social.coop is one of the many independent Mastodon servers you can use to participate in the fediverse.
A Fediverse instance for people interested in cooperative and collective projects. If you are interested in joining our community, please apply at https://join.social.coop/registration-form.html.

Administered by:

Server stats:

502
active users

I took the following AI in the OC meeting: "start a thread or book a meeting to discuss next steps in the messaging space per-WG and cross-WG".

@edumerco @edsu @sam @Matt_Noyes I think you are interested in this topic :) would you like to start by discussing here? I can also start a time slot picking process to meet and discuss synchronously.

Also, anyone else that you think would be interested?

Eduardo Mercovich (él)

@flancian @edsu @sam @Matt_Noyes @mako @ntnsndr @dazinism
On another different line of thought/software, I had really great experiences with Zulip (zulip.com/).

What makes it great from my POV is the serious thread model. It helps to know what is happening, keep context, focus on what matters for you (including notificacions), mark as done, etc.

It is not federated, but it is Free Software and we can self-host it.
Anyone else tried it? Do we want to make a test?

ZulipZulip — organized team chatZulip is an organized team chat app for distributed teams of all sizes.

@edumerco @flancian @edsu @Matt_Noyes @mako @ntnsndr @dazinism I was thinking before that May First has chat and we have May First access but that it probably wouldn't be ideal for us because it is *just* an account and they don't provide a web client or anything (as far as I know?), but it occurred to me later that we could also just use the May First infrastructure and host our own web client (or just a page recommending actual apps for different operating systems).

@sam
@flancian @edsu @Matt_Noyes @mako @ntnsndr @dazinism

There is (or was, at least) a simple but effective web client really, I've used it a long time ago.
I will search for it. :)

@edumerco @sam @edsu @Matt_Noyes @mako @ntnsndr @dazinism thank you all for your interest and participation!

Ideally I think we could pursue the following using the fact that one of our VPSes is under-utilized currently:

- Install some or all of: a lightweight Matrix server (e.g. conduit), an XMPP server (something prosody adjacent?), a Zulip instance.
- Share test credentials with people in the working groups and gather feedback/see what catches on and productionize that?

@edumerco @sam @edsu @Matt_Noyes @mako @ntnsndr @dazinism on the topic of what the goal would be, I think the ideal outcome is we offer an extra service to the community *with the initial warning that it might have a worse SLO than the mastodon instance*, meaning they shouldn't rely on it for anything critical yet. Then we see how expensive it is to run and we vote as a community on running it or not?

@edumerco @sam @edsu @Matt_Noyes @mako @ntnsndr @dazinism of course it is the working groups that need this the most I believe, but I think if we run an instance at all we should try to make it useful to all the community, because why not? :)

@edumerco @sam @edsu @Matt_Noyes @mako @ntnsndr @dazinism from all the options, I still believe Matrix is the most future looking and the platform where most of our members *already have accounts*; but I'm OK being swayed by data and other proposals, and don't get me wrong: I liked Zulip a lot when I used it, and I used to help run XMPP for Google and loved it back then.

@flancian @edumerco @sam @edsu @Matt_Noyes @mako @dazinism I think it would be great to expand our offerings this way and would love to see us host a Matrix server, esp. since we use Matrix anyway, and to support an important project.

