[protege-discussion] Bad interaction between Facet Constraints Tab, template values in slots?

[Richard Dunlap]

To follow up with a diagnosis and fix:

The problem occurred when the slot had multiplicity 1; when multiple 
values were allowed, the problem went away.  This suggested that in the 
single-value case, the template value was being pulled twice, once as a 
template value and once as an own value: changing 
FacetConstraintsValidator.GetTemplateAndOwnSlotValues, which reads

{
       Collection val, val2, t, t2;
       t = c.getTemplateSlotValues(st);
       val = new ArrayList(t);
       t2 = i.getOwnSlotAndSubslotValues(st);
       val2 = new ArrayList(t2);
       val2.addAll(val);
       return val2;
}

to read

{ return i.getOwnSlotAndSubslotValues(st); }

appears to have solved the problem.

-- Richard Dunlap

Richard Dunlap wrote:
> In a new Protege-Frames project:
> 
> * Create a class TestClass.
> 
> * Add to TestClass a new template slot testSlot, leaving the value type
> as String, and set a template value for the slot (e.g. "testValue").
> 
> * Create an instance of TestClass -- as expected (I hope :-) ), the
> instance has one value slot with the unchangeable value "testValue".
> Good so far.
> 
> *  Open up the Facet Constraints Tab and check the facet constraints of
> all classes.  The previously created instance is tagged as an instance
> with a violation.  Not so good.
> 
> Has anyone run across this behavior or something related before?  Is 
> there a reason, perhaps, that the observed behavior is correct?
> 
> We've also noticed that the FacetConstraintsTab is marked as no longer 
> supported on the Wiki and that the source code apparently needs to be 
> downloaded separately from the various packages.  Is there a different, 
> better tab widget to solve the problem "give me a list of my red boxes"?
> 
> Thanks!
> 
> -- Richard Dunlap
> 
> 
> 
> _______________________________________________
> protege-discussion mailing list
> protege-discussion at lists.stanford.edu
> https://mailman.stanford.edu/mailman/listinfo/protege-discussion
> 
> Instructions for unsubscribing: http://protege.stanford.edu/doc/faq.html#01a.03 
> 
>