Re: Missing in Booch: range operations

[porton@ex-code.com (Victor Porton)]

In article <x7vd6pe344s.fsf@smaug.pushface.org>,
	Simon Wright <simon@pushface.org> writes:
> porton@ex-code.com (Victor Porton) writes:
> 
>> I don't understand what is "the STL/Charles view of iteration" and
> 
> I meant, the idea that two iterators designate a range within a
> container over which some algorithm is to be applied.

You cause me to wonder that you don't want accept this!

Programmers often need to deal with a range of elements of
a container. I now deal with exactly this (modifying a
HTML stream I consider parts of a HTML element as ranges in
the buffer to which I read from a file).

In any case I use this range, but now I just designate it by
natural numbers instead of iterators.

Well, e.g. many sorting algorithms deal with ranges.

Why?!!

> This modified code nearly compiles:
> 
>    with BC.Containers;
>    generic
>       with package Containers_Base is new BC.Containers (<>);
>       For_Container : in out Containers_Base.Container'Class;
>    package BC.Containers.Container_Instance is
>       Container : Containers_Base.Container'Class renames For_Container;
>       type Iterator is new Containers_Base.Iterator with private;
>       function New_Iterator return Iterator'Class;
>    private
>       type Iterator is new Containers_Base.Iterator with null record;
>    end BC.Containers.Container_Instance;
> 
>    package body BC.Containers.Container_Instance is
>       function New_Iterator return Iterator'Class is
>       begin
>          return Iterator'Class (Containers_Base.New_Iterator (Container));
>       end;
>    end BC.Containers.Container_Instance;
> 
> but fails with
> 
>    bc-containers-container_instance.ads:10:09: type must be declared
>    abstract or "RESET" overridden
> 
> and when you think about it it's hard to see how this type Iterator
> could be of any use (the concrete iterators in BCs are derived from
> BC.Containers.Iterator, not from this new one).

I don't see your point about how this scheme of derivation would 
disturb its use.

P.S. I want iterator ranges. May be I will even create modified
version of Booch at SourceForge. I think this split would not
cause any harm to the community as in any case we now only in the
process of standartization of Ada Standard Containers.