- From: Richard Tobin <richard@cogsci.ed.ac.uk>
- Date: Fri, 21 Mar 2003 16:01:10 GMT
- To: www-xml-linking-comments@w3.org
The fact that an element can be identified by a child that is a schema-determined ID makes it impossible to search for a match in a fully streaming manner. For example, if the <id> element is of schema type ID, then in <foo> ... millions of child elements ... <id>chapter1</id> <foo> a processor that is asked to return the "chapter1" element will have to buffer the millions of child elements. Worse, it would have to buffer them even if <foo> turned out not to have an <id> child with the right value. A processor capable of schema type analysis may be able to optimise this (e.g. an element that does not allow a child that can be of ID type, and which is known to be valid, cannot have such an ID), but this requires a lot of work. -- Richard
Received on Friday, 21 March 2003 11:01:13 UTC