W3C home > Mailing lists > Public > public-owl-wg@w3.org > August 2009

Re: problematic literal-related rules of OWL 2 RL

From: Ian Horrocks <ian.horrocks@comlab.ox.ac.uk>
Date: Mon, 3 Aug 2009 16:14:00 +0100
Message-Id: <AA5D3FE5-C6DB-4B6B-AEFD-0BE81CD961DF@comlab.ox.ac.uk>
Cc: Michael Schneider <schneid@fzi.de>, public-owl-wg@w3.org, Zhe Wu <alan.wu@oracle.com>
To: Ivan Herman <ivan@w3.org>
The rules presented in Profiles-4.3 constitute "a partial  
axiomatization of the OWL 2 RDF-Based Semantics" and are "a useful  
starting point for practical implementation"; they are NOT themselves  
a set of rules that can be directly used in an implementation -- this  
was never the intention. Moreover, even with the suggested change the  
rules would be unsuitable for (non-naive) implementation as they  
could/would lead to the generation of an unfeasibly large set of  
conclusions in realistic applications.

Therefore, I think that it is better to leave these rules as they are  
-- they clearly state what implementations need to achieve, while at  
the same time making it clear that implementers need to do something  
more than just pouring this set of rules into a rule engine.

Ian



On 1 Aug 2009, at 08:11, Ivan Herman wrote:

> FWIW, my RL implementation can implement those rules only for those
> literals that appear in the graph. Also, the analysis made by Dave
> Reynolds on the implementability of RL with RIF, looking at 4.4.2
>
> http://www.w3.org/2005/rules/wiki/OWLRL#Datatype_rules
>
> led to the same conclusions, as far as I can understand.
>
> I will on vacations starting, well, now:-), ie, will not be at the
> telco. For the records, I am very much in favour of Michael's proposal
>
> Ivan
>
> Michael Schneider wrote:
>> Hi!
>>
>> I know this comes very late in the process, but I want to state it  
>> anyway, after having had a private discussion on this topic with  
>> Ivan and Zhe this month.
>>
>> While creating the datatype related testcases of the RL-specific  
>> fraction of my testsuite (/not/ in the WG-approved subset of  
>> testcases) and playing around with implementing RL reasoning, I  
>> started to consider several of the rules in Table 8 of the RL  
>> specification (Section 4.3 of Profiles) to be problematic, e.g. rule:
>>
>>   <http://www.w3.org/2007/OWL/wiki/Profiles#dt-diff>
>>
>> The RL rules generally do not introduce new nodes such as bNodes  
>> in the consequent, when it isn't already in the premise. This  
>> "principle" contributes to the simplicity of the language. But the  
>> rules "dt-type2", "dt-eq" and "dt-diff" in Table 8 are different.  
>> For example, rule "dt-diff" states that
>>
>>   FORALL literals lt1 and lt2 with different data value:
>>       IF
>>           /* nothing */
>>       THEN
>>           T(lt1, owl:differentFrom, lt2)
>>
>> This leads to infinitely many triples even when starting from the  
>> empty RDF graph. For example, via this rule the following  
>> entailment holds for RL:
>>
>>   (*) {} |= { "42"^^xsd:integer owl:differentFrom  
>> "21.0"^^xsd:decimal . }
>>
>> This is very different from, for example, the reflexivity rule for  
>> owl:sameAs, which only "fires" for nodes that exist in the  
>> original graph. I think that these rules in Table 8 significantly  
>> add to the complexity of the OWL 2 RL/RDF ruleset, and in this  
>> form the usefulness as a "starting point for practical  
>> implementation using rule-based technologies" is IMHO much reduced.
>>
>> My suggestion would be to weaken the rules "dt-type2", "dt-eq" and  
>> "dt-diff" in a way that literals on the right hand side must occur  
>> on the left hand side of the rules. So, (*) would no longer follow  
>> from the RL rules, but (**) would still be a valid conclusion:
>>
>>   (**) {
>>          ex:s1 ex:p1 "42"^^xsd:integer .
>>          ex:s2 ex:p2 "21.0"^^xsd:decimal .
>>        }
>>        |=
>>        { "42"^^xsd:integer owl:differentFrom "42.0"^^xsd:decimal . }
>>
>> But this wouldn't be a big problem anymore, since there would then  
>> only be a finite set of conclusions from the premise graph.
>>
>> AFAICT, this change would also need to be propagated to theorem  
>> PR1. Take for example the following entailment query:
>>
>>   (***) { ex:s ex:p "42"^^xsd:integer . }
>>         |=
>>         { ex:s ex:p "42.0"^^xsd:decimal . }
>>
>> This actually follows from the current OWL 2 RL/RDF rules because  
>> of rule "dt-eq", via an "intermediate" owl:sameAs triple between  
>> "42"^^xsd:integer and "42.0"^^xsd:decimal (and via the  
>> substitution rule for owl:sameAs). It also follows from the Direct  
>> Semantics and, AFAICT, (***) satisfies the conditions of PR1. But  
>> (***) would not follow by the weakened version of rule "dt-eq",  
>> since the literal "42.0"^^xsd:decimal is not mentioned on the left  
>> hand side.
>>
>> What's your opinion on this last minute change request?
>>
>> Regards,
>> Michael
>>
>> --
>> Dipl.-Inform. Michael Schneider
>> Research Scientist, Dept. 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
>> ===================================================================== 
>> ==
>>
>
> -- 
>
> 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
>
Received on Monday, 3 August 2009 15:14:44 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 3 August 2009 15:14:45 GMT