Re: WebDAV spec should explicit say WebDAV XML is not valid

On 23 Jul 98, at 19:04, Jim Davis wrote:

> And lest you have any doubt, the spec contains examples of invalid
> (although well-formed) XML.  For example the prop XML element is defined
> as ANY, and the  XML spec (section 3.2) says that ANY means "any declared
> element type", but in general, client properties will not have been
> declared in the DTD, hence this XML will be well formed but not valid.

The rest of the document could, however, be valid.

> It's also a pity that there's no way in XML (that I see) to declare an
> element's contents to truly be "any".  This would allow you to validate
> WebDAV XML in those places that mattered (e.g. propertybehavior) and not
> in those places that are open ended (resourcetype).  But this can't be
> helped.

This is one of those things that the (I hope) forthcoming work on 
schemas in XML will address. I certainly think it would be generally 
useful to be able to validate part of a document, and let the content 
model of the rest of the document be relatively open.

> It would be a pity if applications tried to validate the XML and rejected
> it.  We can forstall some confusion by stating plainly that the XML is not
> expected to be valid.  Of course this also makes providing the XML DTD
> somewhat pointless.

Not really, since the DTD is at least a guide for the rest of the 
document, and a sufficiently sophisticated application should be 
able to validate that part of the document that does match the DTD, 
even though it would not be able to validate the entire document.


Lauren

Received on Thursday, 23 July 1998 23:41:47 UTC