Hacker News Reader

Matrix will hit exactly the same problem in a few years. The solution is certainly not to restart with a clean slate every few years and fucks to you if you wanted compatibility. That's Google's approach, where every few years they trash an old service and - maybe - replace it with a new one that works in a different way.

There's no simple answers to this. I can tell you it's absolutely not a protocol issue - the protocol issues are ensuring graceful degradation remains possible during advances, which XMPP does well - but a political one. Profile specifications, which indicate groups of XEPs which the community expects to be supported, are part of this. Certification might also need to happen. Maybe monetary awards, even.

One of the biggest things currently driving server implementors to get the XEP support, and server maintainers to deploy it, though, is folk like inputmice driving the market with Conversations.

Sorry, you haven't thought this through. As long as XMPP exists in a state where there are public servers with varying XEP support then adoption rates will always languish. A user would actually have to do considerable research to understand why features they've come to expect from centralized IM aren't working should they happen to choose the wrong server or client. Without a large degree XEP parity, XMPP has ZERO chance to ever compete with centralized IM. The average user simply won't put up with it.

Matrix has a shot because it's consistent right now. If they stay consistent, they will have an even larger shot.

> There is no easy way to tell if your server supports them, either.

False. XMPP has excellent feature negotiation capabilities.

> everyone using Conversations essentially has to use a server that is maintained by the developers

No, you just need to use a server maintained by someone who actually maintains it.

> Otherwise, most of the XMPP featureset they offer simply doesn't work.

Feature negotiation and graceful fallback, is it really that hard?

> I think there needs to be incentive built into the protocol for all servers to maintain a baseline of XEP parity

Incentives, yes. In the (core) protocol? No. I think publishing a new "XMPP Compliance Suite $year" XEP is better than throwing the entire thing out, in a couple of years when everyone suddenly wants a different feature set.

> At least Matrix can provide a clean slate.

A clean slate which will be obsolete in just a few years, and will have to be invented from scratch all over again.

I'd be surprised if Matrix is obsolete in a few years. The difference between the Matrix and XMPP spec structure is simply that Matrix is a single curated monolithic spec rather than a set of XEPs. So there's only one true standard at any given point, at a given version (unless someone forks it). We've been evolving it quite rapidly so far, adding modules for the various different use cases out there (http://matrix.org/docs/spec/client_server/r0.1.0.html#featur...), and I don't see that rate of evolution slowing down any time soon.

And even if we did go and entirely rearchitect Matrix (e.g. making it entirely P2P, which is certainly a thought experiment we keep in mind), one would just go bridge it into the existing Matrix ecosystem and migrate as desired.

> We've been evolving it quite rapidly so far, adding modules for the various different use cases out there [...], and I don't see that rate of evolution slowing down any time soon.

> And even if we did go and entirely rearchitect Matrix (e.g. making it entirely P2P, which is certainly a thought experiment we keep in mind), one would just go bridge it into the existing Matrix ecosystem and migrate as desired.

Are you saying you might end up in a situation where "fragmentation of features between clients and servers is common"?

I think you missed the thrust of my argument which is that XMPP sucks from a _user_ perspective. A non-technical user perspective which is 90% of the users in the real world.

> False. XMPP has excellent feature negotiation capabilities.

User don't care what the protocol can do. "How come persistent group chat isn't working? Ah, I'll just use WhatsApp"

> Feature negotiation and graceful fallback, is it really that hard?

This is dependent on the client. Does your client support that? Who knows? Users don't care about any of this.

> Incentives, yes. In the (core) protocol? No.

There's no greater incentive for public server providers than the danger of having their user base get isolated from the rest of the network. That can be done through the protocol and is the only incentive that will actually _accomplish_ anything.

> There's a reason that everyone using Conversations essentially has to use a server that is maintained by the developers.

That is completely untrue. First of all there are public servers that have exactly the same feature set like jabber.at second of all we are running ejabberd on conversations.im. ejabberd is open source can be run by anyone. You just have to do it. Granted the list of public servers with a full feature set is not as long as we'd like. But the best way of changing this is to run one yourself.

It's impossible to get reliably statistics in a federated system. (Which of course is the entire point.) but looking at the little data I have from people sending me bug reports over XMPP at seems that a lot of Conversations users are running their own servers. The problem is not that the extensions are not implemented (Both prosody and ejabberd in fact support all the extensions Conversations does.) but that a lot of public servers out there are second class services for their operators and are not properly maintained. In that case you can either contact them and ask them to upgrade or ask yourself if you want to have an XMPP account on a provider who doesn't maintain their servers.

You need to think about this from a non-technical user perspective. They'll be using the Conversations server, guaranteed. Because

1) They're not going to run their own server.

2) They're not going to bother researching which servers support Conversations entire featureset.

That's why XMPP is garbage for the majority users and why XMPP weenies will stay baffled as to why there aren't a flood of users using it.