Monday, January 17, 2005, 6:39:49 AM, you wrote:


> On Jan 15, 2005, at 9:30 AM, David Powell wrote:

>> PaceExtensionConstruct hopes to provide a basis for a mapping Atom to
>> models such as RDF [1], ER, and OO. I'm currently doing some work to
>> see
>> whether there is anything in Atom that makes this unnecessarily
>> difficult.
>>
>> I think that Atom's use of xml:lang is likely to be a significant
>> problem to many implementors. I've posted PaceSimpleLanguageTagging
>> explaining the problem.

> I'm fairly unconvinced that this is will loom very large in the 
> universe of implementor problems.

For an atom implementation where content is entered via HTML forms, or
from an existing CMS - then this isn't a problem.  The user won't be
able to scatter xml:lang's everywhere.  But for an implementation
based on Atom Protocol, the server needs to preserve all of this
stuff.

> If it turns out that it's really an
> issue, couldn't you achieve the desired effect by restricting xml:lang
> to appear only on atom:entry and atom:feed?

See my reply to Martin.  Restricting the position of xml:lang could
make things worse, unless we also restrict its inheritance - which I'm
not sure that we are allowed to do.

What do you think about only allowing it on atom:content and on Text
constructs?

> I have to admit that I'm having trouble imagining a scenario where 
> you'd want to scatter xml:lang's randomly around an atom instance, so
> maybe this is a handy simplification.

I think that if we allow xml:lang then it should definitely be
restricted. The current "xml:lang everywhere" situation is only simple
to implement if you assume that your implementation stores all of it's
data in an XML DOM.

> Obviously, all bets are off once you get inside atom:content, right?

Well, HTML and XHTML can include lang, and xml:lang according to their
specs.  They inherit the default language from their envelope.  I
think that is pretty well defined.


-- 
Dave

Reply via email to