OASIS Mailing List ArchivesView the OASIS mailing list archive below
or browse/search using MarkMail.


Help: OASIS Mailing Lists Help | MarkMail Help

relax-ng message

[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]

Subject: Re: [relax-ng] Re: Associating documents and schemas

> As far as I can tell, the W3C XML Schema spec did a reasonable job of
> claiming that the xsi:schemaLocation is merely a hint and a processor is
> allowed to do anything. Everyone in the WG has the same interpretation
> on this area of the spec.
> Yet somehow parsers don't really seem to provide a nice way to ignore
> hints, and people seems to think as if @xsi:schemaLocation is the only
> authoritative way to use XML Schema validation.
> I understand that neither James nor Murata-san want the association
> mechanism to be treated like that. And I suppose that means we need to
> do more than what the W3C XML Schema spec did, to make sure that they
> are not the only way.

I don't think there's much danger of our ending in W3C XML Schema's 
situation at this point, for a whole variety of reasons:

- If we do this, it will be a separate spec.   I think this is 
important: the XSLT spec mentions xml-stylesheet, but implementors don't 
typically support only xml-stylesheet as a document/stylesheet mechanism.

- The association spec will come out at a different time.

- There are current implementations and APIs which all provide a way for 
the user to explictly specify a schema.

- Knowledge about xsi:schemaLocation *has* to be built into a W3C XML 
Schema implementation, because it has to validate these attributes 
specially (which also means it loses the ability to use the schema to 
control usage of xsi:schemaLocation).

- We will, I hope, provide more than one mechanism (I think processing 
instructions and root element attributes are complementary: one is not 
controllable by the schema; one is.)

- The wording in the W3C XML Schema spec is much stronger than just a 
hint: "Unless directed otherwise, ..., processors should attempt to 
dereference each schema document location URI in the ·actual value· of 
such xsi:schemaLocation and xsi:noNamespaceSchemaLocation [attributes]"

- With W3C XML Schema, the information specifying the schema is not just 
a single URI, it's complex structure with a a namespace URI/schema 
document and the no namespace schema document.  W3C XML Schema doesn't 
provide an interchange syntax for this other than the xsi attributes. 
With us, the only information needed for the schema is a single URI.


[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]