thoughts on ontology discussion at last telecon

Dear URW3 Participants,

Unfortunately, I will not be able to attend the telecon this week but  
I have been reading through the minutes from 21 November and wanted  
to provide my thoughts.

3. Report on Busan F2F
Kathy noted that Thomas gave a summary in email. http://lists.w3.org/ 
Archives/Member/member-xg-urw3/2007Nov/0012.html
Claudia said that the uncertainty ontology is useful for annotating  
use cases, but not that much for exploration and mining of uncertain  
knowledge.
Ken: I don't think the uncertainty ontology was meant for mining  
values.  It was to help us characterize uncertainty and give us a  
means to identify uncertainty in our use cases.  Given the  
uncertainty we saw in the use cases, we could posit approaches to  
dealing with those types or aspects of uncertainty and finally look  
at what information needs to be conveyed for someone to analyze real  
instances of our use case problem categories.
ACTION: Authors annotate use cases with uncertainty ontology.
Kathy said that Discovery use case has been annotated. http:// 
www.w3.org/2005/Incubator/urw3/wiki/Discovery
Peter isn't sure how far we need to go on annotating use cases, he  
suggests a more detailed annotation which shows that the ontology can  
help in finding data.

Ken: Identifying specific data from our use cases would require much  
more detail in each case and go beyond our charter. The scope was to  
identify the types of data needed and see whether there is a set that  
could be standardized.  While means to extract uncertainty values is  
indeed important, it goes beyond the resources for our one-year effort.

Kathy said that this upper level ontology is not sufficient, but  
shows necessity for future work on uncertainty interchange format.
Claudia thinks that we need also to define in a more precise way what  
we can do with such annotations.

Ken: The annotation of the use cases was supposed to help the XG  
identify the important properties by which we would suggest real  
problems be characterized.  While there may be some overlap, the  
uncertainty ontology was intended to be a tool for the XG and not  
necessarily the final answer for use by everyone else.

Kathy asks: For instance, how do these annotations help a software  
tool to work with the use case? Mitch, can you comment on this?
Peter notes: GRDDL.
Thomas asked whether Peter thinks annotation should be in RDF.
Peter is not so worried about the syntax. He is concerned about the  
semantic content.

Ken: I agree with Peter.  We are interested in the semantic content  
of the uncertainty information we identify; follow-on groups can  
worry about syntax.

Kathy thinks that it should support automated sharing of data.
Mitch agrees that ontology is not detailed enough for some purposes.  
His intent is that it is a top level structure.
Kathy said that Mitch developed the ontology to help us evaluate  
whether we were covering all the essential aspects of uncertainty. If  
we annotate use cases, we can uncover holes in our use cases. We can  
find kinds of uncertainty for which we don't have a use case, and  
develop a use case to fill that gap.

Ken: This was the original intent.

Kathy asks whether we need to fully work out the ontology to meet our  
charge?
Mitch thinks we do not have time to fully work out an ontology.
Claudia and Peter think we need to go lower, identify where in the  
ontology the interchange format fits.

Ken: I would be interested to see where this would lead.

Mitch thinks we can do this with the speaker scenario
ACTION: Mitch, Peter and Kathy to develop a speaker use case and link  
to ontology.
Please continue the work.  I am planning to be back for the 19  
December telecon.

Ken

------------------------------------------------------------------------ 
-----
Ken Laskey
MITRE Corporation, M/S H305      phone: 703-983-7934
7151 Colshire Drive                         fax:       703-983-1379
McLean VA 22102-7508