[
Lists Home |
Date Index |
Thread Index
]
- From: "Arnold, Curt" <Curt.Arnold@hyprotech.com>
- To: "'xml-dev@xml.org'" <xml-dev@xml.org>
- Date: Tue, 24 Oct 2000 09:56:44 -0600
First of all, if you are making elements that were required in the base type to be optional, that isn't a restriction that a relaxation.
Restriction of element content has been a contentious issue, C. M. Sperberg-McQueen summarized the different approaches that the schema WG considered in
http://lists.w3.org/Archives/Public/www-xml-schema-comments/2000OctDec/0025.html
The current approach requires, as you said, repeating the entire content model (which may have to be concatenated from multiple ancestor types). My feeling is that this is awkward to the point of
avoidance for complicated content models and extremely difficult to implement (to make sure that your derived content model is a true-subset of the base content model).
You may be able to define named groups to capture the unchanging fragments of the content model and build the types using these groups and element references for the few elements that are being
changed.
p.s. Cross-posting is discouraged, but for some reason I didn't see your message show up on xerces-j-dev
|