- From: Don Chamberlin <chamberl@almaden.ibm.com>
- Date: Thu, 29 Apr 2004 16:53:31 -0700
- To: mrys@microsoft.com
- Cc: public-qt-comments@w3.org
- Message-ID: <OF1138236C.6BA4BD89-ON88256E85.00823AF7-88256E85.00832856@us.ibm.com>
Dear Michael, Thank you for your comment MS-XQ-LC1-056 on XQuery Version 1.0. The Query working group considered this comment on 4/15/04. The working group observed that an implementation is allowed to infer a more specific static type for any expression than the static type indicated in the XQuery specification (this feature is called a "static typing extension"). The next edition of the XQuery specification will include the following non-normative note: "It is not recommended for a static typing extension to change the static type behavior of the 'treat' expression, since static type behavior is central to the semantics of this expression." Despite this note, you may wish to consider using a static typing extension when processing path expressions with a leading slash. Regards, --Don Chamberlin ----(referenced comment)---- Section 3.2 Path Expressions Technical fn:root(self::node()) treat as document-node() is not precise enough for meaningful static typing of path expressions of the form /a/b/c. We should use a type that is given via the static context as a new property such as the static type of a built-in context root item (similar to the notion of a context item).
Received on Thursday, 29 April 2004 19:54:05 UTC