- From: Jonathan Rees <jar@creativecommons.org>
- Date: Tue, 2 Sep 2008 17:13:20 -0400
- To: Shane McCarron <shane@aptest.com>
- Cc: Ben Adida <ben@adida.net>, public-rdf-in-xhtml-tf@w3.org, Noah Mendelsohn <noah_mendelsohn@us.ibm.com>
I may not have made it clear that my recent comments were about the CURIE draft, not about the RDFa draft. It appears the TAG is taking up the CURIE issues, and will certainly provide better guidance than I alone would be able to. Regarding RDFa, I am less concerned because it says nothing normatively about XML Schema datatypes. There is only one mention each of "lexical space" and "value space", without any hint that these terms have anything to do with XML Schema. I would be happier if there were no allusion at all, but I don't think anyone will get too confused by this sentence, especially if later the CURIE document justifies it and makes it rigorous. Likewise, Appendix B is only informative, and I think you will be forgiven if it needs to be refined by the CURIE spec. There is a possible inconsistency in that XML Schema datatypes defined using an XML Schema definition like the one given cannot, as far as I know, specify novel lexical-to-value mappings, and my guess (again, I'm not expert on this stuff!) is that absent anything that modifies the mapping (prose? code?), the mapping would be inherited from the type (xs:string) of which the new type is a restriction. In this case the mapping would be the identity function and the value space would be the same as the lexical space - not limited to IRIs. But as the appendix doesn't even hint at anything about the mapping or the value space, I think most readers will simply accept that this issue simply isn't addressed. I know how close you are to publication and I would not worry too much about this. The CURIE spec will fix anything that needs to be fixed. Again, speaking for myself only. Best Jonathan
Received on Tuesday, 2 September 2008 21:14:05 UTC