[Council] VOTE: JEP-0020 (Feature Negotiation)

Peter Saint-Andre stpeter at jabber.org
Mon Oct 21 15:22:59 CDT 2002


OK, I've been thinking about this some more. I'm not so sure of the re-use
argument in the MUC instance because *all* of the forms (except the room
configuration form) use the same structure:

<iq type='result'>
<query xmlns='muc#admin'>
<item [affiliation | role]='foo'/>
...
</query>
</iq>

So you don't need a custom GUI builder in order to edit each list type,
you just need one. Therefore we've got re-use across all of MUC at that
point a much simpler protocol just using muc#admin and muc#owner rather
than x:data.

Thoughts?

Peter

--
Peter Saint-Andre
Jabber Software Foundation
http://www.jabber.org/people/stpeter.php

On Mon, 21 Oct 2002, Peter Saint-Andre wrote:

> Well, it seems there is more to it than that. For example, there is the
> issue of re-use, which pgm brought up to me regarding MUC. We could
> definitely edit all the admin/voice/ban/etc. lists using the MUC format,
> but if we've already got x:data we might as well send that data to the
> generic form-builder, even though the data format is well-defined (it's
> just a list of <item/> elements with various role and affiliation
> attributes). By the criteria discussed here, using x:data for this is a
> bad thing.
> 
> Peter
> 
> --
> Peter Saint-Andre
> Jabber Software Foundation
> http://www.jabber.org/people/stpeter.php
> 
> On Mon, 21 Oct 2002, Joe Hildebrand wrote:
> 
> > Yes, we decided on these guidelines.  Mostly I just wanted to have this
> > discussion in public.  :)
> > 
> > -- 
> > Joe Hildebrand
> > 
> > > -----Original Message-----
> > > From: Peter Millard [mailto:me at pgmillard.com]
> > > Sent: Sunday, October 20, 2002 4:42 PM
> > > To: council at jabber.org
> > > Subject: Re: [Council] VOTE: JEP-0020 (Feature Negotiation)
> > > 
> > > 
> > > Temas speaks for me well in this case :)
> > > 
> > > I thought we decided that x:data was great for cases where 
> > > the schema of the
> > > data was unknown. But for such a simple case, it seems like 
> > > we don't need to
> > > add the xtra layer of abstraction which x-data provides. We 
> > > know we just
> > > need to pick options... we know the schema. There is no need 
> > > for x-data as
> > > the primary transfer layer.  JoeH and I discussed this at the time wrt
> > > disco, and I thought we both agreed at that time that this 
> > > was a reasonable
> > > rule:
> > >     - Use x-data if we don't know the type of data or schema 
> > > which is used
> > > to present/capture the data.
> > > 
> > > pgm.
> > > 
> > > 
> > > _______________________________________________
> > > Council mailing list
> > > Council at jabber.org
> > > http://mailman.jabber.org/listinfo/council
> > > 
> > _______________________________________________
> > Council mailing list
> > Council at jabber.org
> > http://mailman.jabber.org/listinfo/council
> > 
> 
> _______________________________________________
> Council mailing list
> Council at jabber.org
> http://mailman.jabber.org/listinfo/council
> 




More information about the Council mailing list