Re: What is it that's wrong with rdf:List [summary]

Another quick comment (In case you havnt had enough from Jeremy....)

On Jun 21, 2010, at 5:31 AM, Michael Schneider wrote:

> Norman Gray wrote:
>
>> So is rdf:List any different from the other parts of RDF(S) excluded
>> from OWL-DL/Lite <http://www.w3.org/TR/owl-semantics/ 
>> mapping.html#4.2>?
>> I can see that other elements of that excluded list are problematic  
>> for
>> reasoning because in RDF(S) they have domains or members which are
>> largely unrestricted, which could clearly complicate reasoning about
>> them.  It sounds (to me, possibly naively) that rdf:List is  
>> included in
>> that list for rather more technical reasons, and that this may  
>> account
>> for the slight air of aggrievedness (is that too strong?) which hangs
>> around it.
>
> Here is my estimation:
>
> AFAIR, the OWL Working Group *postponed* the issue to allow RDF  
> lists in OWL
> 2 DL (see [1]), since it did not know how to handle RDF lists in the  
> reverse
> OWL 2 Mapping to RDF Graphs [2].
>
> I guess, if someone would take that job on himself to analyze the RDF
> mapping and finds out how to extend it to (more or less) general  
> list usage
> without breaking the mapping and without making it considerably more
> complicated, then there would be a chance to get RDF lists into OWL  
> DL in a
> future version. At least, tool builders would then have a solid  
> foundation
> to extend their OWL DL tools to cope with RDF lists in OWL DL  
> ontologies in
> RDF graph form. And, having several tools with mature RDF list  
> support would
> be an even better reason for a future OWL working group to  
> reconsider the
> decision being made for OWL 2 DL.
>
> In any case, RDF lists would have no impact on OWL DL reasoning.  
> They do not
> have any semantic meaning in OWL DL, and they can be fully  
> reconstructed,
> anyway, only by using a different namespace for the different URIs  
> (e.g.
> "ex:List", "ex:first", "ex:rest", "ex:nil").

Well, yes, of course they can. It is in a sense trivial to simply  
transcribe the rdf:List vocabulary into another namespace and just use  
that instead. And as you say, RDF assigns no formal semantics to this  
vocabulary (just a kind of vague recommendation to use it properly,  
and then interpret it as denoting the list structures it seems to  
describe), so using your own home-rolled version is just as  
semantically good. But this ignores the social utility of there being  
a *single* list vocabulary that every one uses. Without this, lists  
are useless. There is no point in my using my list vocabulary to make  
well-formed lists if you can['t query my RDF without knowing my list  
vocabulary. There needs to be a *single*, *normative* list vocabulary  
in order for RDF to function as an interchange language. And it is the  
using-up of the obvious one by OWL which is the issue with which we  
began. Its not really a technical problem at all: the technical  
problem is completely trivial. Its a political/social/standardization  
problem which can only be solved by the W3C.

I disagree with jeremy about ho hard this is to solve. If we had  
collectively had the foresight to see this problem coming, the obvious  
thing to have done would to have introduced *two* LIst-style sets of  
primitives into RDF, one for general use and the other reserved for  
encoding OWL syntax. And this is still a viable alternative, given  
that RDF is being re-thought. Its a pity that the obvious, intuitive  
names have been used up for syntax encoding, so that the 'real' list  
vocabulary for general use will have to be different, but that is a  
minor matter. The world will rapidly get used to calling them  
'sequences' or some such name. (I hereby vote for rdf:Sexpession,  
which is historically accurate and is likely to give rise to an  
entirely new generation of amusing misunderstandings; and rdf:car and  
rdf:cdr also have a delightfully retro flavor, combining Web 3.0 with  
the architecture of the IBM 704.)

Pat

>
> Cheers,
> Michael
>
> [1] <http://www.w3.org/2007/OWL/tracker/issues/129>
> [2] <http://www.w3.org/TR/2009/REC-owl2-mapping-to-rdf-20091027/>
>
> --
> Dipl.-Inform. Michael Schneider
> Research Scientist, Information Process Engineering (IPE)
> Tel  : +49-721-9654-726
> Fax  : +49-721-9654-727
> Email: michael.schneider@fzi.de
> WWW  : http://www.fzi.de/michael.schneider
> = 
> ======================================================================
> FZI Forschungszentrum Informatik an der Universität Karlsruhe
> Haid-und-Neu-Str. 10-14, D-76131 Karlsruhe
> Tel.: +49-721-9654-0, Fax: +49-721-9654-959
> Stiftung des bürgerlichen Rechts, Az 14-0563.1, RP Karlsruhe
> Vorstand: Prof. Dr.-Ing. Rüdiger Dillmann, Dipl. Wi.-Ing. Michael  
> Flor,
> Prof. Dr. Dr. h.c. Wolffried Stucky, Prof. Dr. Rudi Studer
> Vorsitzender des Kuratoriums: Ministerialdirigent Günther Leßnerkraus
> = 
> ======================================================================
>
>
>

------------------------------------------------------------
IHMC                                     (850)434 8903 or (650)494 3973
40 South Alcaniz St.           (850)202 4416   office
Pensacola                            (850)202 4440   fax
FL 32502                              (850)291 0667   mobile
phayesAT-SIGNihmc.us       http://www.ihmc.us/users/phayes

Received on Tuesday, 22 June 2010 01:01:09 UTC