←back to thread

“I'm basically giving myself a permanent vacation from being BDFL”

(mail.python.org)
2024 points randlet | 1 comments | 12 Jul 18 15:22 UTC | HN request time: 0.206s | source
Show context
rurban ◴[12 Jul 18 16:57 UTC] No.17516478[source]▶
Now this is a very dangerous situation, looking how perl5 was handled after Larry wall left them and stepped over to perl6.

Constant infighting of the core devs, useless communication on the mailing list, not a single worthwhile feature being implemented in 15 years, even if perl6 and Larry designed tons of them suitable for perl5, total destruction of the syntax and the core. The better counterexample was perl6 with stricter processes and some capable devs remaining, but still it's another minefield.

Good luck, tons of work to do. python 3 could get 10x faster and catch up to php7 or javascript, types could get properly integrated but I don't see it coming. Usually you have to expect the worst.

replies(3): >>17516719 #>>17516736 #>>17517977 #
1. zbyte64 ◴[12 Jul 18 17:23 UTC] No.17516736[source]▶
Can tell you performance has never been an issue for me and python. Usually when I see the claim come up it is usually a poor test (ie "let's see how fast we can read from a database") or not knowing the performance options available to python. You can use pypy for JIT compilation performance. You can use eventlet to add co-operative threading / evented processing. Use multiprocessing if you want to use more than one core. The point of python is to optimize for human understanding 1st, and if your abstractions are good then you can enable extra performance with other libraries.