Re: longdesc URLs and RDFa

Leif,

just thinking out loud here, this is not an 'official' answer of the WG in any way. My colleagues may have very different reactions...

It is of course _possible_ to extend the (X)HTML+RDFa attribute set by interpreting @longdesc the way you suggest. Ie,

<img src="*" longdesc="long.html" alt="short description" />

could be considered as a short hand for

<img src="i.src" rel="longdesc" resource="long.html" property="alt" content="short description" />

Generating triples for the i.src URI. As see, I extended that to the @alt attribute, too; I am not sure how I would justify doing that for longdesc and not doing it for @alt.

The only thing I am a little bit concerned, I must admit, is not the technical issue but the approach taken, ie, that we push RDFa into interpreting new attributes for a very specific purposes. If we go down that line, I am afraid of opening up flood gates (well, very small ones, but nevertheless), which would bloat the RDFa spec. Eg, we do not interpret the various attributes of the <object> element, I am not sure what the new <video> and <audio> elements would bring, etc.

Note that registering the longdesc (and alt) as 'terms' as we call them in RDFa, ie, as specific and predefined property/rel types, is much easier to do and we are in the process of defining the mechanism and the content to do that for HTML or XHTML. Ie, that can be done easily, and no matter what. But what this means is that the users would have to manually write the <img> elements the way I wrote them up there. 

Cheers

Ivan


On Aug 14, 2010, at 04:58 , Leif Halvard Silli wrote:

> There is a debate in the HTMLwg about the longdesc attribute. And in 
> that debate, the thought has surfaced that @longdesc,
> 
> 	<img src="*" longdesc="long.html" alt="short description" />
> 
> is simply a *shorthand notation* for this (hypothetical) micro format
> 
> 	<a href="long.html" rel="longdesc">	
> 		<img src="*" alt="short description" />
> 	</a>
> 
> This has resulted in an attempt to register the 'longdesc' as link 
> relation. [1] Simultaneously, a search for a better documentation of 
> implementations of @longdesc has been (re)started. And in that context, 
> I raised a question about RDFa and @longdesc. [2]
> 
> 	Observation 1: The XHTML+RDFa DOCTYPE supports the @longdesc 
> attribute. 
> 	Observation 2: Mark Birbeck states in his 'Introduction to RDFa' 
> article that «The @rel and @href attributes are no longer confined to 
> the a and link elements, but can also be used on img to indicate a 
> relationship between the image and some other item.» [3]
> 	Observation 3: A longdesc link indicates a relation between a short 
> description and a long description. (See the quotes from HTML4 provided 
> in [1]: 'link to long description (complements alt)' and 'link to long 
> description (complements title)'.)
> 
> Questions: Would it not be logical if RDFa treated the @longdesc link 
> as a semantic link - e.g. equivalent to the micro format I described 
> above? What needs to be added in order to make this happen?
> 
> [1] 
> http://www.ietf.org/mail-archive/web/link-relations/current/msg00047.html
> [2] http://lists.w3.org/Archives/Public/public-html/2010Aug/0140
> [3] http://www.alistapart.com/articles/introduction-to-rdfa/
> -- 
> leif halvard silli
> 


----
Ivan Herman, W3C Semantic Web Activity Lead
Home: http://www.w3.org/People/Ivan/
mobile: +31-641044153
PGP Key: http://www.ivan-herman.net/pgpkey.html
FOAF: http://www.ivan-herman.net/foaf.rdf