Reconsider Slack for HF?

Haskell Foundation
#32

I think focusing on the important features would help the discussion. To my understanding we want the following features (not particularly ordered):

  1. Accessible: it should be dead easy and fast to click a “Discuss Haskell here” button, and get into a chat and start communicating.
  2. Auditable: The chat log is an important piece of low grade documentation and can be a great resource for people solving a Haskell related problem or for people looking to learn more Haskell techniques.
  3. Modern Feature Set: As pointed out by @Nycticorax features such as notifications are important to the quality of the conversation. The chat client should have a reasonable feature set for the modern era (I cringe at these words :stuck_out_tongue: )

Please feel free to add a Feature if you think I’ve missed something important. With this list we can rate our options, for example I rate slack as good in Accessibility and Features but bad in Auditable:

Slack:

  • Accessibility: Very good, its on all platforms including mobile
  • Auditable: Bad, the log is locked in a walled garden hence the move from Open Source Projects from slack instances, and the owners of the channels do not have full control over the channel instance.
  • Features: Very good, there are gifs, we could recreate lambdabot in slack (I think someone has done this before?) and there is a system of notifications through user accounts.

FWIW I like the libera.chat channels and IRC but I understand that is not to everyone’s taste. I’m +1 for Matrix, but agree with @hasufell RE the CEO

1 Like
#33

Thank you, that’s a constructive remark.

I think it would be fair to add “easy to migrate from Slack” as a (at least soft) requirement, since migrations go smoother when they are nice to people who dwell in the previous location.

I also think a (stronger) requirement is: Leans toward a “work groups” workflow, with easy groups, channels and side-threads management, because it’s important to be able to do actual work there (something you can do easily on Slack/Mattermost/Chanty but cannot do as easily on mere IM apps like Matrix and the likes).

#34

Well, one of the things that came up in this thread is that we should wait for our new ED. @david-christiansen is now revealed, and will formally begin in the new 2 weeks IIUC!

What’s the best way way to move this forward? (Presumably just putting David in the hot seat and asking him to make a snap decision in the thread is not what people had in mind!)

#35

What’s the best way way to move this forward?

David can, of course, speak for himself, but my answer is: Wait a few months. There are many great things that we want from the HF, but my hope is that we can move quickly to delivering improvements to the Haskell ecosystem and community. Having a debate – about which people will manifestly not all agree – and then executing a potential change has significant costs in both time and annoyance. This change might be a good idea in the long run. But spending yet more time organizing and reorganizing ourselves is not a good look, in my opinion. I’d rather us focus on steps we can take in the near term to make Haskell a safer choice for your projects and more pleasant to work in than to switch from one imperfect chat platform to another.

6 Likes
#36

Oh, I don’t to rush or force a contentious decision. I had the impression the conversation mainly stopped here because people said “wait”, and so I was checking in now that the “wait” was dying down.

For me, facilitating good communication is not a distraction from good work, even in the short term, but precisely how that good work gets done. But if others don’t see it that way — see instead we are already full steam ahead just getting reoriented with new staff and getting existing projects out the door — then so be it.

My caring about how we work always has still getting out of the way of HF staff doing the day-to-day job as a top priority.

2 Likes
#37

I absolutely understand the concerns about Slack that are expressed earlier in this thread, but my instincts point me in the same direction as @rae. Meta-work spent figuring out how to do our work is important, but there’s definitely diminishing returns as more is done, and I worry that opening this can of worms right now will mean not getting other things done.

Perhaps we can have it in the back of our heads and then revisit the question when I’ve been working at the HF for some months and have more insight into where our efforts are going?

6 Likes
Could GHC remove old extensions?
#38

OK! It’s your call. I can wait until then.

1 Like
#39

This is a good reason for not using Matrix at the moment. Element’s search isn’t great.

(but I still like Matrix)

#40

Well note we are just using Slack free tier, and so it has already pay-walled some past messages. Surely that is worse!

I suppose one thing I missed in that framing of this is emphasizing just how much the HF slack is basically dead, so this is really less “should we switch chat apps?” than “do we want to try to resuscitate having a chat at all?”

If the old thing is not in use, then switching is less disruptive because no one needs to stop doing what they were doing.

3 Likes
#41

All good points. In my experience the most socially enticing platforms are Discord and Telegram. Telegram is a non starter for work and data ownership. Discord is a non starter for the latter of these two reasons. Matrix is not ideal for searchability. So again, I cannot narrow down on anything else on the market than Chattermost. (Sorry if I sound like I am obsessed with it; I just think it tick all boxes.)

