Sure. Experiments are cool. On Thu, Jul 28, 2022 at 7:12 PM Ross Turk <ross.t...@astronomer.io.invalid> wrote:
> > Perhaps it doesn’t have to be something that only committers are > responsible for. If we use comments instead of tags, anyone can do it. > > We could a) add a bit of text underneath the PRotM section in the > newsletter that says something like “if you want to nominate a PR, flag it > by adding #protm to a comment”; b) update the script so that it heavily > weighs PRs with this comment tag; c) add a comment containing that tag to > PRs that have been selected, so we start to build a dataset. > > In general I think it’s good to provide mechanisms for people to > participate, and this is one potential mechanism. If nobody uses it, then - > by definition - it has had no ill effect on anyone’s time. Well, anyone > except me and Michael. And if someone nominates a PR just one month out of > twelve, it’s worth it. > > We’ll give it a try this next month. Why not? :) > > Ross > > > On Jul 26, 2022, at 5:36 PM, Jarek Potiuk <ja...@potiuk.com> wrote: > > Might be difficult. Adding new behaviours (especially for people who have > plenty to do) is difficult. I think it's better to make smart use of > existing behaviours that are already happening regardless. > > On Tue, Jul 26, 2022 at 11:28 PM Ross Turk <ross.t...@astronomer.io.invalid> > wrote: > >> >> > On Jul 22, 2022, at 3:08 PM, Jarek Potiuk <ja...@potiuk.com> wrote: >> > >> > Side comment: I wonder when we will start employing AI to select "best >> pr”… >> >> For that we would need a training dataset of some sort. A big one. >> >> Perhaps we can start small by tagging PRs that are potentially >> noteworthy? That would be useful in the short term as well - the script >> could look for them and weigh appropriately. >> >> Cheers, >> Ross >> >> >
