ISSUE-95: Proposal for recursive DatatypeRestrictionS [WAS: Agenda for teleconference Wednesday February 27th, 2008]

Hi!

Peter F. Patel-Schneider answered to Evan Wallace:

>> Boris' action ( documented in 
>http://lists.w3.org/Archives/Public/public-owl-wg/2008Feb/0129.
html) updated the text of
>> the structural specification document to allow 
>datatypeRestrictions only on datatypes.  I am uncomfortable 
>with this new
>> constraint, although I understand this was decided at last 
>week's telecon (I was out sick).  
>
>What would you prefer?
>
>The syntax restriction was put in to prevent constructs like:
>
>   DatatypeRestriction(DataComplementOf(xsd:integer) 
>                       minExclusive "1"^^xsd:integer)
>
>whose meaning is, at best, unclear.

W.r.t. ISSUE-95, I just wanted to bring up again the idea of (recursively) allowing DataRestrictionS on DataRestrictionS [10], which was based on a suggestion by Peter [20].

OWL authors would then be allowed to write expressions like:

  DatatypeRestriction(
    DatatypeRestriction(
      xsd:integer 
      minInclusive "18"^^xsd:integer )
    maxExclusive "65"^^xsd:integer )

This would be semantically identical to:

  DatatypeRestriction(
    xsd:integer 
    minInclusive "18"^^xsd:integer
    maxExclusive "65"^^xsd:integer
  )

Considerations:
---------------

* Usecases: This would allow the definition of named partial data ranges, which can be reused (I think that this is what Peter suggested in [20]). For example, the first example above could then be rewritten as:

  DatatypeRestriction(
    ex:Adult
    maxExclusive "65"^^xsd:integer )

where:

  ex:Adult := 
    DatatypeRestriction(
      xsd:integer 
      minInclusive "18"^^xsd:integer )  

* Semantical aspects: There would be no changes in semantics in comparison to the current proposal by Boris [30]. A recursive expressions would just be syntactic sugar for the respective "flat" DatatypeRestriction on a datatype URIs. The mapping from recursive to flat expressions would be unique modulo facet order. 

* Computational aspects: A simple preprocessor stage could do the linear recursive expansion into a "flattened" DatatypeRestriction. Afterwards a reasoner, which only knows about flat DatatypeRestrictionS, can work on the result.

Any comments?

Michael

[10] <http://lists.w3.org/Archives/Public/public-owl-wg/2008Feb/0033.html>
[20] <http://lists.w3.org/Archives/Public/public-owl-wg/2008Feb/0031.html>
[30] <http://lists.w3.org/Archives/Public/public-owl-wg/2008Feb/0129.html>

--
Dipl.-Inform. Michael Schneider
FZI Forschungszentrum Informatik Karlsruhe
Abtl. Information Process Engineering (IPE)
Tel  : +49-721-9654-726
Fax  : +49-721-9654-727
Email: Michael.Schneider@fzi.de
Web  : http://www.fzi.de/ipe/eng/mitarbeiter.php?id=555

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 Regierungspräsidium Karlsruhe
Vorstand: Rüdiger Dillmann, Michael Flor, Jivka Ovtcharova, Rudi Studer
Vorsitzender des Kuratoriums: Ministerialdirigent Günther Leßnerkraus