- From: Anne van Kesteren <annevk@opera.com>
- Date: Mon, 23 Mar 2009 18:05:17 +0100
- To: "Henry S. Thompson" <ht@inf.ed.ac.uk>, "Dan Connolly" <connolly@w3.org>, "C. M. Sperberg-McQueen" <cmsmcq@blackmesatech.com>
- Cc: "HTML WG" <public-html@w3.org>
On Mon, 23 Mar 2009 17:04:45 +0100, Henry S. Thompson <ht@inf.ed.ac.uk> wrote: > Any reason why you don't reference the LEIRI Note [1]? It was > published by the XML Core WG in the face of delay to a new IRI RFC, > which will define Legacy Extended IRIs. This somewhat cumbersome name > is what XML, XLink, XML Base, the XML Stylesheet PI and so on all use > or will use for what is allowed when an IRI (or URI) (reference) is > wanted. A simple escaping algorithm is specified for converting LEIRIs > to IRIs, from whence the existing IRI can be converted to a URI (and > the conversions can of course be composed.) > > At the very least, could you please consider replacing the "IRI > reference" with "LEIRI reference" in the last three of the four > bullets at the beginning of section 1. Any 3986-defined URI is a > 3987-defined IRI, and every 3987-defined IRI is a LEIRI. Those bullet points are defining what is "valid". I thought LEIRIs are not supposed to be used? > At best, please specify the mapping as a (much simpler) mapping of > arbitrary strings to LEIRIs, and let the LEIRI specification take over > thereafter. That would not work because RFC 3987 has rules that depend on the character encoding of the input document that this specification is trying to avoid. (Ironically it does introduce such a dependency of its own.) Well, it could work, but it would essentially be the same set of steps and therefore not be much simpler. > [1] http://www.w3.org/TR/2008/NOTE-leiri-20081103/ -- Anne van Kesteren http://annevankesteren.nl/
Received on Monday, 23 March 2009 17:06:12 UTC