https://drewdevault.com/2021/04/07/The-next-chat-app.html
Drew DeVault has some correct criticisms of Matrix. Although I haven't run a server myself directly, I've been in close contact with people who have, and I know what a task it is.
However, DD seems to think this disqualifies Matrix all together. I'd like to think that it can overcome these difficulties, maybe with upcoming implementations like Dendrite.
@makeworld The biggest problem with Matrix is the huge and rapidly changing API, as well as a lot of bad design decisions (e.g. JSON/HTTP polling) that would take creating a different protocol to solve.
Dendrite does not really solve any problem, it rather makes the existing ones less noticeable. Being written in Go means a lot more performance, efficiency and reliability (to the point where it can run in your phone or web browser); but complexity and bad protocol design decisions are still there.
@avalos @makeworld Matrix *already is* the different protocol. And totally needlessly aswell. There is no technical reason why they couldn't simply have gone with XMPP
@waweic @makeworld True, XMPP is probably a better choice. The XMPP ecosystem needs to grow and evolve. OMEMO needs to be fully standardized.
@avalos @makeworld It basically just needs some funding. That's the single reason why Matrix is proliferating. The protocol isn't better, it's arguably worse in many aspects. But Matrix.org got the money
@waweic @makeworld Actually, XMPP is still one of the most used chat protocols, but users are not aware because Big Tech doesn't bother talking about it.
@avalos @makeworld True, that's usually my talking point though :D