• To: =?utf-8?B?J0RhcmUgT2Jhc2Fuam8n?= <kpako@y...>, =?utf-8?B?RXJpYyB2?==?utf-8?B?YW4gZGVyIFZsaXN0?= <vdv@d...>, xml-dev@l...
• Subject: =?utf-8?B?UkU6IFt4bWwtZGV2XSBXaGVyZSBoYXZlIHRoZSBlbGVtZW50IHR5?==?utf-8?B?cGVzIGdvbmU/?=
• From: =?utf-8?B?SmVmZiBMb3dlcnk=?= <jlowery@s...>
• Date: Fri, 7 Dec 2001 16:25:00 -0800

The point I would make is that schema types give one the freedom to name
their element according to semantic context, while allowing said elements to
be processed as the same definitional type. Fer instance:


<rectangle>
	<vertices>
		<upperRight>23 34</upperRight>
		<lowerLeft>11 44</lowerLeft>
	</vertices>
</rectangles>

<peanuts>
	<count>2</count>
</peanuts>

Here, all numbers can be processed as positiveInteger types instance (or
lists thereof), regardless of context. On the other hand, the application
can attach different semantic meaning to the data represented by those typed
values. I can separate the validation of the type from it's meaning in a
particular context.


 

• Prev by Date: RE: images and urls in xml
• Next by Date: RE: Attribute Order
• Previous by thread: List-Unsubscribe
• Next by thread: RE: Re: validating hairy data models (was Attribute order)
• Index(es):
  • Date
  • Thread