[
Lists Home |
Date Index |
Thread Index
]
Dare Obasanjo <dareo@microsoft.com> wrote:
> Type derivation in XSD comes in two flavors; extension and restriction.
> Extension is easy to mimic even with DTDs. However how does RELAX NG
> mimic restrictions (especially the nested ones) that involve facets like
> xs:minInclusive, xs:maxInclusive?
For most projects, I would wager that complex types are almost exclusively
derived be extension since that's typically how we derive classes in our
object-oriented programming languages, wouldn't you agree? Especially when
we're talking about the complex types generated from C# (or Java) classes
for WSDL documents.
So if all we're doing is data binding to object-oriented classes (which is
what I see as the most useful aspect of XSD) then we might also find
ooRelaxNG [1] a perfectly suitable replacement for XSD. ooRelaxNG should be
easier to convert to XSD than standard RELAX NG since it's based on types
(classes) and not patterns.
According to the WSDL Note, its types element does allow for content from
any namespace. It would sure be nice if some toolkits stopped hardcoding
support for XSD in that space.
Jason
[1]
http://www.asahi-net.or.jp/~eb2m-mrt/oo.html
|
