[Versa] Versa issue: drawing the line on entailments

[Uche Ogbuji]

On Sun, 2005-08-14 at 22:23 -0600, Michael Olson wrote:
> On Aug 12, 2005, at 10:20 PM, Chimezie Ogbuji wrote:
> 
> > True.  I think the specification should *encourage* inferencing and
> > suggest where it can be done, but be clear not to make it a requiment.
> >  Though I think the language as it is currently doesn't make it a
> > requirement:
> >
> > type(resource)
> >
> > Returns a list of all resources of a specified type, as defined by
> > RDFS and optionally DAML schema specifications..
> 
> In my mind, I think we need to leave off the last little bit.

Agreed.

> a)  what id DAML is not supported?
> b)  what if OWL if supported?
> c)  what is some new funky technology is supported?

Heh.  I doubt you need to elaborate further :-)  The spec has lots of
cruft we should just lop off.

> I think it is best to just say "what ever the underlying model"  
> considers as all of the resources of the specified type.

Exactly.

> >
> > traverse(...) - regarding vtrav:transitive
> >
> > ... If omitted, no transitivity is applied except any applied by the
> > underlying model (e.g. if rdf:type statements are traversed, or
> > statements marked as daml:transitiveProperty). ...
> >
> 
> or subClassOf given rdfs8, or subPropertyOf given rdfs5 etc.
> 
> 
> Again, I think it is best to say something like
> 
> "with out the transitive flag, versa will return all direct objects of  
> the statemtne {resource}, {property}, *.
> 
> If the transitive flag is specified, then the versa implementation will  
> attempt to perform the transitive completion of the property and return  
> the computed set.

I think that we should allow the underlying model to apply entailment
for transitive properties if supported.  Of course, in such a model,
your statement of "direct object" technically covers it, but I think it
might be a bit misleading.  Perhaps it's enough to just drop "direct".


> It should be noted that
> 
> a)  this will take longer

I don't think this belongs in the spec.  It's an implementation detail.
True it's likely to hold for all implementations, but still...

> b)  if this is performed on a property that the model already considers  
> transitive then the results will grown exponentially"

I think a good implementation should catch this case, and handle it
elegantly.  It's pretty trivial to think of strategies for doing so.  I
don't think we need this statement in the spec.


-- 
Uche Ogbuji                               Fourthought, Inc.
http://uche.ogbuji.net                    http://fourthought.com
http://copia.ogbuji.net                   http://4Suite.org
Use CSS to display XML, part 2 - http://www-128.ibm.com/developerworks/edu/x-dw-x-xmlcss2-i.html
XML Output with 4Suite & Amara - http://www.xml.com/pub/a/2005/04/20/py-xml.html
Use XSLT to prepare XML for import into OpenOffice Calc - http://www.ibm.com/developerworks/xml/library/x-oocalc/
Schema standardization for top-down semantic transparency - http://www-128.ibm.com/developerworks/xml/library/x-think31.html