- From: Francois Bry <bry@ifi.lmu.de>
- Date: Wed, 08 Mar 2006 11:52:49 +0100
- To: Hassan Aït-Kaci <hak@ilog.com>
- CC: public-rif-wg@w3.org
Hassan Aït-Kaci wrote: > Now PRs are NOT driven by names. Since a rule is not using a relational > or functional name in its head to drive the rule's call, it is simply > not possible to have recursive rules. +1 > Does it mean PR's cannot compute iteratively? Certainly not: there is > an underlying loop that acytivates the rules based on the data present > in the working memory (or Extensional DB, or Fact Base, etc...). Viz., > > WHILE [some rules match some objects in the WM] > DO [choose a rule and all the objects it matches] > [do the action of the rules on all these objects.] +1 > In conclusion, while it is possible to simulate one system in the > other (e.g., by mere Turing equivalence), it may be contended that > the translations to and from each side (PRs and Pure Horn) are, IMHO, > non-trivial and non-intuitive. +1 > So, Frank is basically right: what Harold et al.'s Road Map defines > as "pure PRs" is computationally uninteresting and its rendition in > Pure Horn is likely to be at odds with a rendition for Full PR. +1. Francois
Received on Wednesday, 8 March 2006 10:52:55 UTC