Suggestion for growing the community around Crystal

Community
21

There’s a Docker image for Zulip. I want to think managing it would be quite straight-forward on an AWS/DO/Vultr instance.

1 Like
22

I think is also about money.

Matrix has also to be considered, with the https://modular.im/ service.
I never talk about adopting one specific solution. This can be voted too.

I talk about visibility and Crystal adoption

23

The first thing to decide is whether a common platform for topics and projects related to crystal can be useful.

After we can think about the plattform:

After someone else must think if want to officially endors the project and based on that we can decide (and vote?) a plattform.
We can also decide that this plattform will be this forum for example!

1 Like
24

Sure, those things have to be solved.
But I wish people would reflexively reach for the Open solution first, instead of reaching for the big free corporate solution. Those Services don’t cost money, but that does not make them free.

1 Like
25

No, really. It is a job.

  • the platform will have to scale
  • solve possible incidents
  • handle storage with backups - docker does not solve this issue

It is just not a random service for a set of guys, but a service with lots of people relying on it.
If down, the core members and the community won’t be able to talk to each others.

I don’t think core members want to spend time on system administration instead of Crystal code.

26

What is specifically wrong with Slack other than “it’s a big free corporate solution”?

And how is it not free? What’s the cost?

27

Freedom

28

It’s not a job. It’s a chat server. I’ve managed XMPP and IRC servers in the past for small communities - they don’t usually have to scale.

Doesn’t the core team already spend time on the website, hosting the docs, etc.? How is this different? Especially with people volunteering to help? There is more to a language with this kind of ambition than writing code, IMHO.

2 Likes
29

I mean administrating a server is a job, with knowledge and time to spend.
They are not hosting the website and docs, this are just static sites.

30

A single-use instance is pretty simple to manage, with a minimal attack surface. It’s not everyone’s cup of tea, to be sure.

31

Mattermost and Zulip are simply better (IMO, at least equal), more flexible, cheaper, plus FOSS.
Even on the proprietary side, Discord has more features. Slack has improved and has some nice features that Discord doesn’t have.

32

So far I see these alternatives:

  • Do nothing. Maybe Gitter is good enough? (I personally don’t think so…)
  • Use Slack, which is basically “do nothing” because it’s already up and running. We would just need to decide which channels to have, and these would grow to add channels for Lucky, Amber, Ameba, etc.
  • Use a hosted solution. It implies a lot of work: hosting a server, paying it, keeping the server up to date, keeping the app up to date.

So I see what “Freedom” means if we don’t use Slack, like, in the pure sense of “we have control over everything”, but on the other hand paying and maintaining an open source platform doesn’t sound like freedom to me… (it sounds more like work).

3 Likes
33

Me neither

I think we only need to create a subscribe page like elixir and others have done, and accept the history limitation

It’s a work like many others in a community…for the price I see we can start with $10/20 on DO and then scale…but if we scale we have had success and probably someone want pay. I think there are already a couple of volounteer for the work…

1 Like
34

Slack is free, and yes we don’t have freedom, but honestly chat is not a good way to search for or find answers. Forums are much better at that. So let’s say Slack goes away tomorrow and we move platforms, do we lose anything that valuable? I personally don’t think so.

So Slack would:

  • Be free
  • Zero maintainence. Even if others are little maintainence, zero is a nice round number :)
  • Allow channels for anything (like Lucky, Amber, etc.). Meaning easier to collab/discover as @Acciaiodigitale suggested
  • Much better UI :tada:
  • People are familiar with it
  • Lots of big communities use it, we aren’t reinventing the wheel
  • They privacy concerns don’t make sense to me. If we have gitter or zulip or any of the others anyone could use that data. Maybe not legally, but most of those companies don’t care. And this is all in public so I don’t think this is a concern.

I’d personally be up for Slack. It is stable, not going anywhere, free, easy to use and familiar for people (wil live right next to their work Slack most likely!). Hell, it’s already up thanks to Ary.

So IMO this is a great direction to go

EDIT: Zulip seems ok too. I think as long as we have one “blessed” chat that is a good option. This is the main reason we use Gitter for Lucky chat

5 Likes
35

Happy to hear we are moving away from Gitter. The Amber team will move to Slack or whatever platform the Crystal Team decides to go with. I have requested a slack channel #amber from @asterite

2 Likes
36

Some Open Source, et al, Slack alternatives.

1 Like
37

Zulip looks like to be a good compromise:

It just feels right that FOSS projects use and support each others.

1 Like
38

Hi Community, I already joined slack and I’ll be happy to contribute to whatever makes crystal-lang better for developers :smile:

I hope you all are doing well working from home :+1:

2 Likes
39

We’ve already ended up talking really fast about very specific details. There are way more options to consider than what has been brought up here.

So let’s please take one step after the other. Thanks @Acciaiodigitale for bring up the topic and trying to keep the discussion focued =)

So far, the responses here have been in broad support of that. That’s a good thing. But it still doesn’t mean we’ll inevitably ending up with that. It’s not only about being useful, but being more useful than the alternative (i.e. what we currently have), considering all positive and negative effects it would have on long term use and short term (switching process, parallel operation etc.)

Exactly. And before that, we need to think about what exactly we need/want from a chat platform.

I have collected few aspects that have so far been mentioned in this thread or I deem noteworthy to add. They’re just a collection of what has been mentioned and don’t represent my or a common opinion:

  • common meeting ground for the Crystal community
  • also integrate communities of Crystal projects/shards => synergy for the ecosystem
  • organize discussions into separate channels but still integrated on one platform (important for growing community)
    • channels can be created for interest groups, projects, local groups, etc.
    • delegate admin power to owners of channels
    • private groups/channels
  • private messaging between members
  • good UI (also on mobile devices)
  • easy onboarding for new/potential community members (user story: “I have a quick question about Crystal to see whether it’s worth checking out.”)
    • join community via registration link
    • federated authentication via Github
  • integration with external services
    • tools for source code management (linking github issues/PRs, maybe notifications from github/CI)
    • custom integration for Crystal resources (API docs, blog)
    • bridge to other/legacy chat platforms (Gitter, IRC)
  • service features
    • no-cost SaaS solution would be ideal
    • open source software is preferred
    • open data access
    • self-hosted option with more (data) autonomy
4 Likes
40

Another one: threaded conversations. Being able to ask a question in a channel and get responses just under that question is very valuable in my opinion (more organized than intermixing questions and answers).

4 Likes