1 Like
#42

Looking at Chattermost’s GitHub it says it’s a (-n early) clone of Mattermost. Why not use that? It’s open source, too.

Actually, I think, with a custom Matrix homeserver, you could achieve searchability.
And there’s already an active Haskell community on Matrix at #haskell-space:matrix.org that could integrate with HF rooms.

1 Like
#43

Woupsie I had a brain lapse because I certainly did mean to write “Mattermost” and not “Chattermost”. I would be glad to help migrating to Mattermost.

As for the other point, when I first entered this thread I was strongly against Matrix for the lack of email notifications. Now I’ve found out that you just need to turn them on, but I am still a bit reluctant because it’s quite deeply nested within the settings structure. On Mattermost it’s much easier to find. Also Mattermost is much easier to bend to a collaborative workflow in my experience. Rooms are easy to set up, you have out of the box global search (relative to all your rooms) and you get subthreads for side conversations.

#44

For the record, the HF Slack started life as a tool for HF board members and executives to communicate, it subsequently morphed into a place for folks who wanted to communicate about HF work, and then further morphed in a general Haskell community forum. One of the things we should bear in mind when making a decision is which of these three things we want it to be (or something else).

2 Likes
#45

That’s a very good point!

For me, all are good, but “HF board members and executives to communicate” is the lowest priority.

The strong push-back I got from @rae makes me think I was interpreted as trying to tell the staff & board how to work, but really I am just hoping for an “official community chat” with the HF’s stamp of approval on it. Come to think of it, I was surprised what the reaction was “wait for new ED” in the first place. But if I was interpreted as primarily being interested in how the HF works, then asking me to wait for new staff-people part makes a lot more sense.

I specifically like Matrix because it is good for these “concentric circles of purposes”. If I am involved with project A and project B, and I would like one to learn from the other, if they are are both using the same medium cross-project dialogue is much easier. But by the same reasoning, if I was doing some general purpose Haskell community chat, and then the discussion veers into talking about an HF project in particular, it is nice if that doesn’t mean hopping a new platform.

I am not on the board or an executive, so how those people want to talk amongst themselves does not concern me!

2 Likes
#46

Ah, I think I can see how confusion arose.

For me saying “The HF’s Slack” is a bit like saying “The HF’s Google Workspace” (i.e. it’s where people formally running the HF do their work). For you I think it’s more like saying “a chat server set up by the HF for the communiy”.

As a point of information, the original reason for setting up the HF Slack was the former. Perhaps the scope has crept to the latter.

2 Likes
#47

I think most people don’t interpret it that way because it’s linked on the website and open.

#48

Which link are you seeing. I’m seeing this one

Joining the Slack is a great way to begin volunteering and participating in HF projects. …

https://haskell.foundation/contact/

which does suggest fairly strongly it’s for HF business, not a general Haskell forum. That’s not to say it couldn’t become a general Haskell forum, but that should be a conscious decision.

1 Like
#49

But it is also explicitly open to the community, not closed for board and staff. So, middle “ring” in your 3 from earlier.

1 Like
#50

What is the goal of the haskell foundation having such a platform? Is it solely for seasoned members of the community or the haskell foundation to collaborate on? Or is it for the entire ecosystem to leverage and collaborate on?

I ask these questions, because I believe it’s important to meet people where they are already. And if it’s just a platform to be used by a few people, then it’s easy to convince those people to use anything.

But if the platform is intended for everyone, then it might be tough to convince people outside our ecosystem to install new software just to keep up with the conversations there.
Eg, a developer who heard about haskell and wants to learn more, is more likely to join a slack or discord channel if they already use slack or discord on their computers, than to go over the mental barrier of downloading and using a new platform only for haskell.

Just looking at the programming ecosystems I am active on, I don’t know of any language ecosystem that uses something non-mainstream.

Here are some of the communities i’m part of:

And these are just the communities I know about. There are loads: GitHub - mhxion/awesome-discord-communities: A curated list of awesome Discord communities for programmers, GitHub - thisdot/tech-community-slacks: Here is a list of all the tech community slacks!

My point with listing this out, is that the haskell community is very small, especially compared to other ecosytems, and I don’t think we benefit from using an obscure tool which few others use, since this would be a barrier to entry fir lots of other developers who can’t justify installing a new software just for haskell.

We would however, benefit from just using the tools that most other developers are already using, and leverage that to attract more developers into our ecosystem.

4 Likes
#51

It has never consciously been that. Perhaps it should be! But as yet, it has not.

2 Likes