Search Mailing List Archives


Limit search to: Subject & Body Subject Author
Sort by: Reverse Sort
Limit to: All This Week Last Week This Month Last Month
Select Date Range     through    

[protege-owl] SWRL doubt: using rules to add new individuals?

Nacho Mayorga nmayorga at gmail.com
Wed Aug 29 09:46:35 PDT 2007


Hi!,

I remember having read in a previous thread on this list that SWRL rules
were not meant to add _new_ and _permanent_ individuals as a result of the
firing of rules (it had something to do with variable instantiation?). Is
that true? Does this constraint apply to built-ins too?

The problem I wanted to solve is that of responding to events that would
require new individuals to be added to the ABox of an ontology. Whenever an
event happens, some information is to be gathered from the event and its
context as the base for creating new individuals. I thought of doing it by
using SWRL: the consequent of a rule would add the individual(s) after its
antecedent had collected the necessary information.  Is this possible?
advisable?

As a bit of explanation, the events I'm referring to are related to users'
actions, which are to provide (a part of) the context. The users would
interact with a web application. Some information would then be collected
(maybe, processed) and then fed to the KBMS. This contextual information
would help selecting the individual's class. The (proposed) rules would act
as a kind of translation system: abstracting contextual information, taking
into account available information and adding individuals representing new
knowledge derived from the event, the TBox and the ABox.

The other way I'd thought of consisted in using the Java OWL API to retrieve
the appropriate information and defining the new individual(s).
Nevertheless, the point of using rules was that a) they would be fired
whenever their conditions were met, either synchronously or asynchronously
with respect to the user events and b) they could be written in a way that
allowed new rules to be fired by the results asserted by other rules.

Comments would be very much appreciated.

Thanks a lot,

                          Nacho
-- 
Nacho Mayorga
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.stanford.edu/pipermail/protege-owl/attachments/20070829/4b317b04/attachment.html>


More information about the protege-owl mailing list