> -1 for QNames in content.

Yeah, I don't really disagree.  But I kinda dislike {uri}lname because it 
looks too much like XSLT's AVT :)

> > Multiple links with the same ³for³ value SHOULD have different ³type²
> > values; processors are allowed to choose the schema language they best
> > support.
> 
> Some questions:
> 
> - What are the types here?

The types of the schema language.
 
> - Which schema languages can validate Atom?

I dunno.  Much of it is probably done via code, as opposed to a schema 
language.  My interest here is in being able to describe, in a 
machine-readable way, the content of an atom:content element.
 
> - What should happen when validation fails (eg in an AtomPub scenario)?

Out of scope; a server MAY reject the post.
 
> - What should an application do when seeing a type it does not 
recognize?

Ignore it, and keep looking in hope of finding a schema language 
(link/@type) that it does understand.
 
> It's not clear what the validation target is - the entry, the content, 
...?

Sorry, I don't understand the question.
 
> Does the rel type requires a co-occurence constraint with @for?
 
Interesting question.  If the @type differs, then no (since a single @for 
target could be described by multiple schema language). If the @type and 
@for are the same at the feed and entry level, then no (since the 
semantics are the entry takes precedence. But is it required that the 
combination of @type and @for be unique at the same "level" (feed or 
entry)?  Not sure.
 
> My concern is that people are going to use this a la DTDs, which means 
> breaking with mustIgnore and applications that worked yesterday not 
> working today.

Right now the only way to define the content of an atom:content element is 
by out of band understanding. I don't see how making it possible to put 
that information in-line breaks anything.  What am I not getting?
 
> If there is associate validation behaviour, can the type not be used to 
> deny service?

Is this a question, or a preference?

        /r$

--
Visiting Member, IBM Academy
STSM, DataPower Chief Programmer
WebSphere DataPower SOA Appliances
http://www.ibm.com/software/integration/datapower/

Reply via email to