RE: Link: relation registry and 303

Hello Mark,

[catching up...]

> -----Original Message-----
> From: Mark Nottingham [mailto:mnot@mnot.net] 
> Sent: 30 January 2009 22:26
> To: Williams, Stuart (HP Labs, Bristol)
> Cc: Tim Berners-Lee; Jonathan Rees; www-tag@w3.org WG; Lisa Dusseault
> Subject: Re: Link: relation registry and 303
> 
> 
> On 31/01/2009, at 3:59 AM, Williams, Stuart (HP Labs, Bristol) wrote:
> > So... lets just leave the registered relation names as absolute  
                                                           ^^relative
> > URIRefs, and lets please continue to take the view that the things  
> > those registered and "non-registered" full URIRefs denote are link  
> > relations rather than there describing documents - and lets just  
> > concede (if we have to) that as a source of information about the  
> > link-relation that description document is more to be believed than  
> > what might be inferred from a 200 or 303 response code.
> >
> > But... please lets not jump through some twisted hoops that have  
> > some of the link rel names being URI (albeit relative URI) and  
> > others not... and please, if we mean the URI to denote link  
> > relations lets say that that's what they denote when we describe  
> > them - in the long run I think that would make life much simpler for  
> > someone 5-10years down the road wondering what planet we were on.
> 
> 
> Perhaps you misunderstand.

Always a possiblity :-)

> I'm not proposing that registered vs. non-registered be split to 
> solve this problem; rather, it was something  
> already in the works, primarily based upon feedback from the HTML5  
> community.

Well... from what you're saying it seems as though some sort of "split" is being considered that disconnects he registered shortnames from URI space.

> They were concerned that, historically, link relations have been  
> compared in a case-insensitive fashion, which makes working with URIs  
> much more complex. Bifurcating it neatly solves this problem.

Well... maybe we just have to allow 2^n aliases for a shortname relation of n-characters to allow case insensitive comparison. Not ideal, but the pragmatics may require it.

> --
> Mark Nottingham     http://www.mnot.net/

Stuart
--
Hewlett-Packard Limited registered Office: Cain Road, Bracknell, Berks RG12 1HN
Registered No: 690597 England

Received on Monday, 2 February 2009 17:17:14 UTC