🚀 go-pugleaf

mcovingt@aisun3.ai.uga.edu (Michael Covington) writes:

>In article <Bwr5A0.2Kr@sunfish.usd.edu> btoben@crappie.usd.edu (B. Toben) writes:
>>			drives(mike,X) :- X <> "camaro".
>>
>>I get an error if I state a goal like     drives(mike,Z)    because 
>>Turbo Prolog says it has nothing to compare against "camaro".  I understand
>>why I get the error, but my question is, is there a way around it, using that
>>general logic?

>No. You must first instantiate X to a value before you can perform the not-
>equal test.  Otherwise, when you state the goal "drives(mike,Z)" you are
>saying "Give me everything in the universe that is not a Camaro"!

Well, this does suggest one solution:

drives(mike,X) :-
	car(X),
	X <> camaro.

car(camaro).
car(falcon).
car(sigma).
  ....

Note that this makes the drives predicate more correct, as well,
since it will fail for queries like ?- drives(mike, teddybear) -
which, presumably, you don't want to happen.

Now, a question to any typing mavens out there: It seems
to me that if you have a typed predicate, such
as drive(person_type, car_type) where the types are regular
types it should be possible to re-write any clause  of the
form 

drive(X, Y) :- Body. 

to be 

drive(X, Y) :- person_type(X), car_type(Y), Body.

where person_type(X) and car_type(Y) are the declarative
deifnitions of the regular types, written in Prolog.
This should kick any problems with an uninstantiated negation
in the head and remove any implicit moding caused by a negation?
Comments?

					- Dug