calling `consult` several times

Can you explain what happens when we call consult several times on the same session? Are the rules from the first call erased or not?

Let me illustrate my question by a code example:

session = pl.create()
session.consult(`p(a).`);
session.query(`p(X).`);session.answers(x=>console.log(pl.format_answer(x)), 5);
// X = a ;

Now I am creating rule q(a), and the p(a) rule is still there.:

session.consult(`q(a).`);
session.query(`p(X).`);session.answers(x=>console.log(pl.format_answer(x)), 5);
// X = a ;

But when I create the rule p(b), the rule p(a) is not there any more:

session.consult(`p(b).`);
session.query(`p(X).`);session.answers(x=>console.log(pl.format_answer(x)), 5);
// X = b ;

The question is: why when I am adding p(b) it erases p(a) but not when I am adding q(b)?

[viebel]

But now I have a different problem. The rules are now duplicated.

session = pl.create();
session.consult("p(a).");
session.consult("p(b).", {reconsult: false});
session.consult("p(b).", {reconsult: false});
session.query("p(X).");
session.answers(x=>console.log(pl.format_answer(x)), 5); // X = a ; X = b. ; X = b.

I’m not sure that what I am asking makes sense but what I’d like is to keep old rules and to add new rules only when the exact same rule doesn’t exist yet.

The context of all my questions is for the integration of Tau Prolog into Klipse as it is demonstated here.

Each code snippets is pass to a pl.consult() call. I want to allow the blog writer to spread the rules among different code snippets.

Does it make sense @jariazavalverde ?

[pmoura]

In the past, Prolog systems provided both consult/1 and reconsult/1 predicates with semantics close to what you discussed and implemented here. But the distinction was lost and several systems today implement consult/1 with reconsult/1 semantics. ISO never standardized a consult predicate, unfortunately.