On Fri, 12 Feb 2021 01:37:35 +1100 Steven D'Aprano email@example.com wrote:
I think it is probably pointless now, but you might get more response if you post actual examples of the syntax and semantics, explaining how your proposal differs from the (now accepted) alternative, and why you made the choices you did.
You should also consider that pattern matching is built on software patterns that are proven to actually be used. People *actually do* write code based on isinstance matches, tuple unpacking, etc., so it seems strange to want to avoid those patterns.
But honestly, the window of opportunity has probably closed.
I wouldn't say it's closed. There was a bunch of pattern matching solutions for Python before PEP634, and there will be a bunch of alternatives after.
But we now have a baseline reference to compare against. And I guess everyone's expectation now when hearing about alternatives will be:
a) Show me differences with the standard pattern matching, including actual side-by-side code. b) (Assuming differences look ok) explain them.
Also, at this time, we all probably need to make homework on adopting the baseline implementation, so I'm not sure how many will be interested to discuss alternatives right away (beyond the expected elaborations which were brought up already while discussing PEP634).