- From: Marcos Caceres <marcosc@opera.com>
- Date: Thu, 10 Sep 2009 12:48:04 +0200
- To: Marcin Hanclik <Marcin.Hanclik@access-company.com>
- Cc: Robin Berjon <robin@berjon.com>, public-webapps WG <public-webapps@w3.org>
On Tue, Sep 8, 2009 at 5:47 PM, Marcos Caceres<marcosc@opera.com> wrote: > > > Marcin Hanclik wrote: >> >> Hi Marcos, >> >> Thanks for your comments. >> It seems we are aligned. >> >>>> UAs will just have to deal with that internally >> >> I assume there could be an easy solution to the normalization: >> The normalization / mandating some equivalence-determining algorithm (from >> RFC3987) could go into P&C. >> Then maybe I18N would not have to be bothered further. > > Ok, basically, to address this all we need to say is: > > If 'potential license href' is a valid IRI, then let 'widget license href' > be the result of normalizing 'potential license href' in accordance with > RFC3987. > > Same for author href. > > (Personally, I think this change is unnecessary because of the fact that > implementers know these are IRIs already). > > Normalization would not apply to feature name or to the id. They are treated > as XML Namespaces (i.e., opaque strings that conform as IRIs). > > In other words: > > id = "http://www.example.org/r%E9sum%E9.xml" > > is not equal > > id = "http://www.example.org/résumé" > > Try it out: > > <test > xmlns:y = "http://www.example.org/r%E9sum%E9.xml" > xmlns:test = "http://www.example.org/résumé" >> > <y:aaa/> > <test:test/> > </test> > Woops, forgot to trim the .xml part :P -- Marcos Caceres http://datadriven.com.au
Received on Thursday, 10 September 2009 10:49:04 UTC