files, data wrangling to learn how to use dplyr syntax with arrow objects, and the function documentation for a full list of supported functions within dplyr queries.

  (note the mixed links to the upstream docs and the individual R vignettes)
  ... is followed by a dedicated "What can the arrow package do?" only a few paragraphs later. Better IMO to consolidate this information in one place and generally prioritize the R package documentation/website above the upstream docs. I'm happy to take a crack at consolidation as part of this PR, but I imagine that you would prefer a separate PR. LMK.

Maybe the PR title is needed updated to something like GH-39638: [Docs][R] ...?

⚠️ GitHub issue #39638 has been automatically assigned in GitHub to PR creator.

Good call @eitsupi, I've changed the title and updated the related issue too. These changes now look good to me @grantmcdermott, thanks. And to your point about potential further improvements, those are probably best as separate issue.

@thisisnic do you want one more look at this PR?

After merging your PR, Conbench analyzed the 4 benchmarking runs that have been run so far on merge-commit 3eb4135.

There were no benchmark performance regressions. 🎉

The full Conbench report has more details. It also includes information about 21 possible false positives for unstable benchmarks that are known to sometimes produce them.