Re: Is Ada the future? [was: Is C++ the future?]

[andrewl@sydney.DIALix.oz.au (Andrew Lees)]

Robin Rowe (cpp@netcom.com) wrote:
: << In the meantime, I have a responsibility to write reliable software.  
: Ada 9X does that for me in ways that C++ folks are now only dreaming 
: about. >>

: What Ada 9X dreams am I supposed to be having? How about a 
: list of ways that Ada 9X can make more reliable software 
: than C++?

This is an interesting point.  I _know_ personally that it is possible to
write software using Ada with an error introduction rate during
design and coding something like
an order of magnitude better than C or C++.  (I have used all the languages
extensively).  However, this improvement is not seen in general, and I cannot
achieve it myself without significantly changing the intellectual tools used
to design and think about the software, and (to a lesser extent) the process
of development.  That is, Ada offers the _potential_ to get _much_ more
reliability and productivity (much, much less time spent "integrating" 
i.e. finding and fixing errors), but this is achievable only by doing things
differently.  Having thought long and hard about getting the same sorts of
improvement with C/C++, I don't think it is practicable (for various
reasons, mainly concerned with complexity in the case of C++).  So Ada
development could show a big competitive advantage over the other
contenders, but in practice it seems the difference (if any) can only
be discovered by careful analysis.

Lets face it, C++ stands a good chance of cornering the language use
market as things stand.  Ada 9X (and gnat especially!) will help to
redress the balance somewhat, but C++ has an enormous push going -
it's going to be difficult to change that.

Are there case studies around that show significantly and consistently 
lower production costs with Ada?  Or have we yet to acheive this sort
of scaled improvement?  I'm not talking about 20%, more like big multipliers.

I understand (only too well) that language is only one of the factors
in the software cost equation, but it seems that with all of the good
properties that Ada has (and now with its serious design flaws fixed),
it offers a great jumping off point for other process improvements.

: Thanks.

: Robin
:                   
:  
Regards,

Andy Lees.


: -- 
: -----
: Robin Rowe                 cpp@netcom.com  408-375-9449  Monterey, CA 
: Rowe Technology            C++ training, consulting, and users groups.