Because of the "stupid X", modern XMPP supports far more than it used to, and isn't stuck in the past like email and SMS.
There is a mismatch between iMessage (Apple) and RCS (Google's flavour of the month). To the point where there is almost no sensible interoperability between the two.
All XMPP does is provide answers to "If I want to implement feature X, how should that look on the wire".
Just as the XMPP Standards Foundation annually publishes the recommended baseline feature sets for XMPP clients, it wouldn't be hard for Apple and Google to follow that or (more likely) agree on their own baseline for interoperability between the two ecosystems.
As I always say when this comes up: the wire protocol is of least concern - it's not the reason these businesses don't prioritize interoperability. No protocol engineering can magically fix that.
> Because of the "stupid X", modern XMPP supports far more than it used to, and isn't stuck in the past like email and SMS.
Yeah, it’s dead. Maybe XMPP supports shiny stuff. But no client or server support them, and if they do it’s like they don’t understand the spec the same way.
A protocol should not be extensible, it should be full featured and regularly updated to include new needs. It should also propose a reference implementation and an official client so that there’s a clear baseline.
I use XMPP because unlike Matrix, XMPP is properly standardized and not a product by a single VC funded startup.
If you ever have multiple independent implementations a single monolithic spec will always only be partially supported (or there is only a single useable implementation like Synapse), so no different than having extensions.
Have you ever considered that your constant FUD against Matrix just ends up hurting open communications in general? How about putting your energy into improving XMPP rather than constantly whining about the existence of Matrix.
Matrix is properly standardised at https://spec.matrix.org by a non-profit foundation: https://matrix.org/foundation. Just because the core team created a startup to fund our work (3 years into the project) doesn’t make it “a product by a single VC funded startup” - especially when there are hundreds of independent companies building on the standard. Meanwhile empirically synapse, dendrite and conduit can all talk fine to each other and we haven’t seen any fragmentation yet.
If you want to complain about something, go attack the closed cabal of RCS or the hypercapitalism silos of FAANG - and leave us the hell alone.
I don't think that promoting existing internet standards over Matrix will hurt open communications.
Instead I think constantly reinventing the same thing with slightly different, incompatible primitives is hurting the ecosystem a lot more.
Bridges also won't solve this. The matrix.org XMPP implementation (besides not being mentioned anywhere, not even in the Matrix FAQ about XMPP) is almost unusable and I suspect it will never allow for encrypted communications or A/V calls.
I don't see any "whining" or "attack" before your comment. Sure the comment mentioning Matrix is a little confused about Matrix the foundation vs the startup, but did it really warrant what you just wrote?
Putting a notification on "matrix" on HN so you can jump on every critic and yell "leave us the hell alone" is not kind.
Matrix ended up hurting open communications in general. Simply because it is another open, federated standard. It splits usage. Creating another standard is a time honoured way to degrade and destroy an existing standard. I am not claiming that Matrix is some sort of conspiracy to destroy XMPP but it serves that purpose anyway. Every time you make a new incompatible thing you take away from the old thing.
There is a mismatch between iMessage (Apple) and RCS (Google's flavour of the month). To the point where there is almost no sensible interoperability between the two.
All XMPP does is provide answers to "If I want to implement feature X, how should that look on the wire".
Just as the XMPP Standards Foundation annually publishes the recommended baseline feature sets for XMPP clients, it wouldn't be hard for Apple and Google to follow that or (more likely) agree on their own baseline for interoperability between the two ecosystems.
As I always say when this comes up: the wire protocol is of least concern - it's not the reason these businesses don't prioritize interoperability. No protocol engineering can magically fix that.