COMMENT ⊗   VALID 00002 PAGES
C REC  PAGE   DESCRIPTION
C00001 00001
C00002 00002	The following disturbing  questions occurred:
C00004 ENDMK
C⊗;
The following disturbing  questions occurred:

Consider a robot climbing problem. Usually a sample(simple)
environment is given with a few  boxes etc. If an iterative
rule is present then hopefully a while-loop will be assembled say
to climb from b1 to b2 to b3(assume three boxes).Then a generalized
plan could be constructed.. to climb a stack of arbitrary height.
As usual the  straight-line program b1-b2-b3 could also be achieved.

Consider now the ssolution to the factorial function(FACT).
The analogous situation would be to ask for say (FACT X0 2)
or (FACT X0 (ADD1(ADD1 0)))(obviously some hack would have to be used
since micro-planner won't allow numbers.. but regardless..)
Then as above, hopefully, a while-loop could be given for (FACT X0 2)
and the generalizer  would create (FACT X0 N); similarly, the straight-line
version of (FACT 2) would be constructible.

But this is not what Jack does. In fact APG has no choice as to what to do
and no properties of N in (FACT x0 N) are used.. none are known.


Other note: I could not understand how Jack could use(MINUS(PLUS X Y)Y)=X
 and (DIV(PROD(X Y)Y)=X, etc in FFAC when my LISP identities didn't
work.  Answer: his don't work either.(sigh)