% notes from JC telecon 6/17/03
% by Benjamin Grosof  

agenda  discussion on ASCII syntax, presentation by Harold Boley

Harold walked thru the whole thread 
starting with 
http://www.daml.org/listarchive/joint-committee/1361.html
ending with 
http://www.daml.org/listarchive/joint-committee/1376.html


issue:  semantics of sequence/list and "rest" variables

Pat:  cf. SCL, sequence/list variables 
can take one beyond 1st-order in general, even violate compactness
(theorem not following from any finite subset of an infinite set of axioms)
- e.g., can axiomatize arithmetic
- depends on universally quantified sequence ("@") variables

Benj Q:  intuitively, seems OK if are doing existentials as queries in LP

Pat A:  probably is OK, esp. since in LP have Herbrand/minimal models


Benj:  observation:
Harold's proposal wrt permitted appearance of sequence (",") in argument
collections simplifies the nesting syntax for tup's and roli's that
was proposed in RuleML V0.8

%%

Benj: desirable often to permit users to specify and employ a
particular sequencing (rather than only a generic
lexicographic/canonical normalization as basis for positionalizing) be
declared for a given predicate's arguments even when it has OO
argument slot names as well

Harold:  yes, can do this if have signature declarations, cf. OO modeling
systems like Protege or Description Logic, e.g., using classes of fillers

Benj:  sometimes have two arguments/slots with the same filler class
-- Pat:  or overlapping classes -- 
so we want a bit different kind of signature declaration

Benj:  also when unambiguous one may want to refer to the slots by their
filler classes

%%

Benj: cautionary note:
",," or ";;" is dangerous human syntax for rest variables --
easily could be a typo

Pat:  yes
Harold:  I'm not attached to these 

%%

agenda for next week:  explicit equality and semantics; and roadmap