Rational Inc. touts C++ = Ada for large projects

Rational Inc. touts C++ = Ada for large projects

[Gregory Aharonian]

    I have long argued that except for very large programming projects,
Ada is not the most cost effective language to use. Desktop computing,
parallel processing, AI, distributed databases, whatever technology,
there is too much available in C/C++.
    However, up until quite recently, I was willing to concede that for
large (say >100,000 LOC efforts, that Ada is the most cost-effective.
I listened to many good arguments on comp.lang.ada, read many good
articles in Ada Letters, and thought that the arguments made sense.
    That is, until I saw a paper from the company with the best Ada
CASE tools, arguing that C++ is just as cost effective for large scale
projects.  Oh, my shattered innocence, my tears in the wildnerness,
my loss of self and purpose (hey you guys don't like my nastiness, let
me try to wax poetic).

    At the 1992 OOPSLA conference, one James Walsh of Rational presented
a paper with the following abstract:

		"PRELIMINARY DEFECT DATA FROM THE ITERATIVE"
		     DEVELOPMENT OF A LARGE C++ PROGRAM

       Preliminary results from the development of Rational Rose,
       a large ( > 100 KLOC ) program written in C++, suggest
       that high product quality can be achieved during a
       telescoped development schedule through the use of an
       iterative-development methodology.  Data on defect density
       and discovery rate gathered on one phase of Rose development
       implies a low error rate in code delivered for functional
       test and supports 80-20 defect clustering, 80-20 distribution
       of noncritical versus critical defects, an association of
       defects detected during functional test with depth in the
       subsystem hierarchy, and the utility of data on rates of
       defect discovery for C++ programs.

==============================================================================

    Look, when the largest Ada contractor, IBM, for the past three years,
has publicly touted only C++ and Smalltalk, and when the best Ada compiler
and CASE company, Rational, argues that for very large projects that C++
is also very cost effective, the public perception in the non-Mandated
world (for those who care to even consider Ada) is that the language is
surrounded by too much hypocrisy.  You cannot tell your customers in the
Mandated world one story, and for the same exact problem, tell your
customers in the non-Mandated world another story.

    If I can't appeal to your ethics, consider this - it's going to be
difficult to get a job in the private sector as the DoD downsizes if you
are associated with this stuff.

Greg Aharonian
-- 
**************************************************************************
 Greg Aharonian
 Source Translation & Optimization
 P.O. Box 404, Belmont, MA 02178

Re: Rational Inc. touts C++ = Ada for large projects

[David Emery]

>    If I can't appeal to your ethics, consider this - it's going to be
>difficult to get a job in the private sector as the DoD downsizes if you
>are associated with this stuff.

This is an interesting quote...  Let me make my position on this quite
clear:  If the choice is between working in C/C++ and not working in
this business, I choose the latter, fully understanding the financial
cost.  

One of the major problems with the software 'business' right now is
captured in the notion that it's OK to hack.  That is not software
engineering, and if I can't do software engineering, I won't do
software.

>       Preliminary results from the development of Rational Rose,
>       a large ( > 100 KLOC ) program written in C++, 

Incidentally, where I work, 100k SLOC is NOT considered a large
program.  100k SLOC is a *small* program around here.

				dave

Re: Rational Inc. touts C++ = Ada for large projects

[Kent Mitchell]

Gregory Aharonian (srctran@world.std.com) wrote:

: 		"PRELIMINARY DEFECT DATA FROM THE ITERATIVE"
: 		     DEVELOPMENT OF A LARGE C++ PROGRAM

:        Preliminary results from the development of Rational Rose,
:        a large ( > 100 KLOC ) program written in C++, suggest
:        that high product quality can be achieved during a
:        telescoped development schedule through the use of an
:        iterative-development methodology.  Data on defect density
:        and discovery rate gathered on one phase of Rose development
:        implies a low error rate in code delivered for functional
:        test and supports 80-20 defect clustering, 80-20 distribution
:        of noncritical versus critical defects, an association of
:        defects detected during functional test with depth in the
:        subsystem hierarchy, and the utility of data on rates of
:        defect discovery for C++ programs.

: =============================================================================
=

:     Look, when the largest Ada contractor, IBM, for the past three years,
: has publicly touted only C++ and Smalltalk, and when the best Ada compiler
: and CASE company, Rational, argues that for very large projects that C++
: is also very cost effective, the public perception in the non-Mandated
: world (for those who care to even consider Ada) is that the language is
: surrounded by too much hypocrisy.  You cannot tell your customers in the
: Mandated world one story, and for the same exact problem, tell your
: customers in the non-Mandated world another story.

I would like to interject a little information from Rational's perspective on
this issue.  Rational is the largest vendor of Ada software and services
in the world.  We remain fully committed to Ada in both the mandated and
non-mandated world.  In fact, we continue to develop the vast majority of
our software products in Ada and project this to continue for the foreseeable
future.  We see the Ada market as strong and growing. 

We have always touted and used a software development methodology called
iterative-development in our software development and have recommended it
to our customers as a viable and cost effective methodology.  The paper
presented at OOPSLA was intended to show that iterative development applies
just as well to C++ development as it does to Ada.  This was an endorsement
of a software development methodology not a language.  We are quite pleased
with the low defect rate we achieved with this methodology in a language
we have much less experience with (C++) than our primary development
language (Ada).

Rational has listened to our customers over the years and many of them are
looking at C++ for a number of commercial and DOD/aerospace projects.  In
talking with sales prospects (many of which are in the commercial arena) we
see the market for OO technology and specifically C++ expanding rapidly.

While Rational continues to support Ada, we also feel that diversification
into more commercial markets and satisfying our customer demands for OO and
the C++ language support is essential for our success in the long term.  We
feel that the powerful capabilities that we bring to the Ada market are also
applicable to other languages and are looking closely at those markets.  In
that light, we felt it was important to implement a medium-large project in
C++ to gain perspective on the language.  We feel this experience with C++
will be vital in designing software engineering support for OO languages.

I read and sometimes agree (sometime I don't but it's a free country) with
Greg's posts WRT the mandate and software reuse.  I did think it was important
to provide some input in this issue from our perspective.

Kent Mitchell
Marketing Technical Consultant
Rational
3320 Scott Blvd. Santa Clara, CA
408-496-3600 kdm@Rational.COM