W3C home > Mailing lists > Public > public-xhtml2@w3.org > February 2009

Re: [Fwd: Using XMLNS in link/@rel]

From: Mark Nottingham <mnot@mnot.net>
Date: Fri, 27 Feb 2009 23:10:50 +1100
Cc: Steven Pemberton <steven.pemberton@cwi.nl>, www-tag@w3.org, Dan Brickley <danbri@danbri.org>, Ben Adida <ben@adida.net>, RDFa <public-rdf-in-xhtml-tf@w3.org>, XHTML WG <public-xhtml2@w3.org>
Message-Id: <B62D4A42-5E9E-43DA-ACD1-FDFB47DA5889@mnot.net>
To: Phil Archer <phil@philarcher.org>
No; I'm not asking for them to have the same syntax (and indeed the  
Link draft goes to some lengths to show that the syntax in the various  
formats isn't the same).

My only real concern here is that the syntax for a widely-used format  
-- HTML4 -- at least have consistent recommendations and indeed  
Recommendations (although of course there is likely to always be a  
difference between either sense of recommendation and actual practice).

RDFa doesn't specify the use of CURIEs in HTML4, it's very easy for  
someone not steeped in the arcana of HTML to believe that they can be  
used there, as we see with the CC document. If there's an expectation  
in the RDFa or XHTML community that they can be, or that it's OK  
anyway (wink, wink, nudge, nudge), that may indeed be the elephant in  
the room, and it needs to come out.

The other issue (RDFa's use of CURIEs in XHTML) needs to be clarified,  
but AFAICT I can live with however that clarification goes...


On 27/02/2009, at 10:52 PM, Phil Archer wrote:

