[Standards] dialback feature

Dave Cridland dave at cridland.net
Thu May 19 11:31:48 UTC 2011


On Thu May 19 08:27:05 2011, Jacek Konieczny wrote:
> On Wed, May 18, 2011 at 10:12:21AM -0600, Peter Saint-Andre wrote:
> > <dialback xmlns='urn:xmpp:features:dialback'/>
> > <dialback xmlns='urn:xmpp:features:dialback:errors'/>
> > <dialback xmlns='urn:xmpp:features:dialback:muliplexing'/>
> > <dialback xmlns='urn:xmpp:features:dialback:piggybacking'/>
> 
> Why four different namespaces for one protocol? Why not four  
> different
> elements in a single namespace? Or an element with subelements
> (options)?

I entirely disagree with the argument that follows, but for the  
record:

The argument is that because different options are introduced at  
different times, and being options may not be implemented, then this  
results in a volatile schema with respect to time.

That is, at different times, and in effect on different  
implementations, the schema may have different elements potentially  
available.

Now, I disagree that this is an issue to be concerned about, because  
the elements present in the schema of whatever time and place all  
have exactly the same semantics. No matter the sender's idea of the  
schema, elements are either unknown or have a defined and well-known  
semantic associated with them. We do not do schema validation in  
XMPP, for good reason, so this will not have any effect on  
interoperability.

Dave.
-- 
Dave Cridland - mailto:dave at cridland.net - xmpp:dwd at dave.cridland.net
  - acap://acap.dave.cridland.net/byowner/user/dwd/bookmarks/
  - http://dave.cridland.net/
Infotrope Polymer - ACAP, IMAP, ESMTP, and Lemonade



More information about the Standards mailing list