p4 rule widget doesn't keep rules?

classic Classic list List threaded Threaded
5 messages Options
Reply | Threaded
Open this post in threaded view
|

p4 rule widget doesn't keep rules?

howard goldberg
build 64, build 101
Entered a SWRL rule into the rules widget (Ontology views->Rules).  Hitting enter, no rule is retained.
Bug, not implemented / placeholder ???

Howard

_______________________________________________
p4-feedback mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/p4-feedback
Reply | Threaded
Open this post in threaded view
|

Re: p4 rule widget doesn't keep rules?

Nick Drummond
Hi Howard,

There has not been any serious effort into implementing the rules view as we are hoping that a more substantial editor will be available when the P3.4 SWRL support is ported. The current view is really just there to display rules that exist in the ontology. Currently no editing is possible. We should have disabled this.
I will check the status of the parsers in the OWL API and see if we can hook this up, and if not possible immediately, will disable the add button.

Nick

On Thu, Sep 18, 2008 at 4:56 AM, howard goldberg <[hidden email]> wrote:
build 64, build 101
Entered a SWRL rule into the rules widget (Ontology views->Rules).  Hitting enter, no rule is retained.
Bug, not implemented / placeholder ???

Howard

_______________________________________________
p4-feedback mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/p4-feedback



_______________________________________________
p4-feedback mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/p4-feedback
Reply | Threaded
Open this post in threaded view
|

Re: p4 rule widget doesn't keep rules?

Bijan Parsia-3
On 22 Sep 2008, at 13:26, Nick Drummond wrote:

> Hi Howard,
>
> There has not been any serious effort into implementing the rules  
> view as we are hoping that a more substantial editor will be  
> available when the P3.4 SWRL support is ported.
[snip]

Or we implement something entirely new. I have some UI ideas....

Cheers,
Bijan.

_______________________________________________
p4-feedback mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/p4-feedback
Reply | Threaded
Open this post in threaded view
|

Re: p4 rule widget doesn't keep rules?

Nick Drummond
I'd like to hear these ideas.

It would let us sketch out some interface reqs for SWRL.

(You will have to bear in mind I know v little about SWRL).

Nick

On Mon, Sep 22, 2008 at 3:09 PM, Bijan Parsia <[hidden email]> wrote:
On 22 Sep 2008, at 13:26, Nick Drummond wrote:

> Hi Howard,
>
> There has not been any serious effort into implementing the rules
> view as we are hoping that a more substantial editor will be
> available when the P3.4 SWRL support is ported.
[snip]

Or we implement something entirely new. I have some UI ideas....

Cheers,
Bijan.

_______________________________________________
p4-feedback mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/p4-feedback


_______________________________________________
p4-feedback mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/p4-feedback
Reply | Threaded
Open this post in threaded view
|

Re: p4 rule widget doesn't keep rules?

Bijan Parsia-3
On 23 Sep 2008, at 10:29, Nick Drummond wrote:

> I'd like to hear these ideas.
>
> It would let us sketch out some interface reqs for SWRL.
>
> (You will have to bear in mind I know v little about SWRL).


With a slight restriction, we can make SWRL fit in nicely with the  
current P4 paradigm. We did an experimental version of this in Swoop.

Essentially, we already have this with role chains. Role chains  
correspond to the body of a rule where the property constrained by it  
is the head. Thus, for any rule with a class or a property atom  
(alone) as the head, we can, in principle, accomodate the body as  
another kind of restriction. E.g., such a rule as:

        C(X) :- D(X), P(Z, Y), Q(X, Y), D(Z), X!=Z.

Could be represented as:

Class: C
...
...
InstanceQuery:
        D(X), P(Z, Y), Q(X, Y), D(Z), X!=Z

The main challenge is that, unlike property chains, we cannot hide  
the presence of variables. Most importantly we need to specify the  
variables shared by the head. We could either make these fixed (e.g.,  
X) or we could have some marker, or we could have a convention e.g.:
        D(aC), P(Z, Y), Q(aC, Y), D(Z), aC!=Z

For rules with compound heads, we could treat them as GCIs (i.e.,  
poorly).

Cheers,
Bijan.

_______________________________________________
p4-feedback mailing list
[hidden email]
https://mailman.stanford.edu/mailman/listinfo/p4-feedback