From dingli1 at cs.umbc.edu  Mon Jul 31 22:52:17 2006
From: dingli1 at cs.umbc.edu (Li ding)
Date: Mon, 31 Jul 2006 22:52:17 -0400
Subject: [Inferenceweb] [Fwd: handling huge conclusion]
Message-ID: <44CEC1E1.5070508@cs.umbc.edu>

Lalana,
we have composed the PML N3 extension ontology at 
http://iw.stanford.edu/2004/07/iw-n3.owl.
I made some changes and do have one question.

Part I. term definition and usage (I have provided rdfs:comment in the 
ontology):

1. N3Formula    (owl:Class)
2.     hasN3string   (owl:DatatypeProperty)
3.     hasN3variable   (owl:ObjectProperty)
4.     iw:hasSource     
5. usesFormula  (owl:ObjectProperty)     domain-  iw:NodeSet   range-  
N3Formula

example instance file

#ns__123   rdf:type    iw:NodeSet
#ns__123   iwn3:usesFormular   #fm_1
#fm_1     rdf:type  iwn3:N3Formula
#fm_1     iwn3:hasN3variable     foo:__g2
#fm_1     iwn3:hasN3string     "@prefix ...... "
#fm_1     iwn3:hasSource        http://blah-blah.com/...


Part II notes and question
1. (note) I have changed the mapping (variable-uri, fomular-uri) , i.e. 
I moved variable ID to the formula definition
2. (question) variable ID is only unique within RDF graph, i.e. foo:__g2 
is a node ID in the serialized N3 formula. The two variables (blank 
nodes) in (i) the object of iw:hasConclusion and (ii) the object of 
hasN3string cannot be "semantically equivalent" after we restore n3 
string back to formula   even both of them serialized the variables 
using  foo:__g2.  This is OK for showing the end user the background KB, 
but not good for restore the original huge conclusion.  In other words, 
when I have two strings obtained from hasConclusion and hasN3string, I 
cannot use CWM to merge them to restore the original huge conclusion. 
Shall we live on that or do you have any good solutions?  
3. how can we obtain the URL of source that derives the n3 string?
4. I also attached some rdfs:label assertions in the same file. I put 
them in this file to keep PML ontology OWL-DL. They are not related to 
huge conclusion representation.

PartIII action items
1. we have update the PML ontology, but still need agreement. Currently, 
I need yours and cynthia's
2. after agreed on the ontology, can you modify cwm-pml.n3 to generate 
corresponding result?
3. after agreed on the ontology, Cynthia will change PML browser to 
support the ontology. But she is on vacation this week.


Best,

Li