Re: c++ vs ada results

[netcomsv!jls@apple.com (Jim Showalter)]

smithd@software.org (Doug Smith) writes:

>As an early attempt to use generics
>for a library of reusable code, they have stood up very well.  But
>as an example of what can be accomplished with generics, they are
>slightly dated.

The C++ version benefitted from several years of 20-20 hindsight.
What would be interesting to me would be for the Ada version to
be rewritten using everything that has been learned about Ada
and C++ in those years, and _then_ comparing them.

I concede, however, that the result is still unlikely to be as
dense as the C++ version, precisely because of the lack of inheritance.
Consider, for example, what happens when you try to implement a
singly-linked list generic and a doubly-linked list generic: since
the data structures are fundamentally different, you wind up having
to write two different packages even though they are quite similar in
a variety of ways.

On the other hand, this still isn't a truly fair comparison. The
C++ version exploits both inheritance and genericity ("templates")
to achieve its density. The Ada version has only genericity at its
disposal, and so naturally comes up short (well, long--but you know
what I mean). But consider this: many of the people who extoll the
virtues of inheritance do so without taking genericity into account.
My guess is that the C++ version written with inheritance but not
templates would be about as cumbersome as the Ada version written
with genericity but not inheritance. And I also think that both
versions would be about the same size (within a small constant
attributable to syntactical differences and naming conventions) if
they were both written using inheritance and genericity. In sum,
I think both of these features are important to a language.
-- 
* Jim Showalter, software engineering consultant *
*         e-mail: jls@netcom.com                 *
*         voice : (408) 243-0630                 *
*         data  : (408) 984-5019                 *