Toshitaka KUMANO <[log in to unmask]> wrote

> This is a proposal of extension to Ada, about
> loop_parameter_specification.
>
> Any comments ?
>
Having used real type loop counter in other languages and found them useful,
I support the idea.
I suggest that the safest basis for such a count is a decimal type, since
this minimises uncertainty about the delta and hence the loop step size.
Floating point loop counters were used in some basics but these were for
less critical purposes than a typical Ada  application.

While we are revisiting the for loop, perhaps we could think about
increasing its flexibility to match other languages, most notably by
allowing a step size to be specified: this would apply to discrete loop
counter types as well as real ones.

In Coral66 we had for loops with integer counters, such as
FOR i:=start_value STEP step_size UNTIL stop_value
and also the form
FOR i:=value1,value2,value3 etc DO
(The values were numeric expressions and I believe dynamically evaluated, so
not necessarily in ascending order)  [Caveat: memory a little vague & no
Coral manual to hand!]

..and in some basics on the old 80's micros you could have, for example,
FOR n=2*PI TO 0 STEP -2*PI/360
where n was floating point - some basics which supported integers
nevertheless insisted on loop counters being floating point.

Rather than introduce another Ada keyword, "delta" could be overloaded in
the loop construct to do the job of STEP. In the absence  of delta usage in
the loop construct, the step would default to 'Succ - which for real loop
counters would be the delta of the  decimal type or subtype of the loop
counter.

So the Ada XX for the specimen line of Coral would be
for I in start_value..stop_value delta step_size loop
e.g.
for I in 1..12 delta 2 loop
would iterate through values 1,3,5,7,9,11 while
for I in 0..12 delta 2 loop
would iterate through values 0,2,4,6,8,10,12

Meanwhile the AdaXX for that line of basic would be
   for Theta in reverse 0.0..2.0*PI delta 2.0*PI/360.0 loop

It is highly desirable that any mumeric expression for step size is an exact
multiple of the delta of the decimal loop counter type:  obviously the step
size is of the same type as the loop counter.

Should the step size be static?  If not should step size modification be
permitted within the body of the loop? Any interest in the value-list
iteration scheme as in the second Coral example?  Discuss!

The behaviour of such  constructs should be easy to simulate in Ada9X with a
generic subprogram instantiated for the loop counter type and reading as
parameters the loop start, stop and step values.

  Regards

      Justin Jackson

------------------------------
Justin Jackson
BAE SYSTEMS New Malden Surrey UK




********************************************************************
This email and any attachments are confidential to the intended
recipient and may also be privileged. If you are not the intended
recipient please delete it from your system and notify the sender.
You should not copy it or use it for any purpose nor disclose or
distribute its contents to any other person.
********************************************************************