Re: Socioeconomics, Ada, C++

[jls@netcom.COM (Jim Showalter)]

Ah, Ted's back. For those of you who missed it, a few months ago he posted
some nonsense about Ada and, after I pressed him on it, he admitted that
he actually knows NOTHING about Ada--he has no experience using it, he has
never read a text on it, etc. His opinions are based entirely on hearsay
and gossip. But that doesn't keep him from parading his ignorance in a
worldwide forum. So now he's back...

Anyway, I tend to ignore Ted Holden these days as a largely irrelevant
annoyance. However, his posting gives me a chance to say a few things
about Ada, C++, and the Ada mandate I've been wanting to say, so this time
I won't ignore him:

>Here's the rub.  40 Years ago or so, the government actually WAS in a
>position top dictate terms to industry in areas such as programming
>languages.  Cobol flew because the industry was in its infancy, and
>there were no better answers floating around out there in industry;  the
>banks and insurance companies bought off on it, and the rest is history.
>Ada backers obviously think nothing has ever changed since then, whereas
>in fact, a great deal has changed.  Not only is government no longer
>either the dominant player or in any position to dictate such terms, but
>there are several FAR better answers out there in common usage, the most
>obvious of which is C++.  

Well, first of all: when it comes to defense contracting, the DoD is by
definition the dominant player, and is therefore in exactly the right
position to dictate its terms. Bjarne Stroustrup once told me in a private
conversation that he is opposed to the Ada mandate because he is opposed
to anything that smacks of the "dead hand of officialdom". And yet, if
Mr. Stroustrup or Mr. Holden hire painters to paint their house and
specify that it be painted in light yellow exterior latex, is this the
"dead hand of officialdom", or merely the perogative of the consumer?
I wager that if the painters decided on their own that they didn't like
the color light yellow or latex paint and decided instead to paint the
house using lime green enamel they not only wouldn't get paid, they'd
get sued. In the realm of defense contracting, the DoD is the consumer.
The DoD is writing the checks and shelling out the money to get stuff
built. Anybody who doesn't want to use Ada has the freedom not to bid
on any of the DoD's contracts. But do, please, stop WHINING about it.

As for Mr. Holden's claim that "there are several FAR better answers out
there in common usage, the most obvious of which is C++", this betrays
his brute ignorance concerning Ada I warned you about at the start of
this post. When I pressed Mr. Holden as to the complexity of the projects
on which he had worked, the answer was that he'd written some "routines"
in C. As in, some little applications. The sorts of projects the DoD is
concerned with are on the order of a couple hundred thousand lines on up.
They involve teams of programmers, often distributed across multiple sites
and companies, at various geographic locations. They involve millions upon
millions of dollars, hundreds of thousands of staff-years, and multiple
calendar years of effort. They are typically at the cutting edge of what is
possible, and sometimes not even that--sometimes it is not even clear that
the project CAN be accomplished, since there is no existence proof. For
such projects, the claim that there are better languages to use than Ada
is contradicted even by P.J. Plauger, the convener and general secretary
of the ANSI C standards committee, who is on record having said "Above
100,000 lines of code, you probably should be writing in Ada". Ada was
designed from the ground up to scale to such ambitious and complex efforts.
This is certainly more than can be said for C (yes, large systems have
been written in C, but this is more a testament to human willpower than
to the software engineering capabilities of the language).

Consider C++ for a moment. Mr. Holden appears to believe that this is some
radical improvement in the state of language design, certainly far superior
to Ada. Yet, as one who is familiar with both languages (as opposed to
Mr. Holden), I see more similarities than differences between the two
languages. Both have strong typing, separation of specification and imp-
lementation, exceptions, and genericity, characteristics I refer to as
"software engineering oriented". Once could probably engineer large complex
systems in either language. But Ada is a far more mature language, with
industrial strength tools proven on some of the largest projects ever
attempted. And Ada is validated, which C++ is not (C++ isn't even a stable
standard yet, but even once it is there's no reason to expect it to ever
be validated). Of course, one would not expect someone who writes "routines"
to understand the value of these things, despite the fact that I personally
witnessed such issues prove the undoing of an ambitious C to C++ transition
effort at a major U.S. corporation last year.

>Reality dictates that DOD cannot afford to proceed with Ada any further
>than it has. Aside from the multitude of unbelievable problems which
>are involved with Ada itself,

Note the cute rhetorical tactic: Mr. Holden raises the specter of Ada
bogeymen but never actually commits to NAMING any of them. WHAT unbelievable
problems? Last time I checked, Ada was being used quite successfully to
engineer extremely complex systems for departments of defense throughout
the world, NASA, the FAA, and many COMMERCIAL organizations in Europe
and Japan. And Ada has not only been a smashing success on hard real-time
embedded systems, but also in MIS and scientific application domains.

>the real kicker is that we are about to
>see the entire mainstream of American computer science using C++,

This seems a rather strong statement to make, particularly considering
the ferocity of some of the anti-C++ backlash on comp.object and in
various OO journals, the continued strength of other OO languages
including Eiffel and Smalltalk, and the simple fact that 70% of all
software is written in COBOL and 15% of all software is written in
FORTRAN, sectors far more likely to migrate to Ada than to C++.

>DOD off in left field in a little toilet-bowl of its own making, paying
>ten times the going rate and taking ten times the time for everything
>they ever do;  they'll get no help other than from small-potatoes
>organizations such as Janus and/or Meridian etc. etc.  And all of this
>in an era of declining budgets.

Hmmmm. Ada was originally conceived as a cost-cutting measure by the
DoD because bitter experience proved that letting contractors choose
their own language(s) led to an unbridled proliferation of languages
and exponential increase in cost. The Ada mandate has brought this
under control, and has been proven to reduce software development costs
over the lifecycle, particularly when maintenance is taken into account.
It seems to me that in an era of declining budgets, the forces motivating
the use of Ada would tend to become STRONGER, not weaker: and this seems
borne out by the fact that Congress recently made the Ada mandate the
law, not just a convention.

As for "small-potatoes" organizations, I guess Mr. Holden has never
heard of IBM, DEC, Sun, Rational, or Borland, all of whom are quite
actively pursuing Ada. But then, his unawareness of this is just more
proof (if any was needed) of his staggering ignorance of the facts,
and of his prejudicial attitude toward Ada.
-- 
**************** JIM SHOWALTER, jls@netcom.com, (408) 243-0630 *****************
* Proven solutions to software problems. Consulting and training in all aspects*
* of software development. Management/process/methodology. Architecture/design/*
* reuse. Quality/productivity. Risk reduction. EFFECTIVE OO techniques. Ada.   *