[Standards] Update to MIX 0.9.7

Evgeny Khramtsov xramtsov at gmail.com
Fri May 11 14:22:36 UTC 2018

Thu, 10 May 2018 12:39:54 +0100
Dave Cridland <dave at cridland.net> wrote:

> XEP-0060 takes a message protocol and builds a pubsub service on top.

I don't want to run into terminology debates, but XMPP strictly
speaking is not a "message protocol", but a protocol for XML routing.

> You're suggesting that MIX should build a message service on top of
> this?

Yes. MIX should build a message service on top of XML packets.

> The problem, in my eyes, is that we have existing, well-defined,
> handling for the concepts of "messages" and "presence", and I would
> hate to have to deal with these entirely differently for group chat
> compared to for 1:1 chat.

There is probably no need to say for the 1000 time that the conception
of "presence" is outdated. I also doubt messages by itself is a good
concept. Yes, they are easy to understand for sure, but seems like in
modern IM we are talking about replicas exchange between endpoints.
The simple concept of a message carrying human readable text 1:1 is
somewhat outdated too. Sorry for the "moot statements", but you started
this discussion in the first place. Frankly, I would rather discuss
technical aspects instead of this philosophic bullshit.

> There are always going to be differences, of course (short of forcing
> every 1:1 conversation into a  2-person groupchat, which has its own
> unique problems), but it makes sense to me to minimize these, and
> reuse existing handling wherever possible.

Well, I actually suggest to reuse existing specification and, what's
more important, existing code. Pubsub is something we have more or less
implemented, and can reuse it, despite it doesn't fit ideology of some
people. Your vision of Pubsub as a "building block" is wrong in my
opinion, because this doesn't help me at all, I need to reimplement
*all* pubsub code to handle MIX, because now Pubsub is not a protocol
between a client and a disc storage, but rather some abstract stuff
used by ad-hoc services which translate Pubsub requests in their
internal state.

More information about the Standards mailing list