W3C home > Mailing lists > Public > www-tag@w3.org > September 2008

Re: rel=CURIE in RDFa, but rel=URI in Link:

From: Julian Reschke <julian.reschke@gmx.de>
Date: Fri, 26 Sep 2008 10:16:39 +0200
Message-ID: <48DC9A67.2000002@gmx.de>
To: Shane McCarron <shane@aptest.com>
CC: Dan Connolly <connolly@w3.org>, "Henry S. Thompson" <ht@inf.ed.ac.uk>, Jonathan Rees <jar@creativecommons.org>, "www-tag@w3.org WG" <www-tag@w3.org>


I noticed that you didn't reply to this email. We discussed this topic 
shortly on yesterday's HTML5 telco.

 From what I hear it seems your position is that either:

1) It's ok for XHTML2 and HTML5 to define conflicting syntax rules for 
@rel (and potentially @rev).

2) It *is* a problem, so HTML5 will need to adapt to what RDFa defines.

Re 1: that would be a problem; it will make conversion of documents and 
document fragments much harder, and will create an inconsistency in the DOM.

Re 2: I doubt that this is going to fly, but if you think this would be 
the right thing, please raise this point over in the HTML WG.

BR, Julian

PS: please note that I'm generally in *favor* of the RDFa approach, and 
I would like it to be fully adopted in HTML5 (and extensions to HTML4, 
btw). I just see a problem with introducing plain (non-safe) CURIEs in a 
place where others already want to use URIs.

Julian Reschke wrote:
> Shane McCarron wrote:
>> FWIW, RDFa is part of the XHTML 2 activity, and DOES own link/@rel.  
>> We believe that the extension of @rel to use CURIE is completely 
>> consistent with the HTTP spec HTTP Link: space. The value space for 
>> CURIE is IRI.  
> I think it's safe to say that there isn't consensus about "who owns the 
> rel attribute" between the XHTML2 and HTML5 working groups. It would be 
> unfortunate if we ended up with different syntax in both languages.
> Furthermore note that the lexical space in the HTTP link header draft is 
> *URI*, not *IRI*. One could argue that this is a problem HTTP needs to 
> solve, but it's worth keeping in mind nevertheless.
>> The lexical space doesn't really matter in this context - since any 
>> processor looking at link / @rel would need the value space version.  
>> What am I missing here?
> Best regards, Julian
Received on Friday, 26 September 2008 08:17:25 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 22:56:24 UTC