[
Lists Home |
Date Index |
Thread Index
]
costello@mitre.org (Roger L. Costello) writes:
>[much excellent material on holistic approaches, f-stops]
>
>There may be untility to treating semantics with a sterilized,
>laboratory approach. Certainly if this was the 17th century, where
>computers weren't available, then such a static, taxonomy-like approach
>would be acceptable. But in today's dynamic, computer-driven world
>surely we can do better ... much better.
17th-century taxonomies were simpler than most of the ones we have today
- partly because of media limitations (books vs. terabyte graphs) - but
the lack of computers itself had a beneficial side-effect. Humans could
still look beyond taxonomies, using their own local expectations, and in
many ways they had (and have) far more flexibility than computers today
have. (That they did so while still claiming to be observing such
taxonomies is perhaps more evidence of their flexibility!)
In some ways, the skepticism of 18th century philosophers (Hume and
Berkeley spring to mind) comes from the mismatches they saw between ways
of categorizing and describing the world and their experience of the
world. During much of the 19th century that skepticism was replaced by
confidence, and that confidence continues at least in Anglo-American
philosophy today.
A lot of my objections to ontologies and in particular visions of the
Semantic Web arise from my concerns about what I see as this sadly naive
confidence in "sterilized, laboratory approaches" to semantics. The
reliance on first-order logic seems fueled by a belief that if we only
had enough assertions, we could make all of these semantics fit
together, uh, better... much better.
>I believe that Didier and Mike Champion made mention of Google as tool
>which provides semantics in a holistic fashion. I found their
>statements extremely enlightening I totally agree with them. Yes, I
>think that Google is the best semantics tool today.
>
>While Google provides semantics in a holistic fashion, it is more or
>less semantics for eyeballs, i.e., the results it returns is intended
>for humans to process. The critical problem is how to create a tool
>which provides semantics in a holistic fashion *for computers*. Would
>someone care to take a stab at characterizing the nature of such a
>tool?
For now, though it's not quite the characterization you're seeking, I'd
propose that such tools combine the speed, efficiency, and repeatability
of computers with the flexibility and unpredictability of humans. That
would mean building systems which have humans in the loop, not just on
the edges of the system as providers or consumers.
I don't think this problem is as simple as computers vs. humans. I think
building systems that can effectively work with semantics requires both.
That's probably bad news for the folks who want to feed huge amounts of
information into a computer (which doesn't receive health benefits) and
get results without having to deal with humans (who don't always show up
for work). It seems, however, like a much richer approach than the
current "reduce it to bits which approximate yesterday's reality and
hope the computer recombines them to reflect tomorrow's reality"
approach.
(I think treating XML as a meeting point between humans and computers is
an important ingredient in making that work, though that often seems an
unpopular approach with both the humans, for whom "Tags are the Real
Enemy" [1] and the folks running the computers, who don't want to have
to deal with error-handling when humans create unexpected situations.)
[1] - http://www.xml.com/pub/a/2003/10/01/deviant.html
--
Simon St.Laurent
Ring around the content, a pocket full of brackets
Errors, errors, all fall down!
http://simonstl.com -- http://monasticxml.org
|
