Re: [RDFa] ISSUE-8: RDF containers in RDFa

Hausenblas, Michael wrote:
> 
> Ben,
> 
> Did you take into account 'RDF features covered by RDFa (RiR)' [1], yet?
> I know, the description on this page is a bit terse and
> might need some more explanations ...
> 
> The main finding of [1] is:
> 
> '... the only requirements - stemming directly from 
>  the RDF Graph Model - are to represent:
> 
>   + URIrefs
>   + bNodes
>   + literals (incl. XMLLiteral) in the triple form S P O
> 
>  all other constructs (as rdf:Seq, or rdf:parseType) 
>  are introduced by RDF/XML .'
> 
> Unfortunately I did not get any feedback if my interpretation is
> correct,

Well, it is not _entirely_ correct

- The RDFS document[1] introduces the classes and properties used to
construct containers and collections (not rdf:li, though, that _is_ an
RDF/XML artefact). The definitions in [1] are _not_ dependent on RDF/XML
or any other serialization.

- The RDF Semantics[2] documents does not defines more than a minimal
semantics to the container terms (Seq, rdf:_1, etc), defining, eg, the
latter to be rdf:Property type. The domain and range of rdf:_1 and the
others are simply defined to be rdf:Resource, ie, anything. None of the
'usual' semantics associated to Seq, Alt, etc, are formally defined.

(Many would like to get those out of RDF altogether because they are ill
defined, but there is some history there that we cannot ignore. Note
also that rdf:_i brings in a very disagreeable feature to the formal
semantics, because the number of terms of the RDF vocabulary becomes
infinite...)

- There _is_ a semantics attached to the collection terms in a separate
'Collection' vocabulary[3]. Eg, the range of rdf:rest is defined to be
rdf:List, that sort of things. Ie, the documents it produces a more or
less controlled and controllable way of constructing, well, collections.

- RDF/XML introduces syntactic sugar for both collections and containers.

- Turtle introduces syntactic sugar for collections: (a b c d e). It
does not introduce anything similar for containers.


[1] http://www.w3.org/TR/rdf-schema/
[2] http://www.w3.org/TR/rdf-mt/#Containers
[3] http://www.w3.org/TR/rdf-mt/#collections


> but maybe we can discuss this in the context of ISSUE-8. 
> 
> My main point here is: vocabulary as rdf:li, or rdf:Seq stem from
> RDF/XML,
> hence we should contemplate on *IF* and *HOW* to support it.
> 

I agree that this background is highly relevant.

Ivan


> Cheers,
>  Michael
> 
> BTW: I get more and more the feeling we focus too much on syntactic
> sugar 
>      while not resolving essential issues as bNodes, etc.
> 
> [1] http://www.w3.org/2006/07/SWD/wiki/RDFa/RDFCoverage
> 
> ----------------------------------------------------------
>  Michael Hausenblas, MSc.
>  Institute of Information Systems & Information Management
>  JOANNEUM RESEARCH Forschungsgesellschaft mbH
>   
>  http://www.joanneum.at/iis/
> ----------------------------------------------------------
>  
> 
>> -----Original Message-----
>> From: public-swd-wg-request@w3.org 
>> [mailto:public-swd-wg-request@w3.org] On Behalf Of Ben Adida
>> Sent: Thursday, July 12, 2007 1:10 AM
>> To: RDFa; SWD WG
>> Subject: [RDFa] ISSUE-8: RDF containers in RDFa
>>
>>
>>
>> Hi all,
>>
>> There's one syntax issue we haven't talked about n a while that we
>> should probably address in *some* form: how to express RDF containers
>> and/or lists. In particular, how do we indicate a list of 
>> creators, or a
>> list of licenses, or a list of authors in a paper.
>>
>> I wrote a proposal on RDFa containers a while ago:
>> http://www.w3.org/2001/sw/BestPractices/HTML/2006-rdfa-containers
>>
>> but that assumed we would do RDF Bags and Sequences. I don't 
>> know yet if
>> that's what we should support, or if we should just focus on lists,
>> denoted [a,b,c] in Turtle.
>>
>> Input from anyone, in particular the SWD WG, would be very helpful! I
>> think for XHTML1.1+RDFa, we should do the simplest thing that doesn't
>> completely prevent us from encoding lists of some kind.
>>
>> In terms of implementation, the right direction is likely something
>> using UL, OL, and LI, possibly involving @href/@resource on the LI, as
>> we had expected @href everywhere in the original XHTML2-based proposal.
>>
>> -Ben
>>
>>
> 

-- 

Ivan Herman, W3C Semantic Web Activity Lead
URL: http://www.w3.org/People/Ivan/
PGP Key: http://www.cwi.nl/%7Eivan/AboutMe/pgpkey.html
FOAF: http://www.ivan-herman.net/foaf.rdf