[Standards-JIG] Re: Re: proto-JEP: Stream Acking

Joe Hildebrand hildjj at gmail.com
Thu Oct 28 16:34:56 UTC 2004


<a:a/> would work ok, if you negotiated a stream feature first, and
restarted the stream with the a: prefix defined in stream:stream.


On Wed, 27 Oct 2004 23:33:35 -0500, Nolan Eakins
<sneakin at semanticgap.com> wrote:
> Justin Karneges wrote:
> > Nothing changed.
> 
> It didn't look like anything did. Maybe the second time around will be
> better.
> 
> > It's funny you bring up single-character acking, considering JEP-0017 came
> > up
> > recently, which also wanted to throw stuff into the <stream> CDATA.  I
> > feel
> > we shouldn't even go there.  We have namespaced elements for a reason.
> 
> I guess throwing in CDATA could cause some problems down the road. Multiple
> enhancements that do this couldn't coexist unless their development was
> coordinated, or if there was a <stream/> CDATA JEP that defined some rules.
> That may not be needed.
> 
> > However we could pick a shorter namespace or a short element name (kinda
> > like
> > JEP-0115) just to reduce size.  You could also declare another namespace
> > prefix in the stream tag.
> 
> Here's some choices using <ack/> as an example and their lengths:
>    <ack xmlns='http://jabber.org/protocol/ack'/>        45
>    <ack xmlns='jabber:ack'/>                            25
>    <ack xmlns='j:ack'/>                                 20
>    <a xmlns='j:ack'/>                                   18
>    <ack/>                                               6
>    <a/>                                                 4
>    A                                                    1
> 
> Take your pick. I personally prefer the later ones since they're small. The
> namespace attribute maybe needed, unless something like '<a:ack/>' or
> '<a:a/>' can be done. I guess that could be done since the XML parser would
> know the namespace. That actually would be preferable, since the 'xmlns'
> attribute uses the most bytes in any of the above examples, and different
> prefixes could be used for different features preventing any element name
> conflicts. Something CDATA can't do.
> 
> - Nolan
> 
> 
> 
> --
> http://www.semanticgap.com/people/sneakin/
> 
> _______________________________________________
> Standards-JIG mailing list
> Standards-JIG at jabber.org
> http://mail.jabber.org/mailman/listinfo/standards-jig
> 


-- 
Joe Hildebrand



More information about the Standards mailing list