On Monday 12 March 2007 13:27, Richard Loosemore wrote:
> Well, I have no problem now, but then it has to be the "concept" level
> that is where the modules live, because they are the Lego blocks.
>
> I thought Minsky was saying they were higher up than that, but maybe I
> was mistaken.
Minsky is an inspiration, not a blueprint. To quote from AI Lab Memo 306 ("the
Frame paper"), "I try here to bring together several of these issues by
pretending to have a unified, coherent theory." Other inspirations include
cybernetics and evolutionary psychology.
The Lego blocks model breaks down because the modules or Minsky's agents do in
fact occur at high levels of abstraction. But the high-level ones operate by
telling low-level ones what to do and by using their pre-digested
interpretations of what's going on. Internally they are just as simple as the
low-level ones.
I would be reasonably happy identifying a concept and a module as I've been
using the term. At the moment I'm working with the assumption that my modules
need to split when they try to model a non-holonomic system, but that will
depend on a LOT of experimentation I haven't done yet.
Josh
-----
This list is sponsored by AGIRI: http://www.agiri.org/email
To unsubscribe or change your options, please go to:
http://v2.listbox.com/member/?list_id=303
