Question

I’m building some relatively simple functions in PROLOG that take one input and one output. For simplicity, something like

func(List, Item, [Item | List]).

Now, I’ve got code that will call several of these functions in a row and pass the result on. The issue is that I have to keep creating new variable names for all of the outputs.

someOtherFunc(List, Item) :-
   func(List, Item, Output1),
   doSomething(Output1).

The issue here is that I actually have several func and several doSomething and would really appreciate not having to bind an Output1 variable explicitly. Is there any way to achieve this?

Answer 1

I’m not sure about what you’re asking, but first of all please note that those are not functions, but predicates. This is a totally different programming paradigm. Variables are not “boxes” where you put in and out some data: they’re closer to the mathematical meaning of variable, since once you bind them to some constraints on their values it’s forever.

To go back to your question, the answer is no, you can’t avoid binding some Output1 like that. Sometimes you can put in an underscore to tell prolog you just don’t care about that value, but doing so you lose the ability to make use of that particular value. In your example you would like to do something like this (in a imperative pseudocode):

var list = ..., item = ...;
doSomething(func(list, item));

There’s no other way in prolog as far as I know, you just have to use intermediate variables as you did. The only improvement I can suggest, is to choose very carefully predicates and variables names.