Le mer. 2 mars 2022 à 18:40, Jarek Potiuk <ja...@potiuk.com> a écrit : > > > > > > > > > 1. make sure foundation itself provides an MVP for at least two > > styles of communication channels ("topic grouped channels" > > and "instant messaging channels"). We've got email and perhaps > > Matrix could be a good enough answer to the 2nd requirement. > > > > That would be great. My slack relationship is a bit of love-hate and > especially when recently they enabled it for a month in "full" version > and then started to annoy us with "do you really want to lose all that > - do pay" was a bit crossing the line (and we can expect more of it).
What did you expect, indeed? How long until GH will go that path? > Having a good "modern" replacement "blessed" by the ASF would > be fantastic. Matrix looks cool and ASF promoting free and distributed > and modern communication tool for that seems like a great idea. > > 2. The best we can do on everything else is simply collect a sort > > of FAQ advising our communities on places they should consider > > monitoring/engaging so that they "go where [users] are" to your > > point. > > Anything else I am missing? > > > > > I thought a bit about this and I think I realized something. > > I think what I miss is something in-between those two emails/async- > at least for some of the communities and for some parts (see below) > > I think email is great to do really "official" communication and > particularly > information that relates to the community. Everything that is related to > the community - > is all fine to be discussed on the mailing list (establishing processes and > communication rules, deciding on project policies etc). > And if we treat it this way, this is good and I never saw any problem > with it. Great. :-) > But I think other media (particularly GitHub Discussions and GitHub > Issues) **might** be better for some communities to make even > important decisions about the *code* not about the community. :-( See below. > > I think we are really at the place where many of the projects require > anywayw GitHub account, otherwise you can neither discuss nor change > anything related to the code. And I think we should embrace the fact > that GitHub tools are simply better suited to discuss any code related > stuff and even make decisions there. This already happens for small > things (basically in every PR) and this is extremely blurry when the > decision is big-enough to bring it to the devlist. The motto "if it did > not happen on the devlist - did not happen" is not even mentioned > anywhere in the official Airflow docs (or I could not find it) but we > keep on hearing (and I was myself saying that multiple times). > If somebody asks me now about a new feature or "Concept" change > in the code - I have no way currently to explain when the feature > is "big enough" to be discussed on the devlist or whether it is > enough to get it approved in PR or issue. > > "Community over code" is actually there in Airflow official docs and main > motto. Although the motto is meant (I assume) to highlight the importance of having a community (of individuals) for the long-term survival of a programming project, use of "over" is misleading as it can actually be used for dividing a community (rather than try and reach consensus). [A better (IMHO) one would be "Community around code".] > > Maybe we should simply (as the official approach of the ASF > make it a totally viable possibility for the projects to use those: > > * when you discuss about community and the way it works "if it did not > happen > on the devlist - it did not happen" > > * but when you discuss code - "if it did not happen on the <choose your > medium here that fits our criteria> it did not happen". That should just be the reverse: * community -> whatever tool is in fashion (for some part of the community) * code -> ML Indeed, this is what happens already for a long time; taking an ASF conference as an example. Some people can attend, and some cannot; yet the former group is not allowed to come back to the official forum (ML) claiming that they decided <something> during their meeting at the conference. > > I think I would be perfectly fine with that. Thus establishing GH preeminence, forcing every ASF contributor to be a GH client and being entirely dependent on it. Not a thing I look forward to. Regards, Gilles > That is maybe not perfectly > black- > whilte criteria (code and community discussions have some overlap) but it > is much "clearer" to me than any other criteria I could apply even now > to what we do. > > WDYT? > > J. --------------------------------------------------------------------- To unsubscribe, e-mail: dev-unsubscr...@community.apache.org For additional commands, e-mail: dev-h...@community.apache.org