Context module and module prefix attached to arguments [Q]

Help!
1

Hi,

Let me give a simple example to explain my question about module.

:- meta_predicate user:a(:).
user:a(X) :- writeln(X).

?- module(b).

?-  a(c).
%@ b:c
%@ true.
?-  user:a(c).
%@ user:c
%@ true.

To me a and user:a in this example should be referred to a same
predicate, and equivalent predicates should behave
equivalently. But, the above example does not fit to this principle.

This question is not new for me, but so far I have not met any clear
explanation about this question. Perhaps I am missing some basics
about module.

Thanks for making clear this in advance,

Kuniaki Mukai

2

Perhaps you are missing that a call Module:Goal does two things: it resolves Goal in Module (i.e., finds the definition there) and it sets the calling context to Module, which causes the user:c. That is how it is defined in Quintus-derived module systems. There is a predicate @/2 that allows you to specify the context explicitly. It is very rarely used though.

Bottom line: the Quintus derived module systems discourage the use of explicitly quantified goals. Import the definition and call it. The qualification is targeted at debugging and special use cases.

3

Thanks Jan,

Personally I am not satisfied with such Quintus derived way of qualifying arguments of meta predicate calls because of what I wrote. However the builtin @/2, which I am not aware of, seems to fill the gap between the Quintus default way and what is necessary for me with my pack/pac, which asserts expanded anonymous predicates into their context modules.

Thank you for clarification and letting me know @/2.

Kuniaki Mukai