Where should the compiler team (and perhaps working groups) chat?

On the rust-lang.zulipchat.com server, you can look for the various streams (e.g., t-compiler, wg-nll, wg-unsafe-code-guidelines)

FWIW over the last couple of months I feel that more and more conversations are getting lost, because people will be logged in everywhere, but will only check one of these places. So for example, I would ping X over a couple of days on IRC, without answer, and then decide to ping X on discord, and realize X hasn’t checked IRC in a while even though it appears as logged in.

I’m basically back at just only using IRC, and github issues or internals for the rest. Tracking discord, gitter, and all other systems simultaneously is a huge pain.

3 Likes

I think most discussion is happening over at Discord nowadays; at least that's where the lang team hangs out. :slight_smile:

I’ve been thinking about this.

Maybe we could take a first step, before settling on Discord vs Zulip, and delete the #rustc IRC channel?

It’d be one less plan to monitor.

3 Likes

:+1:

While we're at it, can we also delete #rust-lang ?

2 Likes

I’m finally getting around to making a comment. Here are my feelings:

  • Zulip and Discord are both significantly superior to IRC (though it would be nice to be able to view a conversation without signing up, like on IRLO or GitHub).
  • Zulip is the better platform: you can’t beat threads and it’s open source.
  • Despite this, I still prefer using Discord. Zulip has several interface quirks (*bugs) that are annoying, it doesn’t feel as intuitive to use casually and it doesn’t look good.

(I haven’t yet got around to using Zulip’s mobile app, but Discord’s is good.)

I think we definitely need to promote a single platform. My tentative feeling is that we should be promoting Zulip (with all the associated corollaries, such as closing the channels elsewhere), but not yet. I think we should focus (in whatever form that takes — probably fixing them ourselves) on addressing the main problems with Zulip before making the move.

I think that the problems with Zulip are composed of lots of small irks, rather than any fundamental problems and that tackling them (ourselves) would not be a mammoth effort. I’m going to split off a new topic to discuss this and see how much work it would take to convert those of us for whom the usability and interface of Zulip is the main sticking point.

2 Likes

This topic was automatically closed 90 days after the last reply. New replies are no longer allowed.