RE: Versioning of XML Schema and namespaces

Personally I think the combination of a namespace and a version attribute in
the document element is a better way of asserting conformance to a
particular version of an external standard than use of a schemaLocation. The
reason for that is that the semantics of xsi:schemaLocation as defined in
the XML Schema specification don't say it's an assertion about conformance,
they say it's a hint about where to find a schema to use for validation.  I
think it's a mistake to overlay different semantics onto an attribute
outside your control; and more practically, using this attribute as a
conformance assertion prevents people using it to point to a local copy of a
schema.

Michael Kay
http://www.saxonica.com/ 

> -----Original Message-----
> From: John.Hockaday@ga.gov.au [mailto:John.Hockaday@ga.gov.au] 
> Sent: 09 May 2005 01:24
> To: mike@saxonica.com; xmlschema-dev@w3.org
> Cc: ":www-xml-schema-comments"@w3.org
> Subject: RE: Versioning of XML Schema and namespaces
> 
> Michael,
> 
> Thanks for your comments.  Please see below one aspect that I 
> disagree with.
> 
> Thanks to you all for being patient and contributing to this 
> discussion.
> 
> 
> John
> 
> > -----Original Message-----
> > From: Michael Kay [mailto:mike@saxonica.com] 
> > Sent: Friday, 6 May 2005 8:26 PM
> > To: Hockaday John; xmlschema-dev@w3.org
> > Cc: ":www-xml-schema-comments"@w3.org
> > Subject: RE: Versioning of XML Schema and namespaces
> > 
> > 
> > ....
> > 
> > I wouldn't expect the instance document to contain a schemaLocation
> > attribute, or if it does, I wouldn't expect a recipient to 
> > trust it when
> > doing validation. If say the sender has decided to leave out 
> > some mandatory
> > elements, and to create a local copy of the schema that makes 
> > them optional,
> > you as the recipient don't want validation to succeed.
> > 
> 
> The metadata gateway that I manage will be able to search metadata of
> multiple types.  For example; 
> ANZLIC version 2 defined by 
> http://www.environment.gov.au/net/dtd/anzmeta-1.3.dtd , 
> extensions to this standard such as 
> http://www.indexgeo.net/dtd/anzmeta-resource-v11.dtd and
> http://www.gso.qld.gov.au/qsiis/dtd/qsiis-1.3.dtd 
> or ISO 19139 metadata records including extensions of that 
> standard, which
> are yet to be defined.
> 
> If I download an XML metadata record (document instance) I 
> don't know what
> standard I should validate it against unless there is a DOCTYPE or
> schemaLocation declaration.  I therefore can't validate that 
> document without
> this information.  I believe that *every* XML document 
> instance should have
> either a DOCTYPE or a schemaLocation so that anyone who wants 
> to look at this
> instance knows with what XSD it complies.  One may not wish 
> to validate the
> document but if one wants to use the document then he or she 
> needs to know
> what this document is about.  That can only be rigorously 
> identified by a
> DOCTYPE or schemaLocation declaration.
> 
> Furthermore, I agree that if a "mandatory" element in an 
> extension of the ISO
> 19139 metadata standard has been redefined as "optional" then 
> validation
> should ignore that change. However, by definition of the 
> extensibility of the
> ISO 19115 metadata standard, an "optional" ISO 19139 element 
> can be made
> "mandatory" in an extension of that XSD for an organisation's 
> or country's
> need.  Therefore, one *must* use that extension when validating an XML
> document instance of that XSD type to check that it meets 
> those needs.  The
> only way I know of to identify that extension is from a schemaLocation
> declaration and therefore is it necessary to validate that 
> XML document
> instance.
> 
> Public identifiers were great.  There was only the need for 
> one DOCTYPE
> declaration in an XML document instance.  It was the inclusion of the
> original DTD in the DTD identified by the DOCTYPE that 
> defined the extension.
> It wasn't necessary to have the original DTD identified in 
> the XML document
> instance.
> 
> However, and I'm not sure about this, it seems that if one 
> extends a W3C XML
> Schema then one needs to not only identify the new XSD via a 
> namespace but
> one also needs to include the original XSD that has been 
> extended.  If this
> is so then XML document instances can become *very* messy 
> when there is an
> extension of an extension of an extension etc..  
> 
> Thanks for your time.
> 
> J.H.
> 
> > Michael Kay
> > http://www.saxonica.com/
> > 
> > 
> > 
>

Received on Monday, 9 May 2005 08:10:16 UTC