Follow

I've said it before and will say it again: email is one of the best things technology has ever accomplished. Such a pity that IM over Email hasn't caught on yet, and the world feels chaffing fatigue like this
youtube.com/shorts/8FrRhXKCA4U

@AmpBenzScientist Yep. Delta Chat does it very literally. But really all IM is very close to email; it's just something about its appearance that leads us to treat it formally.

@worldsendless @AmpBenzScientist The expected low-latency of instant messaging is what drives the behavior and assumptions around it, I think.

@lispi314 @AmpBenzScientist funny thing is, in practice, the latency is pretty much the same for email (hence so much 2FA)

@worldsendless @AmpBenzScientist Yeah, the design of the SMTP protocol is *very* lenient by today's network reliability and uptime standards.

@worldsendless @AmpBenzScientist Personally I'd like to see more delay-tolerant message-oriented protocols. They're much easier to port onto meshnets & darknets.

@AmpBenzScientist @worldsendless It's kind of designed to build its own over a number of different transports isn'it?

Regarding XMPP, is it not subject to the usual low-latency expectations common to most instant messaging protocols?

@lispi314 @worldsendless Yes to the first question. XMPP was used to control industrial equipment. It is rather robust but delays could cause problems.
At a certain point it is helpful to realize that the delays are the absence of a connection that is suitable for communication. Matrix had lowered bandwidth required for messaging to be well within dialup speeds.

@AmpBenzScientist @worldsendless The thing is that you can't really implement a mixnet resistant to observers without delays, which is why I care so much about the message-part as a self-contained unit and the delay-tolerance as an assurance that the protocol won't just keel over when I try to use it in such a way.

Low-latency networking inherently leaks a large amount of timing-related metadata in its operation that greatly facilitates deanonymization of users.

@lispi314 @worldsendless I guess there's always sending and receiving simple alive messages to the connections. It maintains that a link is established and also the delay time to expect. A simple protocol could be expanded upon by simple changes to work in these ways.

@AmpBenzScientist @worldsendless I suppose it could, though it's simpler and more convenient for it to have no real expectations related to timing. Much like email, for the most part.

Sign in to participate in the conversation
Qoto Mastodon

QOTO: Question Others to Teach Ourselves
An inclusive, Academic Freedom, instance
All cultures welcome.
Hate speech and harassment strictly forbidden.