Re: problems with not and unasserted facts

not/1 is not a built-in in gprolog. Instead, use \+ / 1. This means 
"fails to solve", the negation-as-failure predicate.
Many prologs avoid the not/1 built-in because it is not possible to 
implement it in the fully general case and having the built-in would be 
misleading. You could define a not/1 predicate as:

not(Goal) :-
	\+ call(Goal).

The message "uncaught exception: 
error(existence_error(procedure,not/1),action/2)" is gprolog's graceful 
:-) way of telling you that not/1 is an undefined procedure (predicate) 
encountered while solving an action/2 clause.
The behavior you were looking for, a missing predicate definition 
causing a simple failure, can be caused by setting the prolog flag 
'unknown' to 'fail'. Solve the following goal on the command line of 
gprolog:
	set_prolog_flag(unknown, fail).

HTH,
Lindsey Spratt

On Aug 30, 2005, at 7:24 PM, Cliff Bender wrote:

> Hi. I'm new to prolog, but am trying to learn the language to fulfill 
> a prereq for a grad school program. I'm currently using Niel C. Rowe's 
> "Artificial Intelligence Through Prolog" for reference and sample 
> programs. I've typed out the traffic program on pages 61-62:
>
>
> /*--------Rules for arrow lights--------*/
> /*A*/   action(car, stop) :- light(yellow_arrow, Direction), 
> safe_stop_possible.
> /*D,E*/ action(car, yield_and_leftturn)  :- light(yellow_arrow, left), 
> not(safe_stop_possible).
>
[snip program details]
>  given the facts:
>    safe_stop_possible.
>    light(yellow,steady).
>    light(green_arrow,left).
>
>  the query action(car,X) returns:
>
>
> X = yield_and_go ? ;
> uncaught exception: error(existence_error(procedure,not/1),action/2)
>
>
>  if I don't put safe_stop_possible in as a fact, it blows up as well.
>
>  can anyone help? this is leaving me stumped. i thought not was a 
> built-in predicate? and if a fact isn't declared/asserted, that just 
> means that querying it would return false, correct?