- From: C. M. Sperberg-McQueen <cmsmcq@blackmesatech.com>
- Date: Sat, 28 May 2011 10:44:09 -0600
- To: Shane McCarron <shane@aptest.com>
- Cc: "C. M. Sperberg-McQueen" <cmsmcq@blackmesatech.com>, public-rdfa-wg@w3.org
On May 28, 2011, at 8:27 AM, Shane McCarron wrote: > It *sort of* helps. We had always assumed that anyURI means IRI. In fact, we built all of XHTML 2 around it. Unfortunately, it doesn't seem that XSD 1.1 is a Rec yet, and my reading of 1.0 is that it allows ipv6 URIs and URIs, but not IRIs. Am I reading that wrong? Yes, I think so. Remember that the term 'IRI' won't appear in XSD 1.0 because it did not exist when XSD 1.0 was issued. At that time, the closest thing to a normative description of what became IRIs was the algorithm in section 5.4 of XLink, which XSD 1.0 cites in describing exactly what the lexical space of the type is. As the definition of IRIs was worked on, some details were changed, so that the set of character sequences in the lexical space of XSD 1.0 anyURI and the set of character sequences accepted by the grammar in RFC 3987 have some differences in what I think most people will regard as edge cases; some more recent specs introduce the concept of LEIRI (legacy encoded IRI) to try to provide a name for the proto-IRIs supported by XLink, XSD 1.0, and some other specs. -- **************************************************************** * C. M. Sperberg-McQueen, Black Mesa Technologies LLC * http://www.blackmesatech.com * http://cmsmcq.com/mib * http://balisage.net ****************************************************************
Received on Saturday, 28 May 2011 16:44:33 UTC