my notes from JC telecon 5/6/03 on roadmap for rules (inline'd and attached)

From: Benjamin Grosof (bgrosof@MIT.EDU)
Date: 05/09/03

  • Next message: Benjamin Grosof: "RuleML's Horn logic program semantics: draft (inline'd and attached)"
    Hi folks,
    Here are my notes from the last telecon.  Mike:  you may find these useful 
    to merge some into the minutes.
    % notes from JC telecon 5/6/03
    % by Benjamin Grosof
    background docs from earlier today:
    see rules roadmap strawman note by Mike Dean
    see rules roadmap note by Benjamin Grosof
    - defer OWL or RDF syntax -- it's less urgent
    - instead focus on Horn LP in XML syntax
    (Mike presented his strawman roadmap plan)
    (Benjamin presented his roadmap suggestions, with plan of features to address)
    consensus:  these both seem good way to go,
    we'll tweak the suggested plan of features as we proceed.
    Pat sugg:
    semantics should clarify what happens when combine
    rulesets from different sources
    modules is crucial for this
    Mike:  what kind of semantics should we expect to
    have in phase I.  E.g., do we want a model theory?
    Benj A:  I think yes.
    There's a standard notion in LP literature:
    entailed Herbrand-universe-instance ground atoms;
    only have differences when intro NAF/nonmon
    Pat:  want to be able to support reasoning about
    same-as wrt URI's
    Stefan and Benj:  can then add explicit theory of equality
    Mike:  should we have semantics defined in
    a way that closely corresponds to that of OWL-DL?
    some discussion ensued, esp. with Pat, Benj, Ian, Stefan
    about the ff.:
    issue:  when does Horn LP entailment / non-entailment / model
    correspond to classical entailment / non-entailment / model
    e.g., "no" as answer from an inferencing system
    sometimes is meant as "don't know"
    vs. sometimes as "not entailed despite having done complete inferencing"
    obs:  Horn LP entailment is weakening of Horn FOL entailment,
    but is complete for an important expressive category
    of queries, e.g., conjunction of ground atomic queries
    plan on next steps:
    todo's for next week:
    Benjamin will  send out doc with the standard definition of Horn LP semantics,
    then he and others will send more detailed sets of comments later.
    Said, Benjamin and Harold will
    send out doc's on Horn RuleML V0.8 esp. its syntax
    Pat will do
    ditto for SCL (Simple Common Logic) esp. Horn case of that
    with a few examples for each
    e.g., can send RuleML ex's to Pat to encode
    then discuss at the next telecon
    walk-thru of Horn RuleML V0.8 by Benj and of SCL by Pat,
    then compare
    no JC telecon week of WWW (2 weeks from now)
    Prof. Benjamin Grosof
    Web Technologies for E-Commerce, Business Policies, E-Contracting, Rules, 
    XML, Agents, Semantic Web Services
    MIT Sloan School of Management, Information Technology group or

    This archive was generated by hypermail 2.1.4 : 05/09/03 EST