Lists Home |
Date Index |
On Wed, 17 Nov 2004 16:44:55 -0500, Jonathan Borden
> Michael Champion wrote:
> > If so, this is a very interesting question. First, Ontologies are
> > unlikely to get much traction, IMHO, as long as they are called
> > "ontologies", and one is forced to be conversant with formal
> > semantics, formal logics, etc. in order to use them in an enterprise
> > IT project. Somehow this approach has to be repackaged in a way that
> > rests cleanly on the foundation of semantic web theory/technology,
> > but exposes only those concepts (and terminology) that are accessible
> > to ordinary mortals.
> You are probably correct. One way to think about it is the fact that
> UML is basically a subset of "OWL" (the formal part) --- more
> concretely that OWL (the language) is an excellent way to write down
> UML models.
> Of course this leads us to ask whether UML it either a) useful or b)
> accessible to ordinary mortals :-)
Interesting, I would have never thought to question this. I find UML
useful and accessible but a bit of a pain to use. I used Booch's
modelling when it was first published and have had a long history with
OO design. It didn't come easy at first, but now seems natural. I
think that OO development and the related modelling takes a bit of
getting used to, but there's certainly a fair percentage of the
developer population for which it works. Don't know if those people
are ordinary, but they seem mortal enough...
> In any case, people who are already conversant in UML should find OWL
> not so intimidating.
I don't find it intimidating, just a whole lot of writing. The
ability to produce it using graphical modelling tools would definitely
make it far more approachable and I believe there are people
(companies?) doing this? Haven't need to look yet.
<snip>discussion on Ontology building approaches and bottom up/top
> I've said this before, but I personally believe that some way to
> combine the two is the key.
Makes sense. Certainly the vertical industries and the government will
produce top down approaches. Left on their own I think it would be a
very long time before these ever find enough intersection to branch
out and meld into something globally traversable. Search driven
approaches do seem to allow that gap to become almost irrelevant for
most humans, but that still leaves the automated machine driven
Bottom up thin horizontal layers like FOAF seem like ideal candidates
to branch out and form a basis for connecting some of the dispersed
vertical domains, in particular if you imagine FOAF (or variants)
being used to build informal linkages between things that aren't
people (which is already happening): then you get people describing
informal networks of relationships in machine readable form. Having
the average Joe being able to annotate their personal Web pages with
"things I Like" and "things I trust" as well as "people I know" could
have a huge network effect, in particular when some of the things they
like or trust turn out to be Ontologies (implicitly or explicitly).