Cf:
loomio.com/d/5yW7iCsJ/new-soci
loomio.com/d/tOBDu5ap/
loomio.com/d/gP67te9Z/

LoomioNew social.coop services?Hey all, as it's come up several times before, and was again discussed in the last call, why not add services other than Mastodon on social.coop? <br><br>We could occasionally install an interesting FLOSS service, and all have the chance to try it out, to see how usable / user-friendly / useful it is. If the software is mature / useful enough for us, we can keep it running and try to provide the developers with contributions (code + financial), otherwise we can just give them some feedback, scrap it, and try another option. <br>Criteria for social.coop member services<br>IMO our criteria would be (in order of importance):<br><br><br>FLOSS<br>User friendly<br>Based on open standards / protocols<br>Federated / Decentralised<br>Social / Collaborative / Co-operative<br><br>Proposed next steps<br><br>Comment with your opinion on this whole initiative. If there doesn't appear to be consensus on moving forward with this, we can first discuss the premise further. <br>Comment with your thoughts / ideas about the criteria. Once there appears to be consensus, I can post a proposal to formalise this.<br>Discuss the financial aspects: <br><br><br> some more lightweight services could run on our existing infrastructure (esp. with the upcoming upgrade), but others would require more resources (notably storage costs would go up with things like file hosting). <br> how to compensate the members who would install and maintain the services, and contribute to developing the software<br><br>Comment with any services you need (or software that may be worth trying) and I'll update the list below. Once we have a semi-comprehensive list of options, we can discuss them more in depth, and then decide where to start!<br><br>Possible servicesIdentity and single sign-on options<br><br>Custom social.coop member database & web app, using OpenID/OAuth, and maybe integrate IndieAuth<br>Look into projects like Portier, Auth0, and OSIAM<br>Look into building on top of Hubzilla, which has a federated system featuring nomadic identity (a user can migrate from one instance to another, unlike with Mastodon). It also has social networking functionality that can federate with Mastodon (though doesn't seem to support ActivityPub). May be be a good option if we want all our services to be tightly integrated, with federation at their core, but would require more hacking on any software we want to add (which could arguably benefit the whole FLOSS community and push it from the current app/platform silos towards more federation & integration).<br><br>FLOSS, federated services<br><br>Matrix / Riot - federated communications platform, incl. group chat and end-to-end encryption, integrates with Jitsi for audio/video<br>Federated Wiki<br>Collaborative workspaces: Apache Wave, Kune <br>Socialhome (also federates with Mastodon)<br>Movim (XMPP based)<br>Blogging (Ghost / WordPress)<br>Cool things like Known, Quill & Bridgy<br>Email<br>Calendar<br><br>FLOSS, not-yet-federated services<br><br>Decision making with our own instance of Loomio and explore others like VoteIt, Orgbook, ukuvota, etc<br>GitLab to help developers avoid centralising everything on Github. Git is already distributed in nature, but certain functionality could still do with more federation (being discussed among GitLab devs).<br>Map creation (umap based on OpenStreetMap)<br>Collaborative document editing (Etherpad / ethercalc / Jetpad / HackMD / onlyoffice)<br>Reddit-like site for sharing & discussing links (Raddit for eg)<br>Bookmarks / links sharing (examples)<br>Feed reader (examples)<br>Scheduling (eg Rallly, others)<br>Video chat (eg Jitsi or Hublin)<br>Secure password & notes saving/sharing (eg turtl)<br>Kanban/Scrum project management (eg Taiga)<br>LaTeX document editing (eg ShareLaTeX)<br><br>Meta-platforms<br><br>Sandstorm (incl. some federation)<br>Nextcloud (incl. some federation)<br>Cloudron<br><br>FLOSS utilities<br>These would be handy for members to have on a platform we control, but don't necessarily meet the Federated / Decentralised nor the Social / Collaborative / Co-operative criteria.<br><br><br>URL shortener (examples)<br>Image hosting (eg mediagoblin or lutim)<br>Photo sharing (eg Lychee)<br>Audio/Podcast hosting (eg mediagoblin)<br>Video hosting (eg mediagoblin, plumi, mediadrop, or clipbucket, and restreamer for live broadcast)<br>Save stuff for later (eg wallabag, tiddlywiki, keeper or turtl)<br>Static website hosting (eg using jekyll, grav or hugo)<br>Pastebin (eg)<br><br>

@ntnsndr @flancian @edumerco @sam @edsu @Matt_Noyes @dazinism I would also be very excited about this and happy to help support the effort on whatever easy of helpful!

@mako @ntnsndr @edumerco @sam @edsu @Matt_Noyes @dazinism glad to read that Benjamin! This is still brewing but mostly unblocked now, next steps are making some time in the tech working group to deploy test instances of ideally several of the options (I'd personally start with a Matrix homeserver though).

If you're already on Matrix and Interested in discussing further feel free to join our room, it's documented in wiki.social.coop!

@edsu @flancian @mako @ntnsndr @edumerco @Matt_Noyes @dazinism meaning no disrespect, but mind dropping me from this thread? As you all know I am strongly opposed to us running a Matrix homeserver, but we don't seem to be honestly considering other alternatives and I really just don't want to have to keep being sad every time this pops up in my timeline.

@sam @edsu @mako @ntnsndr @edumerco @Matt_Noyes @dazinism none taken! I disagree with the assessment of "not seriously considering other alternatives" as we have been exploring Matrix, Zulip and XMPP (this can be seen in e.g. chat room logs) but I respect your choice to sit this one out of course.

@flancian @edsu @mako @ntnsndr @edumerco @Matt_Noyes @dazinism my mistake then; everything I've seen has been about Matrix specifically.

@edsu @mako @ntnsndr @edumerco @Matt_Noyes @dazinism dropping Sam as per request above, if we want to keep discussing Matrix over here let's use this branch of the thread?

Dear @sam
I didn't knew about your distaste with Matrix and I'm very much with you, that's why proposed zulip (or XMPP). But of course, the last thing we want to do is to bother.

Do you want to be included when/where we discuss other possibilities?

Sorry for the sadness... (≧≦)

@edsu @flancian @mako @ntnsndr @Matt_Noyes @dazinism

@edumerco @edsu @flancian @mako @ntnsndr @Matt_Noyes @dazinism oh yah, that's fine, I just get sick of it always coming back to matrix; apparently it's being discussed *on* Matrix (where I can't see it) though, so my bad for suggesting that alternatives weren't being considered.

