[duskwuff]

> Yeah, it'd be nice if IM just settled on some interoperable basis.

And critically: that standard needs to not be XMPP. It's a mess of a protocol that's difficult to implement, and which fails to adequately support many common use cases (e.g, mobile clients, file transfer, and group chats).

[jalami]

I think we should coalesce around a clean modern modular standard that's not XMPP for the perfect world scenario. It's not perfect, but I think you overstate how bad it is. Conversations (a mobile client) allows pictures and (file transfers) and with multi-user-chats MUC (group chats). It takes implementing some XEP's on the client and the server in order for everything to play nice. Implementation is a bear and some XEPs are simply too complicated for their footprint. However, if no one is interested in the open federated standard communication protocol we have, even less are interested in adopting another one we haven't written a line of spec for yet.

[Arathorn]

As mentioned elsewhere in the comments here, Matrix is seeing quite a lot of adoption as an attempt at a clean modular federated standard (albeit very different in philosophy to XMPP) - there are around 30 clients, lots of bridges, services and bots, and multiple servers in development. There is very much an appetite and interest in open comms still, as the popularity of articles like these attests!

[jalami]

Matrix is interesting. I like where they're going and once their encryption-layer gets out of beta, I'll take a harder look.

Right now there seems to be a significant Axolotl/PEP/ratchet/OMEMO/olm schism[0] that I hope doesn't fracture decentralized IM any more than it is. App-store DRM v. Copy-left purism (not being derogatory) seem to be at the root. At least multiple parties are trying to work together here instead of forging ahead alone. I'm hopeful.

[0] https://github.com/anurodhp/Monal/issues/9#issuecomment-2080...

[Arathorn]

The point of Olm as an independent Apache-licensed double ratchet is absolutely not to fragment E2E, but just provide an alternative clean-room implementation for those who want it (and to be one that we can easily audit and build on for Matrix). We will do everything we can to ensure that Olm directly interops with whatever the OMEMO team ends up doing. Given Olm is written to be compatible with TextSecure this shouldn't be an issue.

Just to be clear: mismatched E2E is obviously the mortal enemy of interoperability. Which is why it's really exciting that so many systems are using Axolotl derivatives, and it's absolutely critical to ensure they can interwork.