Re: [Tools-discuss] Trial chat services: matrix and zulip

[Dave Cridland <[e-mail]>]

On Mon, 5 Oct 2020 at 21:23, Tom Pusateri <pusateri=
[e-mail]> wrote:

> In my opinion, we should not be switching from XMPP to a new chat service
> without first opening a working group to look into what failed and how to
> move back to an IETF standards protocol for chat in the future.
>
> According to Robert above, it’s a client and server availability problem
> (not necessarily a protocol problem). Why do vendors not want to support
> XMPP? Likely because a walled garden is more profitable for them. Is that
> what the IETF should be encouraging?
>
>
XMPP is very widely used - almost every military uses it, and if an online
game uses any kind of open standard for chat it is XMPP - including
Fortnite, for example. These uses are not generally very visible, and often
not on the internet itself at all - but some are federated cases (such as
the military's use), and others are very large. Some are smaller, like that
of my own employer.

Servers are readily available, the majority of XMPP services use open
source servers, and there at least 6 actively maintained and well used
interopable implementations (for those that want to count them: ejabberd,
MongooseIM, Openfire, Tigase, Prosody, and Isode's M-Link - the latter
being the only one not to be open source). It wouldn't surprise me if there
are more; there certainly have been others.

Clients are harder, because "consumer" and "enterprise" markets are the
weakest the weakest area of use, and these tend to drive client
development - but really this is no different to email, where email
clients, and indeed the protocols they use to connect to services, are not
interoperable standards. The XMPP world normally uses a bespoke client, but
it still speaks XMPP's C2S protocol. This means that productized general
client experiences are niche, and generally driven purely by the open
source community (which does a pretty good job given the circumstances) -
again, this is essentially the same as the email world.

But Fortnite, for example, *is* an XMPP client - albeit with some kind of
game attached. If having 350 million users counts as failure, it's the kind
of failure that I'm happy to aspire to.

If XMPP can be considered to have failed, then it's that it has indeed not
persuaded the likes of Google, Facebook, and so on into open federation and
interoperability. Google did it several IM systems back, but abandoned it.
No matter what protocol the IETF chose to replace it with, it seems
unlikely that this direction, at least, will reverse - but that is not a
technical problem but a business one. Sadly, I believe the email world is
headed the same way - witness the mangling of your email address.


> If the XMPP protocol is not sufficient (I understand Matrix is quite well
> designed), then the IETF should be standardizing something better (even if
> that means standardizing Matrix).
>
> This should be the first step.
>
> Tom
>
> ___________________________________________________________
> Tools-discuss mailing list
> [e-mail]
> https://www.ietf.org/mailman/listinfo/tools-discuss
>
> Please report datatracker.ietf.org and mailarchive.ietf.org
> bugs at http://tools.ietf.org/tools/ietfdb
> or send email to [e-mail]
>
> Please report tools.ietf.org bugs at
> http://tools.ietf.org/tools/issues
> or send email to [e-mail]
>