RE: [ubl-dev] SBS and Restricted Data Types

[stephen.green@systml.co.uk]

re: http://www.oasis-open.org/archives/ubl-dev/200605/msg00010.html

David

Thanks for this. I'll digest it carefully. It exemplifies well my
view that as much as possible implementation has to be left to
the implementers - except where state alignment requires more than
that. So we provided the specification in a neutral way (in SBS's
case using an XML instance rather than prose with a design to
facilitate transformation to machine readable formats as needed,
or to prose if necessary). Then we let folk generate Scehamtron,
CAM or code such as SAX and perl according to their preferences
and needs. This seems the thing to do unless we find the need (for
better state synchronisation and interoperability) to also provide
the Schematron, CAM, etc (as with the provision of the UBP for
ebXML use - or as an example for the equivalent in any other means).

Any thought of a CAM profile specifically for codelists and subsets?
These seem to be, regarding UBL at least, key areas of need and it
might be good to optimise CAM for those two areas with a profile.
The provided examples are great though. Refining these to fit the
exact use cases seems a great idea and perhaps even easier to follow
for implementers than a profile.

Ultimately I stiil have a small hope that XSD might be improved to
cater for 1. derivation of enumerations and 2. a way to say (and maybe
validate/verify) that a schema is a strict subset of another schema
of which it therefore uses the same namespace. Too late for XSD 1.1?
Anyone in earshot on the W3C working group? :-)

All the best

Steve