[
Lists Home |
Date Index |
Thread Index
]
At 05:57 PM 5/7/2002 +0100, Ian Fitchet wrote:
> Then, if two systems want to work within a different strongly typed
> context, what should there be to stop them? Consider:
>
><xml type-context="some-common-standard">
><date>7/5/2002</date>
>
> and
>
><xml type-context="my-cool-types">
><date>Jurassic +500</date>
>
> Both ``dates'' are perfectly valid *within their own contexts* and
> neither has any reason to claim superiority over the other. They can
> operate quite independently of one another and there's no reason why
> they can't interoperate (given some context).
XML Schema allows you to define your own date types, but the semantics of
these date types will not be known. We discussed this at length, and came
to the conclusion that there are many calendars, but most people who use
these calendars knows how to convert between them and the modern Western
calendar. Having a common interchange format with known semantics is
extremely useful in practice.
XQuery knows how to sort XML Schema dates, but it does not know how to
compare <date>Jurassic +500</date> to <date>Paleolithic -500</date>.
Personally, I would rather not make our type system any more complex than
it is right now, so extending it in this direction is not something I would
support.
On the other hand, I *do* support use of schema extensibility mechanisms
like the Schema Adjunct Framework to define mappings among date systems.
Jonathan
|
