Re: Booch Components question

[Simon Wright <simon@pushface.org>]

"Randy Brukardt" <randy@rrsoftware.com> writes:

> "Simon Wright" <simon@pushface.org> wrote in message 
> news:m28vnlb9ry.fsf@pushface.org...
> ...
>> If anyone has objections/suggestions, now would be a good time to raise
>> them; thanks in advance.
>
> [I don't use your components and haven't looked at them in many years,
> but I can't resist commenting. Feel free to give this comment the
> value you think it deserves... :-)]

Quite a lot, actually, Randy: thanks for the comments.

> I'm of two minds of this (maybe even more).

Me too.

> First, if someone needs *simple* components, they are best served by
> using the Standard ones added in Ada 2005. Even if they have to use
> Ada 95, they can use the Ada 2005 components with a few small
> modifications (change the name to Ada_Containers, and replace the
> anonyomous access types with named ones). [Ada 95 use of the
> containers will limit instantiations and call-backs to library-level,
> annoying sometimes but not a major hardship.]  Using these containers
> means that they will be available on (most) future compilers and
> presumably will decrease the learning curve for future maintainers
> (most Ada programmers will be at least somewhat familar with them).

Now that would be an interesting little project! Rather similar to
[1]. But I think there wouldn't be that many customers; surely anyone
starting a new project would just go for the standard Containers.

I think the only BC that I've used that doesn't have an equivalent in
the Standard is the Bag.

> So, for alternative components to be valuable, they need facilities
> beyond the basic ones supported by Ada 2005 and even Ada 2012. This
> funny "list" container seems to be in that category. So it would seem
> to be a shame to lose it.

I think I'll move it to a new 'deprecated' section. But the real trouble
with it is that it has quite serious bugs which I don't propose to fix.

> OTOH, the name "list" seems highly confusing. "List" seems to have
> gained a fairly standard meaning in the use of containers, and I'm not
> surprised that most people look there first. Having something that
> works rather differently with that name is asking for trouble.

Well, I think it's rather a sort of LISP-y list, where the List
variables in your program are rather like cursors, with the added
feature that if there are no more cursors pointing to a part of the
list, making it unreachable, it gets reclaimed.

> OT3H :-), if people move to the Standard containers as we would hope,
> it doesn't really pay to maintain alternatives at all. So it might
> make sense to just deep-freeze the thing (don't waste time on
> enhancements or bugs unless awful), but keep it around for interested
> parties.

That would be "the gripping hand" :-)

[1] https://sourceforge.net/apps/mediawiki/gnat-math-extn/