80 likes | 94 Views
DAML Language Breakout Deborah L. McGuinness Knowledge Systems Laboratory Stanford University www.ksl.stanford.edu. 2/15/01. Topics. Reification Defaults/Non-Monotonicity Proof Checking Concrete vs. abstract domains (pre-defined types vs. user-defined types) (datatypes vs. classes).
E N D
DAML Language Breakout Deborah L. McGuinness Knowledge Systems Laboratory Stanford University www.ksl.stanford.edu 2/15/01
Topics • Reification • Defaults/Non-Monotonicity • Proof Checking • Concrete vs. abstract domains (pre-defined types vs. user-defined types) (datatypes vs. classes)
Reification (from dictionary: To regard or treat (an abstraction) as if it had concrete or material existence.) • Requirement: Need to make statements about statements. • Extensive discussion concerning potential pitfalls, usage in rdf, relationship to RDF, requirements • Conclusion – if we include reification, we will expect problems • Speculation – we may not be able to avoid modalities as long as we might have thought.
Defaults/Rules • Currently no default/rule/negation as failure capability in DAML+OIL • Agreement that some solution is needed • Consider some mechanism on top of the core language • Observation that negation as failure in rule systems has long history of usage, tractability, etc. • Action Items – form rule mailing list (Ben Grosof will seed the list) • Small group followup meeting on topic
Proof Checks • Discussion on non-monotonicity led to a discussion of possibly checking non-monotonic proofs • Discussion followed on proof checks of (possibly partial) deductions • Can either trust reasoners or just check their proofs • Proof checking requires tractable proof checking systems • Followup among interested parties
Concrete Data Types • Consensus that concrete data type (numbers, strings, etc) are necessary • Consensus around basic proposal initiated by Horrocks and Patel-Schneider • Technical discussion about rdf syntax. • Peter (and Ian and Frank) will provide a modified proposal taking into account Connolly’s document http://www.w3.org/2001/01/ct24 • Expectation is that this will be included in the language prior to semantic web activity
Other issues? • Send email to rdf-logic mailing list (or dlm@ksl.stanford.edu)
Background • Attended by most of ad-hoc committee (physically or virtually) • Solicited feedback on current language and wishes for modifications/extensions • No feedback on current language other than extensions