W3C home > Mailing lists > Public > semantic-web@w3.org > December 2010

Re: Rule usage description - or how can I associate related rules to an ontology/RDF graph?

From: Bob Ferris <zazi@elbklang.net>
Date: Mon, 13 Dec 2010 11:15:36 +0100
Message-ID: <4D05F248.5050507@elbklang.net>
To: Semantic Web <semantic-web@w3.org>
Hi Ivan,

thanks a lot for your reply. As far as I understand the semantics of 
rif:usedWithProfile* (or rif:imports), it defines a strict application 
of the rules that are included in the associated profile. I'm especially 
interested in varieties of such a relation e.g.,

- "recommendation": it depends upon the reasoning engine settings, 
whether the related rules would be applied or not
- "strict": please apply these rules to achieve the intended meaning
- "applied": the reasoning engine of the information provider applies 
these rules in reasoning tasks

Or can I describe such a rule usage description in a RIF profile 
(machine-processable)? If so, please could you (or someone) provide an 
example.

Cheers,


Bob


*) is similar to spin:rule

Am 12.12.2010 12:09, schrieb Ivan Herman:
> Bob,
>
> this issue did come up in the course of the SPARQL 1.1 work, which contains a separate entailment document[1]. That entailment document also includes RIF entailment, which, in turn, requires the answer to the question you are asking. And (good timing!) Axel Polleres and Sandro Hawke just described a mechanism whose goal is to be included the SPARQL standard[2]. This is by no means final, will be discussed by the SPARQL Group I presume, but will be finalized eventually...
>
> I hope this helps
>
> Ivan
>
>
> [1] http://www.w3.org/TR/2010/WD-sparql11-entailment-20101014/
> [2] http://www.w3.org/2009/sparql/wiki/Importing_RIF
>
> On Dec 9, 2010, at 14:36 , Bob Ferris wrote:
>
>> Hi,
>>
>> someone might say now: "Hey, I heard that question already somewhere else." - Yes, you are right. I asked this question at semanticoverflow.com[1]. However, I'm sure that I'll hopefully reach here a broader/different audience. Furthermore, I think more and more that this is and will be a very important issue (requirement) for the 'main' purpose of the Semantic Web - information integration.
>>
>> So here we go:
>>
>> As I think the need for assigning specific rules to Semantic Web ontologies/RDF graphs to enable intented inferences is getting more and more important, we need possibilities to semantically related these rules to Semantic Web ontologies/RDF graphs. The publication of the Rule Interchange Format (RIF)[2] this summer was a (huge) step into that direction. However, am I right that they missed a (from my point of view) very important functionality? To quote a part of an answer of an RIF FAQ[3]:
>>
>> "This lets you physically embed RIF in an RDFS/OWL document, but notes that the embedded RIF is merely described, not asserted. There is not currently a standard vocabulary saying, in RDFS/OWL, that you also want some RIF rules as part of your ontology. Instead, for now, you must have RIF import RDFS/OWL."
>>
>> So how can I associate rules that should be/could be applied to a specific Semantic Web ontology/RDF graph? - e.g.,
>>
>>     * Information Service A applied rule B,C,D to it whole knowledge base that can be identified by URI Z
>>     * Information Service E suggest rule F and G to be applied at RDF Graph Y
>>     * Ontology H should be used with rule I for proper reasoning
>>
>>
>> Already proposed ideas:
>>
>> 1. the SPIN framework[4]
>> 2. the Rulz vocabulary[5]
>>
>> Where the first one offers spin:rule and spin:constraint to associate rules/constraints to RDF/OWL models, the second offers a quite simple mechanism to embed rules, that are described in a certain rule language, in an RDF graph.
>> However, I'm looking for rule usage description*, i.e. I do not simple want to associate a rule by using a quite static property e.g., spin:rule that has quite interpretable semantics. I want relations to 'suggest' or 'prescribe' rules. Maybe also by explaining their benefits etc. Another attribute would be 'applied', so that I can express that the information service where the information comes from uses this rule (/these rules) in its reasoning engine. I guess there might be more use cases.
>> I think that this mechanism is really necessary, if we want to share proper semantics to interpret the sense of an information. I believe that we cannot achieve a quite good interpretation (intended meaning/purpose) of a message, when we use a 'simple' description of an applied concept (here a description without relations to related rules).
>>
>> What do you think about this issue? I think it is crucial.
>>
>> Cheers,
>>
>>
>> Bob
>>
>>
>> *) afaik RIF includes also some attributes to describe rules/ usage of rules. However, all descriptions I've seen so far are natural language text, which is quite bad to interpret at the moment
>>
>>
>> [1] http://www.semanticoverflow.com/questions/2293/how-can-i-associate-related-rules-to-an-ontology-rdf-graph
>> [2] http://www.w3.org/TR/rif-overview/
>> [3] http://www.w3.org/2005/rules/wiki/RIF_FAQ#How_do_I_embed_RIF_in_an_RDFS.2FOWL_schema_or_ontology.3F
>> [4] http://spinrdf.org/
>> [5] http://vocab.deri.ie/rulz#
Received on Monday, 13 December 2010 10:16:06 GMT

This archive was generated by hypermail 2.3.1 : Tuesday, 26 March 2013 21:45:40 GMT