Re: deductive databases

See below:

"alex goldman" <hello_at_spamm.er> wrote in message news:2406031.I2Scl5LCZd_at_yahoo.com...
> Christopher Browne wrote:
>
>>
>> I expect he's talking about the ML notion of functor, which is the
>> type signature of a 'module.'

No, Prolog's notion of functor is just a fancy way to refer to the atom at the start of a stucture (aka compound term). E.g. in

date(14, may, 2005), 'date' is a functor.

>>
>
> And since we are competing in misunderstanding, by ML you mean machine
> learning, right? :-)
>
> No, functor is usually defined in BNF grammars for first-order logic. For
> example, in
> car(cons(X,Y), X)
> car is a predicate, cons is a functor. .

That is not correct. In Prolog, structures can be nested, therefore, both 'car' and 'cons' (which by the way do not possess any special meaning per se) are 'functors'.

 Sometimes the
> term "function" is used instead of "functor", but I don't like it, because
> it creates confusion.

You right in not liking it since the word "function" simply does not exist in the Prolog vocabulary.

"Function" is better used for predicates that possess
> certain properties (determinism)

Please read , like, a book or something on Prolog 101 before trying to mislead a lot of people. Received on Sun May 15 2005 - 18:03:32 CEST