Re: using `use' : what is the best practice ?

["Dmitry A. Kazakov" <mailbox@dmitry-kazakov.de>]

On Tue, 7 Jun 2011 08:32:25 -0700 (PDT), milouz wrote:

> Maybe a not so anecdotic question : when using the 'use' clause ?

Always, and packages has to be designed to be "use-friendly."

> I only write the short name (`Foo') :
> 
>    with P1; use P1;
>    with P2; use P2;
> 
>    procedure Main is
>       My_Val : Integer;
>    begin
>       Foo(My_Val);  -- P1 or P2 ?
>    ...

Why do you want to know this?

The program must be understandable without this knowledge. Surely if you
choose names like "Foo" that won't happen. What kind of useful information
"P1.Foo" carries compared to "Foo"?

Furthermore location of Foo can be viewed as an implementation detail.
Especially when Foo is a primitive operation with many bodies scattered all
around. In general, the idea of fully qualified names is incompatible with
generic programming, i.e. when some operations are defined on a class of
types with different bodies for different members of the class.

P.S. All Ada users are subdivided into use-haters and with-haters. The
former are in majority.

-- 
Regards,
Dmitry A. Kazakov
http://www.dmitry-kazakov.de