Re: updated document with possible directions for SWRL

From: Peter F. Patel-Schneider ([email protected])
Date: 12/08/03

  • Next message: Mike Dean: "(no subject)"
    Some comments.
    
    peter
    
    
    > % suggestions on directions for SWRL 12/02/03
    > % by Benjamin Grosof
    > 
    > Here's a quick list of possible directions for SWRL:
    > 
    > o expressive features to extend to incorporate:
    > 
    > explicit equality
    
    How is this different from sameAs and differentFrom, which are already in
    the proposal?
    
    > n-ary predicates
    
    How can a rules proposal that builds on RDF do this?
    
    > logical functions
    
    What do you mean by logical functions?
    
    > sensor built-ins, 
    >   e.g., datatype comparisons and arithmetic conversions cf. Mike Dean's note
    
    Umm, I don't view datatype comparisons as ``sensor'' built-ins.  
    
    > some of Lloyd-Topor transformation, 
    > e.g., disjunction in the body (and not just within complex class descriptions)
    
    This was discussed at length in last week's call and could be problematic.
    
    > modules and importing
    
    What is missing here that cannot be handled by owl:imports?
    
    > skolemization, existentials, blank nodes:  there's a set of issues here
    
    Agreed.  There are lots of possibilities. 
    
    > more general sensors, e.g., to call databases or knowledge bases via
    >  web service protocols
    
    Hmm.  There are two issues here.
    1/ Accessing information that is not in RDF/OWL/OWL Rules.
    2/ Accessing information using services.
    It seems to me that both of these are the province of a broader effort
    within the W3C SW effort.
    
    > effectors, i.e., actions triggered by conclusions
    
    I wonder how this can be reconciled with a logical view of rules.
    
    > negation-as-failure, prioritized conflict handling, scoping for nonmonotonicity
    
    There are a number of things that can be said here, but most of them have
    more to do with RDF and the SW as a whole.
    
    > classical negation and strong negation 
    
    Ditto.
    
    > o syntax:
    > 
    > revise RuleML syntax to update it to superset the simplifications and revisions
    > made in a subset of it for SWRL
    > 
    > more on RDF and OWL syntax for SWRL
    > 
    > more in direction of FOL, coordinate that with SCL
    > 
    > 
    > o in document:
    > 
    > more on implementation
    > 
    > more on design rationale
    > 
    > more on warning label
    > 
    > more on use cases and examples
    > 
    > more on theory of integrating the rules with OWL within FOL / classical logic
    


    This archive was generated by hypermail 2.1.4 : 12/08/03 EST