RE: [xml-dev] Separate data from rules ... is the XML Schema 1.1 <as

[rjelliffe@allette.com.au]

Michael Kay wrote
> I don't think the split between
> structural rules and business rules equates at all closely to the split
> between grammar and predicates.

I agree.

But I would say that the demarcation between what constraints to put in
XSD (whether in the grammar or in assertions) or RELAX NG and in a
subsequent Schematron schema (or in derived XSD schemas) should reflect or
align with the institutional arrangements of schema ownership,
responsibility and maintenance.

At one extreme, if a code list (e.g. a product code list or country code
list) is maintained by another organization, then it may be that the XSD
schema should not enumerate the values but just give the regex pattern
they conform to (e.g.  [A-Z]{2-3} ) with the external code list validated
by the separate Schematron schema.

Or, it may be that the purpose of the schema is to collect as many
different constraints as possible in order to be the prime documentation
for developers,regularly updated and maintained. In that case, an
enumeration would be appropriate.

Cheers
Rick Jelliffe