[Reaction-tg] advantages of Situated LP for Reaction RuleML

[Benjamin Grosof]

Hi folks,
Partly in prep for discussion of the TG's directions and the current white 
paper,
Asaf asked me to post a summary of what are the advantages of Situated LP 
for Reaction RuleML.
I talked about these points briefly at the end of the 2/17 telecon that 
discussed Situated LP overall.

Advantages of Situated LP for Reaction Rules work:
- highly declarative semantics for procedural attachments, much more so 
than previous approaches;
       moreover, as a direct expressive generalization of declarative 
pure-belief/derivation LP rules
- extends to (already) include nonmonotonicity
       -- NAF and courteous priorities -- and identifies
       simple/easier vs. complex/trickier aspects of that;
       concepts, use cases, expressiveness, formalization, algorithmic 
techniques;
       more general and clean than current production or ECA systems
- treatment of sensor binding restrictions:  concepts, use cases,
       expressiveness, formalization, algorithmic techniques;
       more general and clean than current production and ECA systems
- translation approach to production rules (Jess) from {core RuleML
       and Prolog and SQL RDBMS}, bidirectionally; this is first of a kind
- does have basic concepts of updating events and incremental processing 
(inferencing+action);
       (Situated LP lacks a concept of an explicit event part of a rule, 
but can represent basic
       kinds of event-based/triggered behavior.)

In addition, more work is needed of course, especially to connect to a 
richer events concept.


Comments are most invited.
Benjamin



________________________________________________________________________________________________
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