Please only vote if you have contributed to PieFed in some capacity (e.g. coding, translation, reporting bugs, discussing bugs) or might in the near future!

Currently, we have two different developer chats on Zulip and Matrix. This means, that new devs will be confused about which channel to join and experienced devs have to actively look around in both channels for questions to answer. This seems like a really inefficient system, so we should at least decide on one official communication channel.

More discussion at https://piefed.social/c/piefed_meta/p/1964543/we-should-rethink-how-the-communitcation-channels-for-piefed-devs-work

Personally I find Matrix to be a pain in the ass and prefer Zulip. But it hasn't managed to attract as many users as Matrix has despite pushing it quite hard for months. Also PieFed development has started to bleed over into PieFed naturally, such as !piefed_css@piefed.social and !piefed_plugins@piefed.social. So we're in an awkward place at the moment.

31% A PieFed community, only
3% chat.piefed.social - Zulip
10% Matrix channel
18% PieFed + Zulip
21% PieFed + Matrix
13% PieFed + Zulip + Matrix
3% Zulip + Matrix (what we have now)
3% Zulip
39 votes - ended 7 months ago
you are viewing a single comment's thread
view the rest of the comments
[–] 5 points 7 months ago (6 children)

Alright, here are my thoughts. I voted PieFed + Zulip + Matrix, but really I think my answer can be a bit broader than that. Here are the three things I think we need and I don't think PieFed covers them fully yet:

1. Somewhat persistent store of knowledge (can be PieFed)

This would be a place where asynchronous discussions can be had or questions can be asked and answered about how to do things regarding development. See something like this exchange I had with DeckPacker as an example. It illustrates everything that this mode of communication needs and PieFed can provide:

  • It was a very technical question, so it would be relevant in a focused dev community
  • It wasn't time-sensitive, so the asynchronous nature of posts/comments works fine
  • Future devs could have a similar question, so having it be somewhat accessible in the future works well

This currently kind of exists within codeberg issues/PRs as well as within our documentation, but someplace with a lower barrier of entry for the dev-curious to ask technical questions might be nice.

2. Realtime support and troubleshooting (mostly Matrix currently)

The needs that this channel of communication needs to meet are realtime troubleshooting and support of other piefed admins that are experimenting or having issues. I personally don't think there is going to be any real way to not have this be matrix. That is where the vast majority of the threadiverse admins and app developers already are. Also, lots of this stuff doesn't need to persist long-term, nor should it.

3. Internal, non-public discussions (currently mainly Zulip)

This channel is one where the access control is much stricter and only intended individuals can view. Currently, most of the piefed.social admins/staff and the core maintainers are active on Zulip. With Zulip making it relatively easy to private message and create private channels/topics, it has worked well for this purpose. I guess that private communities would work for this in PieFed, but it would mean that everybody that is invited needs to make a piefed.social account since private communities don't federate. In practice, that's not that different than people needing to make a zulip account I guess.

  • source
  • hideshow 6 child comments
  • [–] 7 points 7 months ago (2 children)

    @rimu@piefed.social - I am kind of answering your MVP chat question from the other thread here:

    If we wanted to dogfood as much of this as possible, then here are the features I think we would need to add into PieFed:

    • Private group chats : This would be a way for admins/staff/mods to have a truly private chat internally. This would essentially enable something like a mod chat, admin chat, or whatever-else-you-want chat. I think an MVP implementation of this doesn't need to federate and only allow local users and it could still be useful. Federation would be a lot trickier due to the much less secure privacy and software incompatibilities with other Apub software.
    • Public group chats : This would be a group chat that is open to any user. Think of something like sticking a chatroom link in the community sidebar that users can use to join. I honestly don't know how to make federation work here since the link would presumably be for the chatroom's home instance. Maybe something like the url-rewriting that we currently do for post links.
    • Wiki Improvements: This would be a bit more of a nice to have than something for an MVP, but some love for the wiki system could help with knowledge management. I think some of the main things here are actually permissions issues. If we could do something like empower users to have wiki privileges just within a community (without making them an outright mod), that could help ease the burden on moderators and distribute wiki maintenance load.
  • source
  • parent
  • hideshow 2 child comments
  • [–] [S] 6 points 7 months ago

    I reckon I could make substantial progress on those in a weekend. Something basically usable.

    The scary thing about contemplating going all-in on PieFed is that we'll be missing some functionality that we need - but that is also kinda the whole point of dogfooding - to highlight what's missing and drive improvements. That discomfort is to be sought and embraced.

  • source
  • parent
  • [–] 1 point 7 months ago

    software incompatibilities with other Apub software.

    You can just make it federate encrypted data. At least for first only PieFed would be able decrypt the private messages, but eventually at least mBin would join for sure. Safe private messaging would be a very useful feature in any case!

    Having to create an account in a specific instance is too much asked, IMO. But having to create an account on some instance supporting a specific feature does feel reasonable, at least to me.

  • source
  • parent
  • [–] [S] 5 points 7 months ago (2 children)

    I guess it wouldn't need to be a sudden change - whatever the result of the poll is could be used as an eventual goal to aim for rather than something we immediately implement today.

  • source
  • parent
  • hideshow 2 child comments
  • [–] 7 points 7 months ago (1 child)

    Yeah, I think we would need to do things gradually. I would be happy to mod a piefed community for development if we want to go ahead and make it, and then go from there. We can transition stuff around when we think the features are there. Open source like this is a race against burnout more than it is against time, so we don't need to rush.

  • source
  • parent
  • hideshow 1 child comment
  • [–] [S] 7 points 7 months ago

    Race against burnout, heh. For sure.

    I think part of the reason we've been on 1.6.x for so long so I've chilled out a bit, trying to find a more sustainable pace rather than banging out new features constantly (and then associated fighting fires).

  • source
  • parent