Mattermost is a pretty niche chat client and I’m also unable to login to seL4’s Mattermost instance on my Pixel phone (it just hangs). Per Metcalf’s law, I believe that the choice of Mattermost is going to limit community engagement.
I can understand if that is an intentional choice, as it seems like this is primarily an internal communication tool. Having barriers to engagement is one way to keep a good signal-to-noise ratio However, a migration to Matrix or similar would probably help grow the community.
What was the rationale for Mattermost and was Matrix an option when the decision was made?
We’ve been using Mattermost for a very long time, I don’t think Matrix was a viable option back then. To migrate now would mean to lose history and search ability, but if MM is really that bad to use maybe we should consider options again.
@wombat what do you think of Matrix?
Based on a quick googling, it appears that you can import existing conversations to some extent. The pattern seems to be to bridge for a while, then archive only Mattermost, then drop it.
Note that Discord has a larger social network and more support focused features. Zulip also has a better UX for organizations and strong uptake in Rust and high-assurance circles.
But given our audience (OS devs, security nerds, and FOSS nuts) Matrix seems like the natural choice.
Zulip is a chat that is trying to be structured like a forum and that does not work at all for me – at least I find it completely unusable as a chat and the UX is terrible. I have to use it for some other projects and I positively hate every minute of it. It sort of works as a help forum that is synchronous, but we already have a forum and we do not have the bandwidth for a high amount of synchronous messaging.
People chatting informally like on Slack/Mattermost/etc is about as much as we can support at the moment.
1 Like