Re Rick's complaints on the Chapman & Agre papers - I found them rather
obscure too, so I borrowed the books from Penni & then got my own
copies. I think it's pretty normal that papers are enigmatic unless
you already understand the background pretty well. I find that AI
papers tend to leave out too much of the substance of what's being done
for me to follow them, but I imagine that people with the right training
would not have this problem. Books present more of the background, &
are more detailed about the substance.
Moving on to a more specific point, Bill's arm demo (the version I
have at least) has to assume that the visual system can track the
fingertip & the target, whereas a major aspect of Sonja is an
visual system that can to some extent actually do this (though he ducks
problems of early vision).
I actually think that together, Bill and Chapman could build a robot
that could play real video games, at least if you wheeled it into the
arcade and plunked it down in the one it was supposed to play. But
I don't think either could do it alone.
And, on PCT presentation, it would be nice to have more in the way of
concrete examples of 2nd order control systems doing useful things.
I think I understand how the vertical angle control inthe arm demo
works, but that's actually the only example I know, & it would be nice
to have more. Maybe they're in the stuff I've ordered from Pat&Greg,
but still it might be a good idea to have it available online.
Avery.Andrews@anu.edu.au