Inc (was: Real data for a change in the assignment operators and Bounded_String discussions. )

[aek@vib.usr.pu.ru (Alexander Kopilovitch)]

Robert I. Eachus wrote:

>Should Inc be predefined as a procudure?  (In addition to as an
>attribute function.)  No.  It is clearly easier to declare the version
>of Inc that I want, than to remember where it is defined, what the
>parameters are and so on.  And I certainly don't want to discuss that in
>ARG or WG9 meetings for something where the desired functionality fits
>in one line.

That's all fine, it would be sufficient to establish Standard Minimal
Requirements for Ada (2005) textbooks, and require them to explain usefulness
of Inc procedure -:)

>I don't expect this to end the discussion on these two issues, but I
>hope it helps.  The three calls in a row to Inc are much clearer than:
>
>       Total := Total + 1;
>       Single(Data(I)) := Single(Data(I)) + 1;
>       Digraphs(Previous,Data(I))) := Digraphs(Previous,Data(I)) + 1;

I think Pascal language was aware of that even before Ada emerged.

>But part of what makes it much clearer is having the definition of Inc
>right there.  And certainly any argument that
>
>       Total +:= 1;  -- or perhaps
>       Total := idem + 1; -- or whatever
>
>is "better" than:
>
>       Inc(Total);
>
>is crazy.

Certainly, Inc is much better here. But "idem" feature is not aimed for this case,
it is for more complex formulae where cant'be anything so simple and universal
like Inc.



Alexander Kopilovitch                      aek@vib.usr.pu.ru
Saint-Petersburg
Russia