[Standards] {Core|Advanced} {Client|Server} 2015

Dave Cridland dave at cridland.net
Thu Jun 18 18:39:53 UTC 2015


On 18 Jun 2015 19:21, "Curtis King" <cking at mumbo.ca> wrote:
>
>
>> On Jun 18, 2015, at 8:45 AM, Dave Cridland <dave at cridland.net> wrote:
>>
>>
>> On 18 Jun 2015 15:40, "Curtis King" <cking at mumbo.ca> wrote:
>> >
>> >
>> >> On Jun 18, 2015, at 7:25 AM, Dave Cridland <dave at cridland.net> wrote:
>> >>
>> >> There's consensus, I would argue, given that it's extremely well
supported in servers, desktop and mobile clients. In fact, finding servers
that didn't support it a year ago is hard.
>> >
>> >
>> > Two servers and maybe 5 clients does not make for well supported in my
book.
>>
>> Which two were you thinking of?
>>
>> Ejabberd has supported it for two years, Openfire for 18 months or so on
trunk, prosody for ages. I don't know about the others, but those three
account for a very high percentage of deployed domains, once one excludes
Google, who don't implement anything anyway.
>>
>>
>
> Ok three servers if count a 3rd party plugin for Prosody.
>

Yeah, I'd misunderstood prosody.

>> > But, how well an extension is supported doesn’t give it special rights
to skip the standardization process.
>> >
>>
>> Supported isn't the same as deployed, and I'm not arguing high
deployment skips the standards process. However, it does offer evidence
that it works, and is desirable, and high deployment does suggest high
consensus.
>>
>>
>
> You keep making general statements with out any supporting facts. From
where I sit, I see zero deployment. I can’t find a single iOS XMPP client
which supports 280, Trillian has rolled their own solution, all I can find
is a few Android clients, some web frameworks, and after 30 minutes of
looking zero desktop clients. Far from an exhaustive survey but from your
statements I figured it would be easy to find clients which support 280.
>

Gajim, for one. And as for iOS XMPP clients, those are thin on the ground
anyway.

>> Arguing against it on the basis that an unwritten perfect protocol would
be better is a much weaker argument.
>>
>>
>
> I did no such thing. I will rephrase.
>
> Is XEP-0280 a complete solution to it’s stated problem? I say no because
it does not support the offline case.

I don't think it's aiming to support offline use, so I'd disagree with this
assertion.

> Should a XEP with an incomplete solution be included in the Compliance
Suites? If yes, at what level?

An incomplete solution to what? 198 is an incomplete solution to the two
generals problem, but I wouldn't say it's a waste of time as a result.

Extensions solve specific use cases, and "mobile" is really broad,
consisting of many different aspects.

> Should a XEP be able to be added to the Core level by-passing the
Advanced level?
>

If we'd had a compliance suite since 2012, and mind you, that one was
deferred, then I'd say no.

I suppose it really depends on what the suits are for. I see them as a way
of saying, if you want a basic server, a sensible minimum is this stuff in
Core. Those are your cut-off features.

If you want a really good all round experience, you want Advanced.

But "you" here is typical internet usage, we're not talking tactical, or
IoT, or any of the other use cases.

I imagine reaching consensus on what the compliance suites mean would be a
useful first step.

> ck
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.jabber.org/pipermail/standards/attachments/20150618/b9071b3d/attachment.html>


More information about the Standards mailing list