From: Benjamin Grosof (bgrosof@MIT.EDU)
Date: 08/12/03
% notes from JC telecon 8/12/03
% by Benjamin Grosof
agenda:
- introduction of new proposal by Peter and Ian
- Rules Lite plans, including work items and volunteering
participants:
Mike Dean
Benjamin Grosof
Peter Patel-Schneider
Harold Boley
Said Tabet
Ian Horrocks
Pat Hayes
Sandro Hawke
? perhaps one or two others
Peter walked thru the proposal very briefly
Peter: body and head are each a conjunctive query
Peter: adding Horn to DL will result in undecidability
Benj:
issue: explicit inequality (i.e., differentFrom) in head goes beyond Horn;
perhaps reducible if there is a complete theory of equality, but is there
Peter:
another issue: only-if direction goes beyond Horn
Ian:
actually can express head equality given Horn plus DL expressiveness
Benj:
(some discussion around desirability/feasibility of decidability)
%%%%
some initial volunteering:
examples and test cases: Mike, Said, and maybe Deborah
warning label: Benj and Pat
concrete syntax: Harold and Said
abstract syntax: Benj ;
would also like later help from someone who worked on OWL abstract syntax
Benj: in addition it seems important to soon resolve also the ff.:
- expressiveness -- esp. what further expressive restrictions or additions
perhaps can defer the ff. additional work items:
requirements and objectives statement
updated roadmap
semantics -- is simple since is just FOL essentially
running of test cases using various existing tools
agenda for next week: schedule of these work items
summary of expressiveness issues, earlier discussed, that will need to
be resolved: [Benj, after the telecon, extracting from telecon discussion]
- Datalog restriction (probable)
- RDF lists permitted, but probably restricted somewhat:
. blank nodes -- avoid in head; perhaps also avoid in body
. avoid KIF's over-ambitiousness (consensus)
- avoid 0-ary predicates (probable)
- limiting equality as predicate
. in manner similar to OWL-DL (probable)
. esp. in head
- permit conjunction within head in rule syntax, using expressive reducibility
. naming of rules might be a complication
- naming of rules (perhaps defer beyond Rules Lite)
- how tight is layering semantically on top of RDF or OWL
- be able to handle distributed facts: import'ing, module union'ing
. Benj: propose to defer this a bit, beyond the basic proposal
Benj:
propose to keep it really simple:
- Datalog, no head b-nodes, no 0-ary predicates, no head equality,
conjunction within head, do permit rule names, do permit simple RDF lists
- but people should develop test cases if they're worried/irked by these
restrictions
Harold: Datalog can be problematic for lists unless have b-nodes
Mike:
would like to do b-nodes; will develop some test cases;
painful not to permit this
Sandro and Benj:
but that's head existential, and thus not Horn; and permitting just in body
would probably be confusing to users
Pat:
can get quite a bit of expressive power probably if do include
Benj:
how about we permit head conjunction
Benj:
is there a way to impose some restrictions on head b-nodes
that enables reducibility
Mike:
maybe the real issue is gensym capability, but can't create new names
based on variables if are restricted to be Datalog
Benj A to Pat Q: the reason to stick to Horn as well-studied expressive class
is that it has available:
proof theory, completeness, algorithms, implementations
that are tighter than full FOL
Benj: suggest the way to resolve this b-node issue is with some use cases
- Mike; Sandro will help review
agenda for next week:
- IJCAI report from Ian
- more on Peter and Ian's proposal
- schedule of our work items
- use cases on b-nodes
________________________________________________________________________________________________
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
http://ebusiness.mit.edu/bgrosof or http://www.mit.edu/~bgrosof
This archive was generated by hypermail 2.1.4 : 08/12/03 EST