Hacker News new | ask | show | jobs
by _yacj 1731 days ago
That last sentiment would be very nice indeed, but I fear this is what we'll get instead:

https://xkcd.com/927/

2 comments

Closed IM protocols aren’t standards so this doesn’t really apply. There’s only three realistic “standards” in instant messaging: irc, xmpp and matrix. IRC is essentially dead, xmpp is on its last leg and mostly irrelevant, and matrix is very promising just not popular enough yet.
XMPP is widely deployed and it's the backbone of WhatsApp, you don't see a lot of servers that openly federate XMPP because there's little business incentive to do so.
Xmpp can continue being a server side protocol all you want, it doesn’t make it a user facing protocol. So in terms of federated or at least cross compatible IM clients it’s irrelevant.
Doesn't Apple implement XMPP for iMessage as well?
> xmpp is on its last leg

Why? Unlike Matrix it's actually an Internet Standard.

It's had about 20 years to hit the bigtime, I think we can safely say it's not going to happen.
A mere protocol will never "hit the bigtime" without some compelling (non-protocol) reason. It's unfortunate that businesses choose to build silos time and time again, but that's where we are - it's not an issue that can be solved at the protocol level.

Should this be fixed (e.g. through regulation), XMPP is absolutely a sound choice - given, as you say, its 20 years of experience, evolution and deployment track record.

honestly, given the current adoption, growth and state of the ecosystem, I don't think it is.

In theory, you could achieve with it what Matrix achieves, but it would probably be more feasible to start from scratch - like Matrix did

That said, I see how it currently has advantages when it comes to resource consumption as well as internal deployment to centralized services.

So what is the bigtime open messaging standard?
there isn't any.

But if I would have to bet on a standard which could become it right now, then it would be Matrix.

It convinces when it comes to adoption, concept, features and growth.

XKCD should make a comic about pre-emptively 927'ing any suggestion to improve something.
It doesn't mean don't bother improving something, just don't expect everyone to switch to it.
Having hardly anyone switch to a new standard when there are existing ones could actually be considered a good thing. The problem comes when the new standard is promoted, inevitably fails to become dominant and ends up making the problem worse.

Note that sometimes new standards are promoted specifically to destroy an existing open standard. The trick is to push hard for the new standard until it gets sort of popular and then back off. Adjust as required to keep things balanced.