Thread Links Date Links
Thread Prev Thread Next Thread Index Date Prev Date Next Date Index

Slides from ICCS'04



At the ICCS'04 conference, which was held from July 19
to 23, Jim Hendler and I were planning to debate some
issues about the Semantic Web.  Unfortunately, Jim had
to cancel.  Fortunately, Jeff Heflin was able to take
his place and give a talk on the Semantic Web.  However,
there wasn't much of a debate:  Jeff mostly agreed with
what I said in my talk, and I mostly agreed with what he
said in his talk.  But there were major differences in
emphasis and point of view.

Following are my slides:

    http://www.jfsowa.com/talks/iccs04.htm
    Graphics and Languages for the Flexible Modular Framework

And following are Jeff's slides:

    http://concept.cs.uah.edu/ICCS04-Web/heflin-semweb.ppt
    Knowledge Representation Issues for the Semantic Web

After Jeff's talk, I made only one comment:

 > I liked your talk.  Those ideas were pefectly fine
 > 20 years ago, and they're still just as good today.
 > And the unsolved problems were unsolved 20 years ago,
 > and they'll still be unsolved 20 years from now.

That last line might sound pessimistic, but I didn't
intend it to be.  Really difficult problems take a very
long time to solve.  Most progress is not made by facing
the hardest problems head on, but by discovering new
problems which are easier to solve, but whose solutions
can have as much or more benefit than a solution to
some of the harder problems.

In my talk, I discused John McCarthy's Elephant 2000
language as an approach that has a much more solid semantic
foundation than the current plans for the Semantic Web.
That does not mean that the current work is bad, but
that more can and should be done.

Furthermore, I believe that Elephant 2000 and other themes
that I addressed in my talk, such as the use of controlled
natural languages, suggest a number of solvable problems
that can make important progress in a relatively short time
(i.e., considerably less than 20 years).

Some references for further reading are included in my last
slide, but anyone who wants to read McCarthy's Elephant paper
can just type two words to Google:  McCarthy Elephant

John Sowa