- From: Elliotte Rusty Harold <elharo@metalab.unc.edu>
- Date: Sun, 11 Apr 2004 11:45:44 -0400
- To: public-xml-id@w3.org
A few random thoughts on a Sunday morning. I was reading through the XML Schema 1.1 requirements <http://www.w3.org/TR/2003/WD-xmlschema-11-req-20030121/> today, when I noticed the following proposal: 2.1.3.4 Allow complex types more than one ID attribute (RQ-105) Relax the constraint that a complex type may contain at most one attribute of type ID. See http://lists.w3.org/Archives/Public/www-xml-schema-comments/2002JulSep/0069.html I'm not sure what sections xml:id exactly this would affect if adopted, but it does seem like it might well affect something in the xml:id spec, so I just thought I'd toss this out here, and see what people thought. There is an implicit assumption throughout the xml:id spec that each element has and should have at most one unique ID type attribute. Indeed, a lot of verbiage is expended considering how to handle the case where this is not true. But what if we were to remove this requirement completely? Currently, it is not possible for a *valid* element to have two attributes with ID type. However it is definitely possible for an *invalid* document to have two such attributes. Neither the infoset nor XML 1.0/1.1 prohibits this. In the future, it seems it may be possible for a schema-valid document to have multiple ID type attributes. What if instead of trying to avoid this case, xml:id embraced it? Anyway, I'm not sure exactly what text this might change in the xml:id spec; maybe the bit about not declaring non-xml:id attributes as type ID. But it might adjust how one thinks about some of these issues. -- Elliotte Rusty Harold elharo@metalab.unc.edu Effective XML (Addison-Wesley, 2003) http://www.cafeconleche.org/books/effectivexml http://www.amazon.com/exec/obidos/ISBN%3D0321150406/ref%3Dnosim/cafeaulaitA
Received on Monday, 12 April 2004 06:22:47 UTC