About the implementation of swrlb:empty in protege 3.4 swrlTab

[edison]

Dear protege-owl users:

Recently we built up the ontology about species of plants and the specimen related to different specie. We also enabled the SWRLTab, set up several SWRL rules and tested them by JESS rule engine. When we try to find out the species with no specimen produced, the rule produce(?x, ?y) ^ swrlb:empty(?y) -> sqwrl:select(?x) was invalid. The rule editor shows that swrlb:empty is not valid, but the protege SWRLTab webpage shows they do have the Builtin swrlb:empty. So Do you know what the problem is? Where in the rule should I modify? Thanks for help!

Feng-chien Chung
Metadata Architecture and Application Team
Academia Sinica, Taiwan



______________________________________________________________________________________________________
付費才容量無䞊限Yahoo!奇摩電子信箱2.0免費絊䜠信件氞遠䞍必刪 http://tw.mg0.mail.yahoo.com/dc/landing

[Thomas Russ]

Post by edison
Recently we built up the ontology about species of plants and the
specimen related to different specie. We also enabled the SWRLTab,
set up several SWRL rules and tested them by JESS rule engine. When
we try to find out the species with no specimen produced, the rule
produce(?x, ?y) ^ swrlb:empty(?y) -> sqwrl:select(?x) was invalid.
The rule editor shows that swrlb:empty is not valid, but the protege
SWRLTab webpage shows they do have the Builtin swrlb:empty. So Do
you know what the problem is? Where in the rule should I modify?
Thanks for help!

Your fundamental problem will be open world reasoning issues. The
only way for an OWL open world reasoner to be able to tell that there
is no value for a property is to either be told that there is no
value, or else have some way of proving that no value is possible.
The latter is very hard to do. So just looking at the fillers doesn't
allow answering such a question under open world semantics.

You may need to resort to using some procedural code to achieve what
you want. That would involve getting all of the species, all of the
species with a specimen and then performing your own set difference
operation.

_______________________________________________
protege-owl mailing list
protege-***@lists.stanford.edu
https://mailman.stanford.edu/mailman/listinfo/protege-owl

Instructions for unsubscribing: http://protege.stanford.edu/doc/faq.html#01a.03

[Martin O'Connor]

Also, as mentioned in the wiki documentation [1], list built-ins are
also not yet implemented in the SWRLTab.

Even if they were, however, they will not work as you expect below - the
open world assumption still applies.

Martin

[1] http://protege.cim3.net/cgi-bin/wiki.pl?CoreSWRLBuiltIns#nid98T

Post by Thomas Russ

Post by edison
Recently we built up the ontology about species of plants and the
specimen related to different specie. We also enabled the SWRLTab,
set up several SWRL rules and tested them by JESS rule engine. When
we try to find out the species with no specimen produced, the rule
produce(?x, ?y) ^ swrlb:empty(?y) -> sqwrl:select(?x) was invalid.
The rule editor shows that swrlb:empty is not valid, but the protege
SWRLTab webpage shows they do have the Builtin swrlb:empty. So Do
you know what the problem is? Where in the rule should I modify?
Thanks for help!

Your fundamental problem will be open world reasoning issues. The
only way for an OWL open world reasoner to be able to tell that there
is no value for a property is to either be told that there is no
value, or else have some way of proving that no value is possible.
The latter is very hard to do. So just looking at the fillers doesn't
allow answering such a question under open world semantics.
You may need to resort to using some procedural code to achieve what
you want. That would involve getting all of the species, all of the
species with a specimen and then performing your own set difference
operation.
_______________________________________________
protege-owl mailing list
https://mailman.stanford.edu/mailman/listinfo/protege-owl
Instructions for unsubscribing: http://protege.stanford.edu/doc/faq.html#01a.03

_______________________________________________
protege-owl mailing list
protege-***@lists.stanford.edu
https://mailman.stanford.edu/mailman/listinfo/protege-owl

Instructions for unsubscribing: http://protege.stanford.edu/doc/faq.html#01a.03