- From: Julian Reschke <julian.reschke@gmx.de>
- Date: Thu, 15 Apr 2010 14:52:57 +0200
- To: Robin Berjon <robin@berjon.com>
- CC: uri-review@ietf.org, public-webapps WG <public-webapps@w3.org>, Arthur Barstow <art.barstow@nokia.com>
On 15.04.2010 14:37, Robin Berjon wrote: > ... > I may be mistaken but my understanding from RFC4395 leads me to believe that reusing the generic syntax (as indicated) is a good thing. > ... Yes, but you still want to state how the generic syntax applies to your particular scheme. The easiest thing to do that probably would be to repeat the ABNF, and then have individial spec sections explain the semantics of each component. >> - It appears that the spec tries to define things in terms of IRIs; my understanding is that what you need to define are URIs (plain ASCII, as per RFC 3986), and then optionally include additional information about how to map from/to IRIs. > > I tried to look for examples of this but failed, which makes me wonder if I'm misunderstanding the request. Would you happen to have a pointer handy? > ... Just define the widget *URI* syntax in terms of RFC 3986 (URI), not RFC 3987 (IRI). Then, state how to map strings that contain non-URI characters to URI syntax (such as UTF-8-encode-then-percent-escape). (This is something that might change when IRIbis is done, but as far as I understand, this is how it works right now). Best regards, Julian
Received on Thursday, 15 April 2010 12:53:38 UTC