Re: [xml-dev] If I want to use catalogs for XSDs, must document instance

[Victor Porton <porton@narod.ru>]

23.04.2012, 23:28, "Liam R E Quin" <liam@w3.org>:
> On Mon, 2012-04-23 at 12:31 +0200, Jirka Kosek wrote:
>
>> šMight be I'm not understanding you correctly Liam, but IMHO namespace
>> šURI is perfect way to *indirectly* say where the schema is located.
>>
>> šGood XML application support mapping file where namespace URI can be
>> šmapped to schema which should be used for validation.
>
> Yes, that aspect is fine. Blindly dereferencing the namespace and
> expecting to find a schema there is what I was saying is not fine.

I am writing a standard about what should be located at namespace URLs.

This should contain not only schemas (of any kind) but also rules for transformations between different namespaces.

I think this is an important standard (when I will release a good draft), but we need work on its promotion to become an agreed industry standard. Your help to promote it could be essential.

http://freesoft.portonvictor.org/binaries/Namespace_Semantic.doc

>> šIn theory on namespace URI could reside RDDL document which can point to
>> šXML schema, but past experience with heavy loaded network resources
>> š(XHTML DTDs) shows that's not wise to depend on such approach in an
>> šapplication.
>
> The XML Plenary failed to get consensus on whether it was acceptable to
> define anything at a namespace URI, or even if it was OK to try to
> dereference them.
>
> If applications did start, you're right, W3C would have to block the
> ones on our servers, as we have to do for the DTDs and XSD files.


-- 
Victor Porton - http://portonvictor.org