conduit VS Synapse

I'd like to suggest Conduit. I found it very easy to install and maintain. https://gitlab.com/famedly/conduit

> "At least as standard" how?

There are 8 people who vote on changes to the Matrix spec (the Spec Core Team), 7 of which are Element employees (including Matthew, Element's CEO). Element also controls the development of clients and servers used by the large majority of users in the public federation.

> A substantial portion of the IRC comunity is actively hostile to the IRCv3 extensions, and in some cases prefer incompatible implementations of the same functionality; Matrix has nothing like that going on.

But any IRC client will work fine on any IRC server, and they can connect to various servers with different implementations.

On Matrix, clients (generally) can only connect to one homeserver at a time; which forces them to converge on following exactly the same spec. And if your server differs ever so slightly from the other ones in how it implements some parts of the spec (room consensus), then it can be split-brained from the rest of the federation. Instead, changes to the room consensus are done by pushing new room versions, and each server implementation needs to explicitly support it or they can't join it. This means Synapse devs (which are a majority of Element employees) get to decide what room versions can get traction.

It is not uncommon for people in the Matrix community to complain about this and Element keeping specs in limbo, and PRs to the flagship clients being stuck in "design review tar".

> And there seem to be more visibly independent implementations of Matrix than IRCv3.

Clients, maybe, at least in the number of implementation. It's hard to find stats of this, but I feel that >95% of people in the public federation use Element even in tech-y rooms; IRC has a healthier mix of major clients (weechat, irssi, IRCCloud, Hexchat, KiwiIRC, The Lounge each have >5% of desktop/web users). But I admit that's just my very subjective point of view.

In terms of servers, Matrix has three open source ones as far as I know: Synapse (controlled by Element), Dendrite (controlled by Element, and almost on par with Synapse according to https://arewep2pyet.com/ ), and Conduit. Based on https://gitlab.com/famedly/conduit/-/milestones/3 , Conduit seems to be far from implementing the spec yet (eg. it doesn't seem to support leaving rooms or respecting history visibility).

> things like: server-side history extensions tended to mess up my client's history implementation (I'd end up with multiple copies of the same messages in my local logs, often with the wrong timestamps)

You can use https://ircv3.net/specs/extensions/message-ids to deduplicate them.

> And if you're in a conversation where people are using embedded gifs, then fundamentally you'll always be a second-class citizen if you're trying to participate in that with a client that can't display embedded gifs.

A conversation where people where people are using embedded gifs will exclude me regardless of client, because they are too distracting. At least on IRC I can expect people not to do it too much, and use words or emojis instead of reaction gifs.

> SSO access control; you just can't do that in a nice way if the client doesn't support it

That's a fair point; IRC is made by hobbyists more than companies, so that's not surprising. There is some discussion around it though: https://github.com/ircv3/ircv3-ideas/issues/74 and Sourcehut is sponsoring implementation (https://emersion.fr/blog/2022/irc-and-oauth2/).

To start, this will be strictly Non-Federated. Just a few friends will be using this. Here: https://gitlab.com/famedly/conduit/-/blob/next/DEPLOY.md is the documentation I am following. It tells me I must "use my server name", but what is this exactly? What do I put in there? Do I have to go out and buy a domain?

Either Tinode (https://github.com/tinode/chat) or Matrix Protocol (https://gitlab.com/famedly/conduit)

Matrix Protocol: Fractal (Client), Conduit (Server)

If self-hosting is an option then I'd say Matrix, you can try Conduit (server) and Elements(client). To simplify deployment you can refer to this repo.

You can babble on and on about how its not how you do it, no one needs it, etc... But its a demonstrable need in this space and its caused me great pain trying to write applications that would be used by such people. It's even bit Conduit to the point they have 5+ DB backends coded in now that the user can choose between based on their local system setups.

Give Conduit a try. It uses way less memory than Synapse. It is still in early stages but works great. I have been running one on a Pi4 for like a year, going great so far.

There's a Rust implementation of a Matrix server that uses sled: https://gitlab.com/famedly/conduit/

What are you thinking of here? Synapse has supported purging room history since 2016: https://github.com/matrix-org/synapse/pull/911, and configurable data retention since 2019: https://github.com/matrix-org/synapse/pull/5815.

Meanwhile, Matrix has never needed the full room history to be synchronised - when a server joins a room, it typically only grabs the last 20 messages. (It does needs to grab all the key-value state about the room, although these days that happens gradually in the background).

If you're wondering why Matrix implementations are often greedy on disk space, it's because they typically cache the key-value state aggressively (storing a snapshot of it for the room on a regular basis). However, that's just an implementation quirk; folks could absolutely come up with fancier datastructures to store it more efficiently; it's just not got to the top of anyone's todo list yet - things like performance and UX are considered much more important than disk usage right now.

some context re the Matrix isses, long history apparently: https://github.com/matrix-org/synapse/issues/14481#issuecomm...

Why not Matrix? Here's one reason: it has incredibly hard-to-debug edge cases, and plenty of bugs. One of my favourites is the one where people are kicked out of your room at random, which was reported a year ago[0]. It wasn't fixed, however, because the head of the Matrix foundation (Matthew) presumably didn't like the issue being posted on Twitter.

This is honestly really disappointing behaviour from a platform owner.

[0]: https://github.com/matrix-org/synapse/issues/14481

> That doesn't make this situation any less bad to the rest of the community.

How is the community suffering here? Let's say Element adds a bunch of baller stuff to their versions over the next few months and then closes the source. Can't the community just fork the last AGPL version? You might say, "well then no one can take the AGPL fork and make their own closed-source business", but do you want them to? Even if you do, they still can with the existing Apache-licensed version, just like Element is doing right now.

You're arguing that Element will lose a lot of contributions, but TFA points out that despite being super open, the vast majority of contributions are still made by Element employees (which seems to be true [0]). It's not the case that Element is looking to monetize the (small) contributions of others, it is the case that others are looking to monetize the (huge) contributions of Element.

And besides, aren't the MSCs the core of Matrix? It's already super possible to build your own compliant client and server.

The situation is that Element needs money to keep developing the ecosystem. It would be cool if there were a big network of donors and contributions, but there isn't. You're essentially saying, "that's fine, go out of business then, and the community will keep developing the ecosystem", but that's not happening now, and it can still happen anyway with the Apache-licensed versions, which again people can still contribute to.

[0]: https://github.com/matrix-org/synapse/graphs/contributors