Recently someone mentioned the difficulty of submitting to CRAN #rstats. In close to 4 years of data, I found that 55% of new package submissions fail on the first try (7202 new packages in the period analyzed) .
Overall, ~70% of submissions fail to pass CRAN checks (15926 different versions of packages submitted).
I am checking the data, but I'll post a blog soon. More or less in line with the latest update from CRAN: https://journal.r-project.org/news/RJ-2023-1-cran/
@Lluis_Revilla
Interesting post. Thanks for sharing. Can't wait for your blog post too.
Is the comments publicly available? I wonder how many of them were about titles, function names, documentation and wording. Basically how many are technical ans how many comments were about wording and naming.
@Mehrad No, exchanges between CRAN volunteers and maintainers are not public: they can contain sensitive data and are private (GDPR, ...).
If they are about wording and naming they are manual actions, as per the link, but those could also be, fix links, change of maintainers or similar exchanges. I could try to retrieve problems in checks and count that, but the historic archive doesn't have it. I tried to document them via self-reporting some years ago in https://github.com/cranchange/cran_histories/discussions ...
Add comment