- From: Lisa Martin <lmartin@ca.ibm.com>
- Date: Sun, 15 Feb 2004 13:28:08 -0500
- To: public-qt-comments@w3.org
- Cc: W3C XML Schema WG <w3c-xml-schema-wg@w3.org>
Dear Colleagues, This comment pertains to the Nov. 12 2003 version of XPath 2.0 [1]. [1] http://www.w3.org/TR/xpath20/ Lisa Martin, on behalf of the XML Schema Working Group ---------------------------------------------------- Section 2.4.4 Sequence Type Matching In the definition of type-matches in 2.4.4, known and unknown derived types are treated differently with respect to derivation by extension and restriction. Implementations are free to return true if they can determine that unknown types are derived by restriction but not if they are derived by extension or a mix of extension and restriction steps, whereas known types will return true if derived by any means. If this non-parallelism is intentional, an explanatory note might help avoid confusion on the part of readers. If it is not intentional, it should be fixed. In addition, the rules are careful to say "derived by one or more steps of restriction or extension" but say simply "derived by restriction". Is it intended that this be confined to a single derivation step? Or would one or more restriction steps be OK?
Received on Sunday, 15 February 2004 13:28:22 UTC