(cannonical model)

Didier PH Martin martind at
Thu May 27 01:26:57 BST 1999


you said:
Do you have any thoughts about the cannonical model spec at

To me it seems as a reversal of all the good things that XML modeling
offer and a move toward network database model. Here is what looks very
offending to me:

1. All relationships are represented as IDREFs.
Apart from the current XML problem, that IDREFs are untyped, this
sounds like a very inflexible approach to me. Many relationships can
be modeled by subelements. That way you can balance replication and
efficiency (checking subelements is a local operation and should be
much more efficient than traversing links).

2. All relationships are materialized.
By keeping explicit IDREF pointers all the relationships of the ER
schema are materialized. This sounds quite alarming.

3. XML elements are not used to model data at all.
I.e. only the entity is an element, with no children. Isn't XML's
goal to avoid these flat unnested representations?

Is it obvious that the cannonical model proposal has some serious
problems, or am I not getting something?

Didier says:
Do you have a particular format in mind like HL7 for instance? as you know,
biztalk is only the envelope as cXML also make some envelope proposal too.
but this envelope contains more specific formats like for instance HL7.

>From waht I know, most XML-EDI format do not convey ER relationship only
transaction documents like for example purchase orders. In this latter case,
containment relationship is provided (as in HL7)

Didier PH Martin
mailto:martind at

xml-dev: A list for W3C XML Developers. To post, mailto:xml-dev at
Archived as: and on CD-ROM/ISBN 981-02-3594-1
To (un)subscribe, mailto:majordomo at the following message;
(un)subscribe xml-dev
To subscribe to the digests, mailto:majordomo at the following message;
subscribe xml-dev-digest
List coordinator, Henry Rzepa (mailto:rzepa at

More information about the Xml-dev mailing list