W3C home > Mailing lists > Public > www-rdf-interest@w3.org > May 2000

Re: Converting SHOE to RDF: about 2/3 done; some gotchas

From: Ken Laskey <KENNETH.J.LASKEY@saic.com>
Date: Fri, 19 May 2000 19:20:00 -0400
Message-Id: <p04310101b54b76c73554@[10.43.242.86]>
To: Jeff Heflin <heflin@cs.umd.edu>, Dan Connolly <connolly@w3.org>
Cc: Sean Luke <seanl@cs.umd.edu>, www-rdf-interest@w3.org
Without going through the details of the give and take over the past 
week, I was struck by a couple points.

- I think there is agreement that someone should not be able to 
change someone else's schema.  However, someone else can make a 
parallel schema and assert a correspondence.  The intent may be 
malicious or it just may be an attempt to add capability or a 
different perspective.  Thus, if maliciousness is not an issue, then 
the issue is not trust but competency and opinion.

- This brings me to the question of data represented by a schema. 
What inferences can be drawn if the data is contradictory.  Take for 
example this discussion. Imagine representing the assertions about 
the correspondence between RDF and SHOE.  Again, the issue is not one 
of trust but of opinion that in some cases has direct contradictions. 
Can more be inferred than a conversation is taking place?

- The concept of trust needs to be extended to a measure of authority 
on who is making an assertion.  Certainly Jeff Heflin's statements 
should probably carry more weight on SHOE and Dan Connolly's more 
weight on RDF, but how do you factor in a new opinion from left field 
that brings insight by asking basic questions that the experts were 
all sure were already answered?  Is there a question of trust?  How 
do you assess validity?

It appears that we're trying to draw conclusions from information we 
have yet to determine how to accurately represent.  Moreover, we are 
trying to provide a means to represent information without 
controlling what that information is.  From a Web perspective, this 
is how it needs to be.  While keeping an eye on a greater goal, might 
it not be helpful to make sure we can assert things in a Web 
environment and unambiguously know what has been asserted before we 
draw conclusions.

Ken
Received on Friday, 19 May 2000 19:20:05 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 7 December 2009 10:51:43 GMT