On Thursday, 6 July 2023 at 23:25:12 UTC+2 Nils Bruin wrote: It looks to me proposal 1 still doesn't make it *mandatory* to drop support outside of the NEP29-defined version window. It seems to me that it just defines when a support-drop PR becomes acceptable to merge, and that such a PR would only be issued once someone feels there's a benefit to its merge.
If that interpretation is right, it seems to me that the difference between proposal 1 and proposal 2 will only be relevant if there is a PR to drop a python version V (outside of the NEP29 window) and there is someone else who want to keep support for that version V. With proposal 2 we'd get a discussion of the relevant pros and cons. With proposal 1, there'd normally not be a discussion at this point: version V is fair game to be dropped. There could still be a discussion if someone claims there's an exceptional circumstance that warrants deviating from official policy, but that would have a much higher bar to clear. Yes, this interpretation is what I had in mind. >From the perspective above, Proposal 1 seems more attractive to me than Proposal 2, since it avoids extra discussion with what seems a very reasonable default policy (the NEP 29 window seems wide enough to me). For me, this is the main advantage of having a policy about the drop schedule. We can now have a dedicated conversation to thoroughly explore the pros and cons, to consider various perspectives, gather feedback, and make an informed decision. Once a policy has been established and agreed upon, it provides a clear framework that eliminates the need for repetitive discussions and debates about dropping specific policies in every PR again. -- You received this message because you are subscribed to the Google Groups "sage-devel" group. To unsubscribe from this group and stop receiving emails from it, send an email to sage-devel+unsubscr...@googlegroups.com. To view this discussion on the web visit https://groups.google.com/d/msgid/sage-devel/56b2148c-7b22-4631-ba7c-2e931ae03bc3n%40googlegroups.com.