ontac-forum
[Top] [All Lists]

Re: [ontac-forum] Rule Interchange Format group

To: ONTAC-WG General Discussion <ontac-forum@xxxxxxxxxxxxxx>
From: richard murphy <rick@xxxxxxxxxxxxxx>
Date: Sat, 17 Dec 2005 12:06:38 -0500
Message-id: <43A4459E.2050205@xxxxxxxxxxxxxx>
Patrick & Chris:    (01)

CM> I am no expert on Topic Maps, but I don't believe they
 > have that sort of expressive power.  Their semantical framework (as seen
 > in Barta's Tau model, for example) is also quite different from that of
 > standard first-order logic (not bad, just different ;-), and its
 > connection to FOL is unclear -- I've seen assertions to the effect that
 > Topic Maps are similar in expressive power to description logics, which
 > means its expressiveness does not exceed that of OWL DL, and hence that
 > they are expressively quite weak, and certainly underpowered for a RIF.    (02)

In Description Logic for Use as the ODM Core:    (03)

http://www.sandsoft.com/edoc2004/HartEmeryDLCoreMDSW.pdf#search='a 
description logic for use as the odm core'    (04)

Table - 1 lists the expressive power of various languages based on 
description logic nomenclature. In this analysis, TopicMaps come in a 
distant last in expressiveness.    (05)

The classification as AL-- means TopicMaps only provide Atomic Concepts, 
Value Restrictions, and Datatypes.    (06)

OWL-DL is classified as SHOIN(D) which provides Atomic Concepts, 
Universal Concept, Bottom Concept, Atomic Negation, Intersection, Value 
Restrictions, Limited Existential Quantification, Full Negation or 
Complement, Full Existential Quantification, Role Hierarchies, Inverse 
Roles, Unqualified Number Restrictions, Enumerated Classes, Transitive 
Roles, Union Constructor, and Datatypes.    (07)

That being said, the stuff above presumes a DL rules engine and after 
browsing some of the papers here:    (08)

http://www.w3.org/2004/12/rules-ws/attendees    (09)

the conference participants haven't standardized on a rules engine.    (010)

Interestingly, ILog's paper has an appendix in which it's rules are 
modeled in UML which has ALHOIN(D) expressiveness which is less than 
SHIN(D), but quite a bit more than E/R which has ALN(D) expressiveness.    (011)

So, if we accept your assertion Chris, that TopicMaps are underpowered 
for a RIF, shouldn't we also ask what it means to be sufficiently 
powered and overpowered; and under what conditions an infomorphism 
respects local logics in a distributed system?    (012)

-- 
Best wishes,    (013)

Rick    (014)

email:  rick@xxxxxxxxxxxxxx
web:    http://www.rickmurphy.org
cell:   703-201-9129    (015)

Chris Menzel wrote:
> On Fri, Dec 16, 2005 at 09:30:34PM -0500, Patrick Durusau wrote:
> 
>>Snipping past most of your post to reach your comments on Rule 
>>Interchange Format group:
>>
>><snip>
>>
>>>You have provided no rigorous demonstration of the limitations, so
>>>what you are saying is idle chatter until you provide one.  And as
>>>I've already noted, W3C itself recognizes the expressive limitations
>>>of OWL.  This has led to the development of RuleML and SWRL, and is
>>>also a driver behind the newly formed Rule Interchange Format group
>>>(http://www.w3.org/2005/rules).
>>>
>>
>>What I find curious about the Rule Interchange Format activity at the 
>>W3C is that the posted use cases look like topic map use cases, for 
>>which there already exists an ISO standard, ISO 13250. While the charter 
>>is obviously a summary document, one would expect acknowledgement of 
>>prior work in an area with some explanation of why it was unsuitable 
>>before undertaking a duplication of prior work.
> 
> 
> Well, for one thing, this is a W3C effort and so the product is being
> designed to work well out of the box with OWL.  Granted, however that is
> more a matter of form than substance.  I think the major issue here is
> that the framework for the RIF will in the long run (maybe the short
> run) have to possess (or be extensible so as to possess) the power of
> full first-order logic (plus some set theory, no doubt) -- see Section 3
> of the charter.  I am no expert on Topic Maps, but I don't believe they
> have that sort of expressive power.  Their semantical framework (as seen
> in Barta's Tau model, for example) is also quite different from that of
> standard first-order logic (not bad, just different ;-), and its
> connection to FOL is unclear -- I've seen assertions to the effect that
> Topic Maps are similar in expressive power to description logics, which
> means its expressiveness does not exceed that of OWL DL, and hence that
> they are expressively quite weak, and certainly underpowered for a RIF.
> Again, that is in no wise a criticism.  Different tools are suitable for
> different jobs.  This one probably isn't for Topic Maps.  Point taken,
> though, that this deserves discussion.
> 
> 
>>My impression from reading the charter document is that the results of 
>>the activity have to be backwards compatible with RDF/OWL.
> 
> 
> Surely not in the sense that anything written in the rule language can
> be expressed in RDF/OWL.  One of the motivations is to provide a
> framework for having expressive power beyond OWL when you need it.
> 
> Chris Menzel
> 
>  
> _________________________________________________________________
> Message Archives: http://colab.cim3.net/forum/ontac-forum/
> To Post: mailto:ontac-forum@xxxxxxxxxxxxxx
> Subscribe/Unsubscribe/Config: 
>http://colab.cim3.net/mailman/listinfo/ontac-forum/
> Shared Files: http://colab.cim3.net/file/work/SICoP/ontac/
> Community Wiki: 
>http://colab.cim3.net/cgi-bin/wiki.pl?SICoP/OntologyTaxonomyCoordinatingWG
> 
> 
>     (016)



_________________________________________________________________
Message Archives: http://colab.cim3.net/forum/ontac-forum/
To Post: mailto:ontac-forum@xxxxxxxxxxxxxx
Subscribe/Unsubscribe/Config: 
http://colab.cim3.net/mailman/listinfo/ontac-forum/
Shared Files: http://colab.cim3.net/file/work/SICoP/ontac/
Community Wiki: 
http://colab.cim3.net/cgi-bin/wiki.pl?SICoP/OntologyTaxonomyCoordinatingWG    (017)
<Prev in Thread] Current Thread [Next in Thread>