[From Bill Powers (920602.2000)]
Penni Sibun (920602) --
(/o^")_ Submitter is taking it lying down. (hint: nose up)
I split my sides -- thanks :-(haw)
My objection to jargon -- sometimes -- is that it makes things seem
important and new that are neither. Consider "situated" modeling. If I
understand this usage, it means models that actually operate in some
specific environment instead of in an abstract space. If so, that's the
only kind of model I've ever used. I didn't know it had a name.
There must be a word for this: the discovering-you've-been-speaking-prose
effect. Maybe the word is "retronym" (analog wristwatch). First it was
just modeling. Then people started making abstract models built
completely of symbol manipulations. Then they needed a word for models
that considered a real environment. So we get "situated models", meaning
modeling the way it used to be done? I suppose I've guessed wrong.
···
-------------------------------------------------------------------
Rick Marken (20602) --
It isn't the "free" part of free enterprise that I think has gone wrong.
It's the people who use their own freedom as a way of getting control of
everyone else. What we need is a theory of economics. I don't mean a
better theory -- I mean just a theory (that works). Preferably it will
have people in it. We can't fix the system until we understand what's
going on.
--------------------------------------------------------------------
Best
Bill.
(from Penni Sibun 920602.2000)
My objection to jargon -- sometimes -- is that it makes things seem
important and new that are neither. Consider "situated" modeling.
....
First it was
just modeling. Then people started making abstract models built
completely of symbol manipulations. Then they needed a word for models
that considered a real environment. So we get "situated models", meaning
modeling the way it used to be done? I suppose I've guessed wrong.
i think you've guessed right. it's happened cause the ai weenies know
*only* about the middle state of affairs. groundbreaking work in ai
in the 70s was *all* abstract symbol manipulation. people got
everlasting fame and fortune for writing theses about things like the
``blocks world,'' a ``world'' consisting of a table and some
(typically 3) blocks. these entities and their relationships can be
expressed in propositional calculus and programs could be given plans
for moving around the blocks. (even doing this sort of thing is
provably intractable (which is what chapman's master's thesis was
about).) it seemed like a good idea at the time, i guess.
i ta'd and then took an intro ai course at two different schools.
this is the stuff that was presented as ai (in the mid-80s). people
in ai needed to be hit over the head. i realized this looks silly
from the outside, but each field has its own course of development,
and ai has been disadvantaged in that for nearly 20 years it was
possible to do brilliantly in the field w/o having any knowledge of
anything else--just by writing a snazzy program.
oh, well. i don't tell people i do ai any more (``computational
linguistics'' is descriptively adequate) because the glitz is gone and
the shallowness has become more evident. oh, and it's not gonna get
me a job, either--the world is quite saturated w/ ai types!
--penni