Re: New Version Notification for draft-nottingham-http-link-header-02

Anne van Kesteren wrote:
> I'd suggest not to pretend that everything is a URI. HTML5 does allow 
> URIs (if that's a concern) as tokens as the only requirement is that a 
> token consists of at least one character and no whitespace characters.

URIs give you the infrastructure for disambiguating relation names 
without a central registry. Some consider this a feature.

>> Is there a technical argument behind that, or is it just personal 
>> preference? IANA is well-recognised, has processes in place for change 
>> control, is accountable for availability, continuity, etc. and is 
>> backed by a stable financial structure. I don't see any benefit to 
>> making an exception for one type of registry when every other one on 
>> the Internet uses IANA, but maybe I'm missing something.
> 
> The reason is that it makes managing the registry less centralized. As 
> in, everyone can easily propose something new that is then subject to 
> community review on whether it will be endorsed or unendorsed. See
> 
>   http://wiki.whatwg.org/wiki/RelExtensions
> 
> for instance.
> ...

Again, there's no need to centrally register a relation name, as it can 
be a URI.


BR, Julian

Received on Thursday, 3 July 2008 13:52:48 UTC