[
Lists Home |
Date Index |
Thread Index
]
dareo@microsoft.com (Dare Obasanjo) writes:
>As a datayping
>language W3C XML Schema is the only game in town unless you count
>defunct technologies like Microsoft's XDR. Most of the folks on
>XML-DEV are interested in document validation. Most of my users are
>interested in datatyping.
I've already expressed that I think data typing is a foolish approach to
XML, but that aside -
It seems like W3C XML Schema really fails at data typing in some
important ways.
On the structure side, it tries to offer too much. Type hierarchies are
at the heart of how WXS defines structures, but those hierarchies can be
cut across by substitution groups and then differently by keys. It
seems to combine the restrictions of single-inheritance with the
complications of multiple inheritance and then adds further
complication. (I also don't think extension and restriction can cope
with both object structures and XML issues, though that at least is a
genuinely difficult problem.)
On the data side, the tools for deriving new types are, well,
pathetically weak, because of the limited primitive types and the
complete failure to provide even the simplest of techniques for defining
new pathways from the lexical space to the value space. The pattern
facet is a Swiss Army Knife, but it only goes a little way.
If I was interested in building type systems around XML documents, I
think I'd be even more frustrated with W3C XML Schema than I am today.
|
