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

[stephen.green@systml.co.uk]

re: http://www.oasis-open.org/archives/ubl-dev/200605/msg00015.html

Thanks Chee-Kai

A very interesting point. At first I thought it would always be
that the subset is strictly a subset of the schemas * before *
redefinition but then I realised that would prevent the use of
redefine for customisation. It might still be prudent to use
substitution or just redeclaration (a new schema which doesn't
use XSD derivation but might still be a logical derivation of
another schema) rather than redefine if one is going to create
a subset based on the resulting schema. That is because, without
a change of namespace (redefine doesn't allow a change of namespace
but the substitution group mechanism of XSD customisation requires
one) one has this problem when defining a subset - what actual schema
is it a subset of? One way might be to require: version, namespace
and schemaLocation attributes all to be provided; or at least namespace
and schemaLocation. This is similar to the requirement we considered
in ebBP when considering how to identify a subset as part of a logical
document's specification. The anwser acccepted was like the last one:
namespace and schemaLocation along with an 'externalDocumentDefRef'
which can be used to refer to the subset definition itself. So as
well as a schema, a subset could be a strict subset of another subset.
That too would need consideration (but not so much if the first subset
was/could itself be defined as a type of XSD schema).

I like the use of the mathematics too.

All the best

Steve