W3C home > Mailing lists > Public > public-rdf-wg@w3.org > March 2012

Re: three kinds of dataset

From: Pat Hayes <phayes@ihmc.us>
Date: Mon, 12 Mar 2012 16:06:18 -0500
Cc: RDF-WG WG <public-rdf-wg@w3.org>
Message-Id: <8C38448B-0055-437F-8E70-0C3D3E1DC879@ihmc.us>
To: Antoine Zimmermann <antoine.zimmermann@emse.fr>, "Peter F. Patel-Schneider" <pfpschneider@gmail.com>

On Mar 7, 2012, at 6:55 AM, Antoine Zimmermann wrote:

> Peter,
> 
> 
> Infinity does not cause problems at all, on the contrary, that's why I'd like to allow dataset-interpretations to assign a "local" interpretation to an infinity of "graph names". Moreover, I'd like literals to be allowed as "graph names".
> 
> The reason is the following. Imagine I want to extend this weak semantics to a temporal semantics.
> 
> I'd like to say things like:
> 
> "[1927,1945]"^^interval { :x  a  :Child .}
> "[1999,+\infinity]"^^interval { :x  a  :DeadPerson .}
> 
> and I want to infer that at any given year <y> >= 1999
> 
> "[<y>,<y>]"^^interval { :x  a  :DeadPerson .}
> 
> which requires that there exists an RDF/RDFS/etc-interpretation Con("[<y>,<y>]"^^interval) for any <y>.

This gets more and more like a modal logic with every email :-)  This is Kripke semantics for "1999"^^date G{ :x a :DeadPerson .} where G is not-F-not, ie the strong future tense "from now on, ...", probably in S4.

I have to admit, the simplicity and one-size-fits-all of your semantics is appealing. But I really am uncomfortable (and I think many others will be) at the idea of making all URIs contextual by default. There has to be a middle way here somewhere.

Pat

PS. Peter, remember that the modal languages are almost all robustly decidable, and indeed that the decideable guarded fragment was originally adapted from modal logic. So even if there are infinite models, I think this will turn out to not be a dealbreaker for the DL world. Just more employment opportunities for logicians :-). 

> 
> 
> AZ
> 
> 
> Le 06/03/2012 19:31, Peter F. Patel-Schneider a écrit :
>> Well, I don't see where the infinity causes problems.  RDF datasets and
>> RDF graphs are going to be finite, after all, aren't they?  Perhaps for
>> mathematical cleanliness you might want to allow interpretations to have
>> an infinite number of names, I suppose.
>> 
>> The wiki seems to be missing the notion that Con(V) is an interpretation
>> of the graph with name V.
>> 
>> In any case, the two formulations appear to be quite close (modulo the
>> issue just above).
>> 
>> peter
>> 
>> 
>> 
>> From: Antoine Zimmermann<antoine.zimmermann@emse.fr>
>> Subject: Re: three kinds of dataset
>> Date: Tue, 06 Mar 2012 19:10:36 +0100
>> 
>>> Why not indeed?  This was my original idea, modulo a little improper
>>> formulation (you can see the previous version in the wiki) but 1) Pat
>>> was very much against this formulation, and 2) the current formulation
>>> allows a dataset-interpretation to assign a "local" interpretation to a
>>> potentially infinite set of terms. This is particularly useful for
>>> reasoning with annotated triples (UC 6.2 in
>>> http://www.w3.org/2011/rdf-wg/wiki/TF-Graphs-UC#.28B_priority.29_Reasoning_over_annotations).
>>> 
>>> Le 06/03/2012 18:52, Peter F. Patel-Schneider a écrit :
>>>> Why not just say that an RDF/RDFS/... dataset interpretation of
>>>> 	D = (G, {<n1,G1>, ...,<nk,Gk>})
>>>> is a structure
>>>> 	I = (I, {<m1,I1>, ...,<mh,Ih>})
>>>> where I is an RDF/RDFS/... interpretation of G
>>>> and for each 1<=i<=k there is a j, 1<=j<=h such that mj=ni
>>>> and Ij is an RDF/RDFS/... interpretation of Gi
>>>> (could also require ni distinct and h=k)
>>>> 
>>>> Then	D = (G, {<n1,G1>, ...,<nk,Gk>})
>>>> entails D' = (G', {<n'1,G'1>, ...,<n'k',G'k'>})
>>>> just when every RDF/RDFS/... datatset interpretation of D
>>>> is also an RDF/RDFS/... datatset interpretation of G'
>>>> 
>>>> peter
> 
> -- 
> Antoine Zimmermann
> ISCOD / LSTI - Institut Henri Fayol
> École Nationale Supérieure des Mines de Saint-Étienne
> 158 cours Fauriel
> 42023 Saint-Étienne Cedex 2
> France
> Tél:+33(0)4 77 42 83 36
> Fax:+33(0)4 77 42 66 66
> http://zimmer.aprilfoolsreview.com/
> 
> 

------------------------------------------------------------
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 Monday, 12 March 2012 21:06:55 GMT

This archive was generated by hypermail 2.3.1 : Tuesday, 26 March 2013 16:25:47 GMT