Search Mailing List Archives
[protege-owl] SWRL to "move" individuals?
martin.oconnor at stanford.edu
Tue Aug 28 02:04:07 PDT 2007
Only a very resticted subset of SWRL rules can be saved as equivalent
OWL constructs. Conversely, SWRL can not directly represent most OWL
contructs - it requires the use of OWL anonymous classes to do so.
For example, to state that individuals of a particular class, A, have
a property hasP with a cardinality restriction of 1 can be stated in
A(?a) -> (hasP <= 1)(?a)
The expression in parenthesis is an OWL expression and expressing it
using SWRL would not give any particular advantages - and one would
still be required to be familiar with OWL expressions.
Non trivial use of OWL requires an intimate knowledge of OWL
constructs. SWRL provides no magic bullet here.
Quoting Kaarel Kaljurand <kaljurand at gmail.com>:
> On 8/27/07, Thomas Russ <tar at isi.edu> wrote:
>> On Aug 27, 2007, at 2:59 AM, Kaarel Kaljurand wrote:
>> > On 8/25/07, Martin O'Connor <martin.oconnor at stanford.edu> wrote:
>> >> In general, if you can do something in OWL, you should.
>> > I'm just curious. Why doesn't SWRLTab just save the rule as OWL it
>> > this happens to
>> > be is possible?
>> Well, for starters, this isn't necessarily an easy thing to figure out.
>> While it may be possible to do this for some relatively simple patterns,
>> in general it would take a theorem prover to figure out and may end up
>> being undecidable. It involves figuring out that a particular logical
>> expression in SWRL entails some equivalent in OWL, and that will not be
>> an easy task, especially given that you have to construct the potential
>> OWL equivalent.
> They do it in:
> http://www.mindswap.org/papers/CautiousSWRL.pdf (section 3)
> which is quite easy.
> protege-owl mailing list
> protege-owl at lists.stanford.edu
> Instructions for unsubscribing:
More information about the protege-owl