RE: question: datatype reasoning?

At 21:50 +0000 1/24/03, Ian Horrocks wrote:
>As Jeremy points out, reasoning with datatypes can be
>hard/problematical. Fortunately, OWL Lite and DL have a cunning design
>such that reasoning with datatypes is separated from "abstract"
>reasoning. This means that:
>
>1. an OWL Lite/DL reasoner is complete iff both the abstract part and
>datatype part are independently sound and complete.
>
>2. we can easily formalise what constitutes an "admissible" datatype,
>i.e., one for which it is theoretically possible to provide a sound
>and complete reasoner (see [1]).
>
>3. a limited (in terms of datatypes supported) and/or incomplete
>datatype reasoner would introduce a limited and easily characterised
>form of overall incompleteness.
>
>I think that we can expect many implementations to support only a
>subset of the available datatypes, but given the above design it will
>be easy to inform users as to just what such reasoners can do and
>where they will be incomplete.
>
>Ian

Ian, that's reassuring (and thanks also to Peter C. for his earlier 
mail on the topic) -- I think that the WG needs to reach a decision, 
and particularly to make sure this gets into some document - JH

>
>
>[1] 
>http://www.cs.man.ac.uk/~panz/Zhilin/download/Paper/Pan-Horrocks-datatype-2002.pdf
>
>On January 24, Jeremy Carroll writes:
>>
>>
>>  My understanding is as follows:
>>
>>  There are many difficult cases
>>  e.g.
>>
>>  <p> range xsd:negativeInteger
>>  <p> range xsd:byte
>>
>>  DL-entails
>>
>>  owl:Thing owl:sameClassAs
>>    restriction(p, maxCardinality=128 )
>>
>>  There are worse cases when we are considering potential compatible types
>>  that do not have the same primitive base type (e.g. try xsd:float instead of
>>  xsd:byte). Such cases are in need of clarification from XML Schema WG.
>>  Nasty questions that can be asked in OWL DL assuming floats and decimals
>>  share the same underlying real values is how many floats are unsignedLongs
>>  (algorithm is keep increasing the maxCardinality in an entailment like that
>>  above until the entailment does not hold).
>>
>>
>>  I am not sure that there can be an effective OWL DL reasoner that cannot
>>  reason about the cardinalities of sets of individuals, and hence it may also
>>  need the ability to reason about the cardinalities of sets of datatypes (as
>>  in the above example).
>>  OWL Lite is largely rescued from this by restriciting cardinalities to 0 and
>>  1. A complete OWL Lite reasoner would I suspect need to simply know the
>>  intersection lattice of the XML Schema builtin datatypes and which
>>  intersections are empty and singleton.
>>
>>  e.g.
>>
>>  <p> range xsd:nonNegativeInteger
>>  <p> range xsd:nonPositiveInteger
>>
>>  Lite-entails
>>
>>  owl:Thing rdfs:subClassOf
>>    restriction(p, maxCardinality=1 )
>>
>>  or
>>
>>  <p> range xsd:unsignedByte
>>  <p> range xsd:nonPositiveInteger
>>  <i> rdf:type restriction( p, minCardinality=1 )
>>
>>  Lite-entails
>>
>>  <i> <p> "0"^xsd:int
>>
>>  OWL with user defined datatypes (that we want) is worse.
>>  OWL Full (with or without user defined datatypes) is much worse.
>>
>>  I was planning on adding such tests under the datatypes issue.
>>
>>  see:
>>  http://lists.w3.org/Archives/Public/www-archive/2002Nov/att-0092/02-index
>>  for long discussion of these issues (RDF not OWL)
>>
>>
>>  Jeremy
>>
>>
>>  > -----Original Message-----
>>  > From: www-webont-wg-request@w3.org
>>  > [mailto:www-webont-wg-request@w3.org]On Behalf Of Jim Hendler
>>  > Sent: 24 January 2003 03:26
>>  > To: webont
>>  > Subject: question: datatype reasoning?
>>  >
>>  >
>>  >
>>  > I was asked the following by a colleague, we tried to find the answer
>>  > in the Semantics document, but we couldn't quite work out the
>>  > details.  Question is, does a complete Owl Lite or DL reasoner have
>>  > to do complete datatype reasoning?  i.e. for all the XML schema
>>  > primitive types, does a complete OWL reasoner have to be able to do
>>  > the correct class reasoning, etc -- knowing integers are numbers,
>  > > URIs are strings, etc. and appropriately applying these.
>>  >   If the answer is that an OWL system must do so, do we have any
>>  > implementation evidence to offer in this space?  If we don't expect
>>  > complete datatype reasoning, what level of such do we expect, and
>>  > where will we specify it (document-wise)
>>  >   thanks
>>  >   JH
>>  >
>>  >
>>  >
>>  >
>>  > --
>>  > Professor James Hendler				  hendler@cs.umd.edu
>>  > Director, Semantic Web and Agent Technologies	  301-405-2696
>>  > Maryland Information and Network Dynamics Lab.	  301-405-6707 (Fax)
>>  > Univ of Maryland, College Park, MD 20742	  240-731-3822 (Cell)
>>  > http://www.cs.umd.edu/users/hendler
>>  >
>>  >
>>


-- 
Professor James Hendler				  hendler@cs.umd.edu
Director, Semantic Web and Agent Technologies	  301-405-2696
Maryland Information and Network Dynamics Lab.	  301-405-6707 (Fax)
Univ of Maryland, College Park, MD 20742	  240-731-3822 (Cell)
http://www.cs.umd.edu/users/hendler

Received on Friday, 24 January 2003 17:26:42 UTC