@sam
No harm done, I open a matrix client to speak with one person really and didn't saw there much conversation about it neither. :)

So, coming back to this talk, which criteria do we use to make a choice?
I'll start with a few to ignite the talk:
. Inclusivity and freedom (amount of possibilities about how to use/access the service)
. Standards (tested, advancements, documentation)
1/2
@edsu @flancian @mako @ntnsndr @Matt_Noyes @dazinism

2/2
. Thread management (what is being talked here now? What is already done and what is alive?
. Federation?
. Security (both data and metadata)
. Low on resources (both for installation and maintenance)
. Possibility to integrate with our -desired- SSO
. What others? :)

@edsu @flancian @mako @ntnsndr @Matt_Noyes @dazinism @sam

@edumerco @flancian @mako @ntnsndr @Matt_Noyes @dazinism it seems like there is some subset of social.coop members who already use Matrix and would like a co-op supported server.

Hi @edsu
I get it, and it is true. And there is other subset that would prefer other solution/s.

Or we could support many services, but now we are limited on people with knowledge so that is for some future. :)

So, in orden not to fall on preferences, let's try to agree on criteria to make a selection for a specific use.

Today matrix is already used by some WGs. Do we want to provide a service to all the coop members or just some WGs?

@flancian @mako @ntnsndr @Matt_Noyes @dazinism

@Eduardo Mercovich (él) @In #Flancia we'll meet @Benjamin Mako Hill @Nathan Schneider @dazinism

The main reasons that I've been interested in social.coop hosting Matrix are 1) it irritates me that a platform is used for official social.coop communications for which we don't provide accounts to our members, and 2) I selfishly would like an account on a Matrix server run by an org I affiliate with.  

I don't specifically think it's a good tool for internal communications, and am pretty sure it's a bad tool for decisionmaking (tools with infinite scroll seem poorly suited for deliberation).

Hi @dynamic
I agreed with almost everything except the selfishly part, which made laugh a lot, so thank you very much... :DDD

Back on topic, I *specially* agree about the infinite scroll. It is positively difficult to talk and focus in such a varying context. That is why I proposed zulip, that manages those things in a spectacular way. :)

@flancian @dazinism @ntnsndr @mako

@flancian @edumerco @edsu @Matt_Noyes @mako @ntnsndr @dazinism I'd like to push back on "why not" as a reason to do anything. More specifically, I think we'll have different requirements depending on if this is for all users or just for a few working groups to use (I suspect if it's for all social.coop members as a service those requirements will be a superset of the requirements for if it's just us). We should decide what we want to do here and have a good reason for it up front IMO.

@Sam Whited  @In #Flancia we'll meet  @Eduardo Mercovich (él) @Ed Summers @Matt Noyes  @Benjamin Mako Hill  @Nathan Schneider  @dazinism

Personally, I would love to have a general purpose Matrix account hosted by social.coop, but also understand if that turns out not to be workable.

@dynamic @sam @dazinism @ntnsndr @mako thanks for the feedback both of you!

- I think it's totally workable if it can run comfortably within the VPS that currently only hosts our wiki and our (under-utilized) single-sign-on service, called Hypha. A test should sort that out.
- I don't think the requirements are going to be too different in the end (continued...)

@dynamic @sam @dazinism @ntnsndr @mako we should set up the services in question using [[coop cloud]], which makes it cheap/quick if the recipes for the services exist. I will check which ones exist and report back.

In any case we would want to set up SSO so people can use the password they use e.g. with the wiki to test.

Which service(s) we try first and which ones we support depends on difficulty and time/interest available within the TWG, assuming no other additional costs for the coop.

@dynamic @sam @dazinism @ntnsndr @mako @edsu @dphiffer @3wordchant

I checked git.coopcloud.tech/coop-cloud for recipes for: zulip, an xmpp server and a matrix server. I found so far:

- the mainline main matrix implementation, presumably slow but well maintained
- snikket (xmpp)

Calix, would recipes for conduit and zulip make sense as contributions?

Co-op Cloud CodeCo-op Cloud RecipesThe Co-op Cloud configuration commons. All [recipes](https://docs.coopcloud.tech/glossary/#recipe) in one place.

@ntnsndr @edumerco @sam @dazinism @mako @edsu @dphiffer @3wordchant this is a bad move by Matrix honestly. Reliability and scalability should not be locked behind a paywall. It seems unethical to leave the community with a substandard implementation, systems that don't scale well tend to be inefficient in many ways.

@ntnsndr @edumerco @sam @dazinism @mako @edsu @dphiffer @3wordchant it's a corpo like move IMHO, would the community that helped you build Matrix so far vote yes on this decision?

