- From: Xan Gregg <xan@extensibility.com>
- Date: Mon, 2 Oct 2000 09:38:50 -0400
- To: xmlschema-dev@w3.org
1) Section 3.2.6 of 9/22 Datatypes contains the following order relation for timeDuration: "for instances x and y of timeDuration, x-y is defined as component-wise subtraction. x > y iff all the components of x-y are non-negative and at least one component is positive. x < y iff all the components of x-y are non-positive and at least one component is negative. x = y iff all components of x-y are zero." For some x and y (for example, P1M and P30D), each of the three relations is false. That goes against common sense, though it's not unreasonable considering the example. If all the relations can be false, I think the doc should say so explicitly. How should a validator interpret min/max facets when all three relations are false? 2) Appendix E defines the syntax for Regular Expressions for use in patterns. The quantifiers allowed are S{n,m}, S{0,m}, S{n,}, and S{0,0}. There is no S{n}. However, S{n} is used in examples in the Datatypes document and the Primer. Which is correct? xan Xan Gregg Senior Architect | TIBCO Extensibility(tm) Office: 919-969-6510 E-mail: xan@tibco.com ____________________________ TIBCO Software Inc. www.extensibility.com | www.XMLschema.com | www.tibco.com
Received on Monday, 2 October 2000 09:42:07 UTC