[danaris]

Hangouts is the one I'm most personally frustrated by, too, but what I'd really love to see rise up to replace it is an open protocol with a reference implementation client that's similar to Discord (broadly, anyone can create a "server" that's independent of actual physical servers, each server can have more-or-less arbitrary channels, most importantly both text and voice/video, supporting inlining of images and some other nice modern QoL features that, say, IRC lacks). It frustrates me no end that all our common "modern" communications platforms belong entirely to one company or another, and will only be allowed to exist as long as those companies can continue leveraging them to maintain an income (one way or another).

I've heard some good things about Matrix[0], but I'm even a bit wary of any "open" protocol that's backed by a single for-profit company. However, I don't know of anything out there that's better right now.

The trouble, as always, is getting the people you want to communicate with to agree to use it, too...

[0] https://matrix.org

[willtemperley]

Matrix is interesting, I'll have a look, thanks. Signal is the only video call capable chat app I know of I would genuinely trust.

Unfortunately the killer P2P protocol, BitTorrent, has been a victim of its own success, I imagine it'd ge tricky to get a BitTorrent protocol based video app into the Apple App Store, for instance.

I suppose it's the App Store in question that defines the limitations of a P2P video/chat app, I realise I don't have enough knowledge of those limitations to comment further.

[danaris]

I'm not sure if BitTorrent is a reasonable solution to the problem of P2P video chatting, since the power of BitTorrent is in having many nodes all sharing the same content and able to send parts of it to a new node seeking that content.

Realtime chat doesn't seem amenable to that type of optimization.

In any case, I don't think the problem here is primarily in the technologies being used; I wouldn't exactly call real-time video chat a solved problem (I'm sure there's room for plenty more optimizations in it over the next few decades), but at least as I understand it, it's definitely one for which there are plenty of available open-source stacks available.

The problem is in a) designing a robust protocol/API that can be implemented by many clients and servers (that fulfills my basic requirements above of "basically Discord, but not controlled by one company"), b) writing a solid reference implementation of both the client and the server, and c) (perhaps most importantly) getting other people to adopt it so it's more than just Failed Competing Standard #27.

[willtemperley]

Looks like BitTorrent tried it, with Bleep [0], but it disappeared in 2015 leaving barely a trace. I agree with all your points though.

[0] https://www.latimes.com/business/technology/la-fi-tn-bittorr...