Hacker News Reader

Comment by zaik | Hacker News Reader

You might like XMPP then - IETF Internet standard, always had decentralization and e2e encryption. I still don't understand why so much energy was invested in building a seperate ecosystem. Matrix claims to be fundamentally different, but all their major selling points (dencentralization, e2ee, open protocol) were already there. And since the Matrix encryption protocol is incompatible with all the other ones there's currently no way to securely communicate with someone outside their network.

View on HN

It hasn't really been there:

- encryption wasn't really there (and still is experimental)

- decentralization does not exist to the degree that Matrix provides it

And Matrix is fundamentally different. It works somewhat similar to git compared to XMPP being more like IRC, which results in a completely different form of decentralization. E2EE is different as well - even if both are based on the Signal Protocol: Matrix encrypts for groups of people instead of single devices

I doubt Matrix would be where it is if they tried to bend XMPP

Decentralization in XMPP exists to a much greater degree that in is in Matrix.

Practical decentralization, where you can talk to your friend that uses a different provider?

Not just using the same client for Facebook and your corporate chat instance?

Practical decentralisation, as in thousands of servers connected to each other in one federation, working for 20 years.

> Practical decentralization, where you can talk to your friend that uses a different provider?

Yes. You can find some lists of servers participating in the federation here: https://xmpp.org/getting-started/

It's just that XMPP providers are not required to participate in the federation. Sadly, large instances like WhatsApp choose not to.

> encryption wasn't really there (and still is experimental)

Clients could encrypt messages using OpenPGP and the more recent OMEMO seems pretty reliable to me (Conversations, Gajim).

But XMPP failed (as a decentralized open protocol in the wider picture).

Maybe it failed in the consumer app space because it couldn’t keep up with user experience expectations. But if that is the metric, then Matrix already failed too, right?

One new server + consumer app combo could revive XMPP for consumers quite easily.

So not keeping up to users expectations - no matter how much - makes a failure?

I think your argumentation is flawed. If your server+consumer app combo can do that, do that

I was giving the person above me the benefit of the doubt. I assume that was their argument, not mine. But i entertained the idea regardless.

Reply on news.ycombinator.com