[Council] [Board] Software recommendations on xmpp.org website

Dave Cridland dave at cridland.net
Fri Feb 20 13:50:10 UTC 2015

On 20 February 2015 at 08:25, Kevin Smith <kevin.smith at isode.com> wrote:

> Note: Council can’t respond to the board list (other than me), so this
> isn’t going to work as a way of discussing very well.
> On 20 Feb 2015, at 03:35, adam at andyet.com wrote:
> > xmpp.org having “recommended” options might be controversial and
> undesirable in the “we’re all equals here” sense of being a standards body,
> but  I really believe it would be a Good Thing™ in being responsible
> leaders of a community. Of course, I believe those recommendations should
> be reflected by council, and perhaps even voted on or ratified by members.
Actually I much prefer to build on Ralph's oft-stated suggestion that we
should have magic-switchy-web-stuff for examples in developer
documentation, so that examples can be shown for a particular language and
library based on the user's choice. I think that would showcase the
flexibility available to developers, as well as allow developers to pick a
"nice" library for their particular language and style.

> > Does anyone think this is a *bad* idea?
> I think it’s a very appealing idea, but I don’t see any way of it working:
> a) Technically. For us to evaluate scaling, for example, scalability is
> decidedly non-trivial. Reliability even more so.

Or correctness, or pretty well anything beyond trivially provable things.

Even relatively simple things like "Supports SCRAM" become horrible when
considering what that actually means - usually such things boil down to "in
certain configurations", and "with limitations".

> b) Politically. I remember the last time the JSF tried running objective
> feature lists for servers, and the games the vendors used to play to get
> their scores higher, and just looking at the lies told on the wikipedia
> comparison page shows that this is a bit of a minefield.
> I’m uncomfortable with using the voice of the XSF for this.
Yeah, likewise.

The problem with the XSF recommending a particular server, client, or
library is that we're implicitly recommending against others - whether
that's the intention or not.

If the XSF chooses, say, stanza.io for the "web" XMPP library, what signal
are we sending to those members of the community working on strophe.js, or
XMPP-FTW, or ...?

> As a counter-proposal, if we feel this is really needed, we might consider
> doing this as individuals. i.e. “Here’s each member of Council’s personal
> opinion on clients and servers. These don’t represent the views of the
> XSF”. I think this would still be useful for users and admins, needs a
> lower bar as it’s opinion, rather than trying to be objective, and a
> /little/ less politically challenging (although I’m still concerned about
> this).
I suspect the way to do this is actually to have Council and Board folk
advocate different clients and servers, probably as part of the "bio" we
tend to skip over. Recasting those bios as mock interviews (as Surevine
does) might make this work.

It might even be worthwhile getting such an interview done with each
developer wanting to have a library/client/server listed.

> I’m not sure how our sponsors that sell servers would feel if we, for
> example, as a majority agreed Prosody was the primary choice for a free
> server and said so, while Matt is on Council. Or even non-sponsors. The
> appearance of an old boys’ club (which I think there is a danger of) seems
> damaging.

Also, I'm not even sure we'd find the right one - because I don't think
there is one.

The real strength of open standards is that it enables choice, allowing
users to obtain the highest suitability for their particular needs - to my
eyes, if the XSF were to recommend a particular server we'd be doing a huge
disservice to our community.

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.jabber.org/pipermail/council/attachments/20150220/d311e7f8/attachment.html>

More information about the Council mailing list