Re: Question on bounded / unbounded strings

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

On 22/09/2016 13:08, J-P. Rosen wrote:
> Le 22/09/2016 à 11:53, Dmitry A. Kazakov a écrit :
>>> => Bounded_String is the way to go.
>>
>> Never. It is very difficult to find cases where
>>
>> 1. There is a hard upper bound, so hard that it would be feasible to
>> mold it into the type.
>>
>> 2. There is no cases where the upper bound implies another type. There
>> is nothing in the first name's length that makes first name different
>> from the second name. These are just unrelated things.
>>
>> The major reason for that is that constraint whatever it be is most
>> likely an implementation detail, which type difference is to reflect the
>> problem space.
>>
> ?? I don't follow you here. I may not have been explicit enough, but of
> course each type corresponds to a different instantiation of bounded
> strings, wich provide different types with unrelated lengths.

I meant that any non-tagged type can be cloned in Ada, String included. 
This is not specific to bounded strings. That a change of maximum length 
requires another type is not a feature, it is rather a design flaw. 
Clearly bounded strings of different maximum length should be subtypes 
of each other, as strings are, unless cloned.

It is thinkable but useless to consider the maximum length a part of the 
contract. In most cases it is an implementation detail that does not 
deserve new type promotion.

Just my explanation of the empiric fact that practically nobody uses them.

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