Re: XSchema validity (was: root element)

Simon St.Laurent (SimonStL@classic.msn.com)
Fri, 3 Jul 98 16:00:30 UT


Michael Kay writes:
>>I meant the barfing semantics. Valid was clearly a poor
>>choice of words.
>
>This is going to get worse. Perhaps we should use:
>
>well-formed: as defined in XML 1.0 (loosely, matching tags
>etc)
>
>valid: as defined in XML 1.0 (loosely, conforms to its own
>DTD)
>
>conforms to XYZ: conforms to the rules of standard XYZ (e.g.
>XML-Namespace). This may of course be an
>application-oriented (anti-barfing) standard
>
>obeys ABC: conforms to the constraints specified in XSchema
>ABC

This is a great start; as I revise the XSchema sections I'll try to make sure
these terms are used _consistently_.

>These are predicates that can be applied to any XML document
>including, of course, an XSchema. For an XSchema [document]
>to be conformant to the XSchema standard if must be
>well-formed, it must be valid under the XSchema DTD, and it
>must meet additional constraints described in the text of
>the XSchema standard.

Here I disagree. I don't want to conflate validation against a DTD with
checking (verification? still fishing for a good word) against an XSchema.
The process is similar, but not identical.

I also need terms to describe the piece of software that checks a document
against an XSchema. So far I've been fumbling with XSchema processor;
Verifier or something else might be better.

>An interesting question: is it an objective to allow all
>[reasonable] "conformance" rules for an application to be
>expressed as XSchema constraints?

I'd like to think so; at least in part that's why there's that crazy XSC:More
element for application designers to experiment with.

Simon St.Laurent
Dynamic HTML: A Primer / XML: A Primer / Cookies