@matrix please answer ^ thank you for building Matrix in any case!

@ntnsndr @edumerco @sam @dazinism @mako @edsu @dphiffer @3wordchant above the best client for xmpp: what's the best web interface? web first is a must for accessibility IMHO

@flancian been a while since I researched / ran one but I think the winning option at the time was conversejs.org/

conversejs.orgConverseConverse XMPP/Jabber Chat

@3wordchant @flancian I'm definitely open to trying XMPP (or using it again after a long hiatus). I honestly like the fact that it is lighter, older, and has survived. As an experiment I tried conversejs.org with my @mayfirst Jabber account and got a connection error, but the same settings work fine in Adium. I'm not sure if this is a conversejs issue or a Mayfirst issue...

@edsu I'm not sure why I can't log in to converse with my may first account. @mayfirst about to upgrade our xmpp server and launch our own web client using movim. Also, we will be explicitly adding xmpp to our control panel. Stay tuned! I'm a big fan of xmpp because, unlike matrix, we don't have a complete store of all sent messages on our server.

@jamiem that is an excellent feature, especially in these times. thanks for the details about the plans for xmpp. And also, thanks for the reminder that I need to be tuning into what is going on in the larger community.

@jamiem @edsu @mayfirst fellow MayFirst member and former XSF council/editor here, give me a hollar if you need a hand with any of this! I've been wanting to be more involved for a long time, but never could figure out how, but I'm always happy to answer XMPP related questions (or hook you up with the people you need if they're super implementation specific or something).

@sam Thanks Sam. Once we complete this change we will be on prosody .12 and I was planning to reach out to you to ask you to put us through all your tests again and see if we can ensure we are up to speed. I'd also love to learn why we can't login to conversejs. Not sure if that's worth debugging now or whether we should just wait for the move.

@jamiem RE converse.js I'm not seeing the same error (mine tries to log in and then spins forever), but it doesn't look like mayfirst.org/.well-known/host- is present. Since the address is "mayfirst.org" but the actual connection spot is "im.mayfirst.org" and web clients can't lookup SRV records, I think you'd need a host-meta with the CORS header set to allow the converse.js domain to access it.

@jamiem @edsu @mayfirst interesting! Would you be up for giving a tour of movim when you're done? xmpp chat hosted on MayFirst could be a good solution for SC.
@sam @hollie @flancian

@edumerco just saw this earlier, thanks for raising it here! I've hosted XMPP before, happy to try to get running a server onto the Tech Working Group 2025 roadmap if folks are interested.

CC @flancian @dynamic @sam @dazinism @ntnsndr @mako @edsu @dphiffer

In other group we are also checking XMPP and it's working super fine.
Some details (image download, reactions, etc) are dependent on the server software/version, but to do the usual chat is perfectly fine.
It integrates calls/vc; clients are abundant; tt's federated, based on a standard, time tested. If we have our own server, seems a great choice.
How much effort and computing resources we need to try?
@3wordchant @flancian @dynamic @sam @dazinism @ntnsndr @mako @edsu @dphiffer

@sam @flancian @edsu @Matt_Noyes @mako @ntnsndr @dazinism

Today I checked with @mayfirst and I was wrong, there is no XMPP web client there.
However, any XMPP program can work with their service. :)
Should we explore with them if is possible to have some private rooms?

@edumerco @sam @flancian @Matt_Noyes @mako @ntnsndr @dazinism @mayfirst I'm not sure they need to be private? I didn't know that Mayfirst provided an XMPP server, thanks for sharing that!

@edsu @edumerco @sam @flancian @Matt_Noyes @mako @ntnsndr @dazinism

yes! workers use it every day so we know it's properly working but not so many members do it so welcome more people there!

Welcome to this talk @mayfirst ! Glad to have you here! :)
AFAIK today social coop has 1 MF account only that it is used to authenticate and log in.
Can we use the XMPP service with many people? If yes, how?

@edsu @sam @flancian @Matt_Noyes @mako @ntnsndr @dazinism

@edumerco @mayfirst @edsu @flancian @Matt_Noyes @mako @ntnsndr @dazinism I do think exploring a partnership here would be beneficial (if it's something other mayfirst folks would be interested in, I can bring it up on the internal mayfirst forums if the team needs more details), but I think for our needs it would require them to run a custom configuration just for social.coop, so it would be more complicated than just using mayfirsts existing XMPP server.

@sam yes! starting a new thread in TIAS or Engangement discourse would be the best so more people from those groups can join in the conversation

@edumerco @edsu @flancian @Matt_Noyes @mako @ntnsndr @dazinism

@edumerco @flancian @edsu @Matt_Noyes @mako @ntnsndr @dazinism if we did end up in a partnership with them we could always run our own web frontend and let them handle the actual chat server too.