- From: Martin Duerst <duerst@w3.org>
- Date: Mon, 05 May 2003 16:42:34 -0400
- To: "Julian Reschke" <julian.reschke@gmx.de>, "Rick Jelliffe" <ricko@topologi.com>, <www-tag@w3.org>
At 18:27 03/04/30 +0200, Julian Reschke wrote: > > From: www-tag-request@w3.org [mailto:www-tag-request@w3.org]On Behalf Of > > Rick Jelliffe >Rick, > >that's correct. My concerns are/were: >- it may break code, in particular if IRI allows some ASCII characters which >are commonly used as delimiters *between* namespace names (I assume that >this is going to be fixed) As Eliotte has reported, and as documented at http://www.w3.org/International/iri-edit#iadditional-00 this is currently fixed. Please have a look at the current wording in http://www.w3.org/International/iri-edit/draft-duerst-iri.txt and tell me if that addresses your issue. And by the way, it turns out that most of the namespace 1.0 implementations, in particular in XSLT-land, already work fine with IRIs. >- the benefit is unclear: if somebody really want's to *refer* to his >namespace using localized characters, he *can* do that. Just use the IRI, >and specify that the actual namespace name is the mapped URI (just like >you'll need to do the same transformation when for instance mapping a IRI to >an HTTP URL). The benefit of using what you call 'localized' characters directly may be unclear to you, but it's very clear to whoever is actually using them. If you had the alternative between using Greek letters, using some weird numbers and symbols (i.e. %-hh), and using the Latin letters you are used to, the benefit would be very clear to you, too. Regards, Martin.
Received on Monday, 5 May 2003 17:24:49 UTC