[mercury-users] Pred defns

[Peter Schachte]

On Wed, 25 Mar 1998, Ralph Becket wrote:
> Peter Phelan wrote on 25 Mar:

> > How come the declarative definition takes just one argument,
> > whilst there are 3 in the type declaration?
> 
> With the best will in the world, RTFM.  Have a look at

I think this is a fair question.  It *is* confusing and
counterintuitive to have a predicate declared with 3 arguments,
defined with 1 argument, and called with either 1 or 3 arguments.
It's also easy to miss the --> where the :- should have been.  And
even if you notice it, how do you know where to look in the manual.

I think it would be worth putting a note in the reference manual where
it talks about pred declarations that a predicate can be defined with
two fewer arguments than it is declared with if it is defined as a
DCG, and refer to the DCG section.

Better still, I think there should be a specialized syntax for
declaring DCG predicates, ideally some variation on the EDCG
notation.  There should also be a special syntax for calling DCG
predicates from non-DCG code; maybe something like phrase/2,3 in Prolog.
Then the current practice of declaring and calling DCG predicates with
2 extra arguments should be phased out.


-Peter Schachte			| The secret of being miserable is to have the
pets at cs.mu.OZ.AU		| leisure to bother about whether you are
http://www.cs.mu.oz.au/~pets/	| happy or not. The cure is occupation.
PGP key available on request	| -- George Bernard Shaw