On Thu, Nov 11, 2021 at 1:38 PM Tom Lane <t...@sss.pgh.pa.us> wrote: > Even if you were excited about it, would maintaining such data be > a good use of project resources? It's not like we lack other things > we ought to be doing. I agree that the lack of reliable versioning > info is a problem, but I can't see that "let's fork ICU and tzdb too" > is a good answer.
You might be right, but I think it's hard to say for certain. I don't think this is one of our top 10 problems, but it's probably one of our top 1000 problems, and it might be one of our top 100 problems. It's entirely subjective, and people are likely to disagree, but based on those numbers I'd say it's not worth 2% of our resources but it might well be worth 0.02% of our resources. Everybody's going to have their own opinion, though. I'm not sure how relevant those opinions are in the end, though. The community has little power to force anybody to work on anything; people work on what they want to work on, or what they get paid to work on, not what somebody else in the community decides is most important. Anyway, from my point of view, if some well-respected community member showed up and wanted to add a new kind of collation that is provided by PostgreSQL itself and had some well-thought-out candidates for initial integration, I don't know that it would be smart to turn that down because solving the whole problem for every case might be more work than anyone's willing to do. The only real issue for the project is if somebody makes a drive-by contribution of something that's going to need continuous updating. That sort of thing would be bad on multiple fronts: not only do we not want to get forced into spending ongoing maintenance effort on something like this, but we want collation definitions that *actually don't change*. -- Robert Haas EDB: http://www.enterprisedb.com