XSchema Spec Section 2.2, Draft 1

Wed Jun 24 14:29:20 BST 1998

>Peter Murray-Rust wrote:
>
>> Unless there is
>> sensible running text whose prime purpose is to be read
by humans there is
>> no particular value in having mixed content (i.e. strings
+ elements mixed).
>
Counter-example from GedML (adapted to make it
self-explanatory):

<PERSON>
<NAME>Frederick BLOGGS
   <SOURCE>Birth Certificate</SOURCE>
   <CONFIDENCE>High</CONFIDENCE>
   <CHANGED>
        <ON>12 Jan 1983</ON>
        <BY>MHK</BY>
   </CHANGED>
   <NOTE>Generally signed himself "Frederik"</NOTE>
</NAME>

The general form is that the "attributes" of an object are
represented by the content of a first-level child element,
and qualifiers to the attributes (source, confidence level,
alternative values, etc) are represented by grandchild
elements.

This is a direct translation of the GEDCOM metamodel. I have
come across similar things in medical data models where the
aim is to record opinions rather than facts: in such a model
we need to record "facts about facts about facts": e.g.
"John says that his mother told him that his father died of
a heart attack".

Unfortunately I am unable to declare in the DTD that the
element contains "PCDATA content followed by zero or more of
the following child elements", XML's peculiar rules mean
that with mixed content I cannot declare any ordering or
cardinality constraints. (I wonder if someone thought that
there was no requirement?)

Mike Kay

xml-dev: A list for W3C XML Developers. To post, mailto:xml-dev at ic.ac.uk
Archived as: http://www.lists.ic.ac.uk/hypermail/xml-dev/
To (un)subscribe, mailto:majordomo at ic.ac.uk the following message;
(un)subscribe xml-dev
To subscribe to the digests, mailto:majordomo at ic.ac.uk the following message;
subscribe xml-dev-digest
List coordinator, Henry Rzepa (mailto:rzepa at ic.ac.uk)