[XML-DEV Mailing List Archive Home] [By Thread] [By Date] [Recent Entries] [Reply To This Message]

RE: RE: Constraining Data Types: Request for use cases

• To: "'James Youngman'" <jay@e...>
• Subject: RE: RE: Constraining Data Types: Request for use cases
• From: "John F. Madden" <john.madden@d...>
• Date: Wed, 31 Mar 2004 19:06:24 -0500
• Cc: <xml-dev@l...>, <strucdoc@l...>
• In-reply-to: <20040331224135.GA7604@e...>
• Organization: Duke University Medical Center
• Reply-to: <john.madden@d...>
• Thread-index: AcQXcV6n1EdwuzHeQaWAJznEUjZlUQAClFqA

James,

> Suppose I have two XML schemas, "Parent" and "Child".  I want 
> some way of determining the answer to the question
> 
>    'Does there exist any XML document which is acceptable to "Child"
>    but not to "Parent"?'
> 

I don't know if there is, but it's a great question for the strucdocs group
to consider. Several HIMSS implementers (including me) found the full CDA to
be an intractable (direct) target for document generation, and instead tried
an implementation strategy where we pared down the full CDA schema to a
subset, and then tried to use this as the target schema for instance
generation.

My experience was, that didn't work very well. That's why I ended up using
the strategy where I collected data into an application-specific container
(in my case, a document-type specific schema of my own design, which
incorporated all my constraints), and then transformed into CDA-valid format
in a second step.

But if you wanted to do it the first way, you'd want to be able to
"meta-validate" your schema, just as you point out.

John

• Prev by Date: XML conventions
• Previous by thread: RE: Constraining Data Types: Request for use cases
• Index(es):
  • Date
  • Thread