How far are modules implemented?

[ghost]

Since when using the sandbox, I see a use_module/1 directive, I
thought most of the concepts of the ISO module standard
are implemented. So I tried a module that is created on the fly:

[ghost]

I just adopted naively what works in the SWI console. But I am
not sure whether it also works in the SWI web sandbox. But in
the SWI console the result would be:

[jariazavalverde]

Sorry for the delay. Currently, the use_module directive only can be used to load libraries of Tau Prolog. User-defined modules can't be used. In the next days I'll start working on the rest of directives.

[ghost]

Would it be possible in the future to have an operator (:)/2
also for static foreign function calls. And maybe even an
operator (::)/2 for dynamic foreign function calls.

I already find these predicates here:

http://tau-prolog.org/documentation/prolog/js/apply/3

http://tau-prolog.org/documentation/prolog/js/apply/4

What would prevent Tau Prolog from insted of having names
and lists, to use compounds as in the ISO module standard.
So instead of for example:

apply(foo, [a, b], X).
apply(Y, foo, [a,b] X).

One would for example do:

JSMain:foo(a,b,X).
Y::foo(a,b,X).

The operator (:)/2 is from the ISO module standard. The
operator (:;)/2 is found in Logtalk for example, but it can
be also bootstrapped Pythonesk from (:)/2.

[ghost]

The JSMain is a module name I made up for a static call.
But if the foreign function is known in the current context,
one could drop this module name, and simply write:

foo(a,b,X).

Even if its a Java Script call, is not specially registered
in the Prolog context, only visible from the Prolog context,
then an automatic synthetic Prolog module could be

created and cached.

[gilbert]

Having ffi using : would be fantastic. It would enable Tau to better integrate with the JS ecosystem. I'm very interested in seeing and helping this happen.