- From: <noah_mendelsohn@us.ibm.com>
- Date: Sun, 8 Feb 2004 19:11:30 -0500
- To: ht@inf.ed.ac.uk (Henry S. Thompson)
- Cc: "'Bruno Chatel'" <bcha@chadocs.net>, "Michael Kay" <mhk@mhk.me.uk>, xmlschema-dev@w3.org
Henry Thompson writes: >> Allowing one 'all' to be extended by >> (merged with) another would lead >> to violations of the Extension invariant. >> As it stands, there is an invariant for extension, just as >> there is for restriction, namely >> Extension: Consider a type definition DE derived by extension from a >> base type definition B. Then any sequence of EIIs valid per DE >> contains an initial sub-sequence valid per B. >> [...] Indeed. That's the right invariant for sequences, but not I think for "all". Just as "all" is already different in many respects from other content models, I had always assumed that when and if we allow extension of an all group, we would not do it in an "all-specific" manner. Specifically: I assumed that we would allow an "all" group to be an extension of another "all" group if and only if the extension subsumed the base. Conversely for restriction. Whether we would require you to restate the original model in the extension, or only state the additions, seems to me to be syntactic sugar. I don't much care, but would lean slightly toward explicitly providing only the new content. I >think< we've actually had some user requests for such extension of "all", but unless Michael Kay is making such a request now (I don't think he is), I can't immediately find the reference. I think it's a nice to have, but have no personal urgent need for such a new feature. In any case the fact that it would violate the invariant quoted above doesn't particularly bother me. -------------------------------------- Noah Mendelsohn IBM Corporation One Rogers Street Cambridge, MA 02142 1-617-693-4036 --------------------------------------
Received on Sunday, 8 February 2004 19:15:11 UTC