> I understand the problem here. It would be nice to be able to use  
> the same @rel syntax in both HTTP Link:, HTML 4 and XHTML Link  
> elements. But you can't:  HTTP and HTML 4 have no way to define a  
> namespace. XHTML does, of course, and that's where the RDFa comes  
> in. So I wonder whether this is really 'all that bad?'
> It's an irritating fact of life that the ATOM registry and the RDFa  
> @rel values [1] are not the same but at least they don't actually  
> contradict each other. Where the same term exists in both, it means  
> the same thing (e.g.  next) and the synonyms (first cf. start) are,  
> well, just irritating, but that's life.
> So although sloppiness may lead some people to put  
> cc:morePermissions in an HTTP Link: header, it's clearly incorrect.  
> Whereas rel="http://creativecommons.org/ns#morePermissions" is fine  
> for everyone.
> Link: can't support Curies - but it doesn't have to. RDFa isn't  
> going to change, but it doesn't have to either. HTML doesn't support  
> namespaces and so shouldn't be using Curies anyway. If people do, it  
> may lead to unexpected results. C'est la guerre mes amis.
> Or am I missing the elephant in the room here? (not unknown I know...)
> Phil.
> [1] http://www.w3.org/TR/rdfa-syntax/#relValues
> [2] http://www.iana.org/assignments/link-relations/link- 
> relations.xhtml
> Mark Nottingham wrote:
>> Hello Steven,
>> On 27/02/2009, at 9:06 PM, Steven Pemberton wrote:
>>> On Fri, 27 Feb 2009 10:31:57 +0100, "Mark Nottingham"  
>>> <mnot@mnot.net> said:
>>>> Creative Commons just released a new spec:
>>>>  http://wiki.creativecommons.org/Ccplus
>>>> that has markup in this form:
>>>>  <a xmlns:cc="http://creativecommons.org/ns#"
>>>> rel="cc:morePermissions" href="#agreement">below</a>
>>>> (in HTML4, one assumes, since they don't specify XHTML, and this is
>>>> what the vast majority of users will presume).
>>> In the link you refer to they don't specify either, but I imagine  
>>> they mean XHTML,
>> I will wager any amount of money you care to name that more than  
>> 99% of the document's readers (as it stands) will assume HTML4.
>>> and I'm sure Ben Adida of CC can speak up here.
>>>> However, it appears that they adopted this practice from RDFa;
>>>>  http://www.w3.org/TR/rdfa-syntax/#relValues
>>>> which, in turn, *does* rely upon XHTML. However, XHTML does *not*
>>>> specify the @rel value as a QName (or CURIE, as RDFa assumes);
>>>> http://www.w3.org/TR/2008/REC-xhtml-modularization-20081008/abstraction.html#dt_LinkTypes
>>>> "Note that in a future version of this specification, the Working
>>>> Group expects to evolve this type from a simple name to a Qualified
>>>> Name (QName)."
>>>> So, that's an expectation, not a current specification.
>>> In fact it is a current specification. RDFa specifies a version of  
>>> XHTML that defines the meaning of CURIEs in rel and rev values.  
>>> Note that this is also not invalid HTML4 (which also allows such  
>>> values in a rel - they are CDATA - but doesn't specify what they  
>>> mean).
>>  http://www.w3.org/TR/rdfa-syntax/
>> refers to
>>  http://www.w3.org/TR/2001/REC-xhtml11-20010531/
>> which contains
>>  http://www.w3.org/TR/2001/REC-xhtml11-20010531/doctype.html
>> which refers to, for the Hypertext module (note 'latest version'  
>> URI):
>>  http://www.w3.org/TR/xhtml-modularization/abstract_modules.html#s_hypertextmodule 
>>  which leads back to
>>  http://www.w3.org/TR/xhtml-modularization/abstraction.html#dt_LinkTypes
>> i.e., the same, albeit most recent (instead of versioned) URI for
>>  http://www.w3.org/TR/2008/REC-xhtml-modularization-20081008/abstraction.html#dt_LinkTypes 
>>  Even taking the other road and going with the contemporary version,
>>  http://www.w3.org/TR/2001/REC-xhtml-modularization-20010410/abstraction.html#dt_LinkTypes 
>>  it's still just short names, with no reference to CURIEs or QNames  
>> at all.
>> What am I missing?
>> The only place I see this defined is in the RDFa syntax document  
>> itself -- do you mean that is the specification of authority? I  
>> note that it specifies /html/@version="XHTML+RDFa 1.0", and it has  
>> its own DTD, so in a way I suppose it's not really an extension to  
>> XHTML, but a re-definition of it...
>>>> Of course, this conflicts with the Link draft;
>>>> http://tools.ietf.org/id/draft-nottingham-http-link-header-04.txt
>>>> which we've worked pretty hard to come to consensus on across a  
>>>> broad
>>>> selection of communities (Atom, POWDER, OAuth, HTTP, and
>>>> optimistically, HTML5).
>>>> A few observations and questions;
>>>> 1) I'm more than happy to specify in the Link that in XHTML, a link
>>>> rel value is indeed a QName, if XHTML chooses to take that position
>>>> (although I believe a URI is a better fit than a QName here, as in
>>>> most other places). Can we get a current reading from the XHTML  
>>>> world
>>>> on this?
>>> A CURIE is a URI not a QName, so you're OK.
>> I haven't paid a lot of attention to them to date, but as far as I  
>> can see, a CURIE is most definitely not a URI; at most, it's a  
>> shorthand for one.
>>> CCing the XHTML2 WG and/or RDFa group would have helped in this  
>>> case if you wanted a response from them :-)
>> I wanted to get a feel from an architectural standpoint before  
>> talking to WGs about potentially irrelevant problems, but point  
>> taken.
>>>> 2) However, it seems like RDFa is jumping the gun by assuming  
>>>> @rel is
>>>> a CURIE right now.
>>> See above. It is already a Rec.
>>> [All the rest snipped since it was based on the assumption that  
>>> XHTML+RDFa isn't a Rec].
>> As I said before, the third point is IME the most concerning.  
>> Having two subtly incompatible syntax for the same attribute in  
>> HTML and XHTML isn't a great situation, but assuming that one is  
>> valid to use in the other is far more troublesome.
>> Cheers,
>> -- 
>> Mark Nottingham     http://www.mnot.net/
> -- 
> Phil Archer
> http://philarcher.org/
> i-sieve technologies                |      W3C Mobile Web Initiative
> Making Sense of the Buzz            |      www.w3.org/Mobile

Mark Nottingham     http://www.mnot.net/
Received on Friday, 27 February 2009 12:11:36 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 19:40:04 UTC