Re: ADA AND THE MANDATE

[csus.edu!netcom.com!netcomsv!nitelog!michael.hagerty@decwrl.dec.com (Mic hael Hagerty)]

On 10-31-92, John Cobarruvias responded to my earlier post disseminating
what I thought to be new information with:

JC> Are we going to go through this thread again?

I don't think so...  The issue I was raising was the duplicity, evasion
and blindness of the government, specifically the Department of Defense and
Congress. After all, this is election time and political arguements are all
the rage.  If this be the subject of a former thread, I apologize to all.

{Flame_Mode=On} {Flame_Level=Low}

Following the discussion at last year's TRI-Ada, I believe the imposition
of the mandate to have been a foolish move. Congress' paper attempts to
control, influence or direct the DoD is analogous to teaching a pig to
sing, that is, it wastes your time and annoys the pig.  However, once the
mandate was put in place, a Hobson's choice was created with neither of the
two options being particularly attractive.  Either it can be continued (in
which case it should be ruthlessly enforced) or Congress can 'fess up that
it, as a body, is ineffective in determining or coercing policy (and suffer
a resultant loss of face).  I have difficulty believing that Congress is
prepared for the latter.

I believe sincerely that Ada83 could have survived and even prospered
without the mandate.  It could have been helped by efforts similar to what
AJPO and the 9X people are now doing with the GNU shop.  It is not helped
by a government edict which is observed more often in the breach.   As a
result of the mandate, DoD commands now waste a great deal of effort
attempting to figure out how to avoid Ada.  Rather like sitting around the
dinner table trying to ignore the crazy aunt upstairs who is singing at the
top of her lungs.  Since there are exactly no teeth in the mandate, it
becomes a source of embarassment only when the issue is raised in the
presence of outsiders.

There are many more effective strategies than that of making a blanket
pronouncement and then turning a blind eye to the actual state of affairs.
I recall several countries "declaring" themselves literate.  Note that
the percentage of the population who could read did not increase; it was
just that the country was now "literate".  Actual literacy increases come
through effort and perserverance (i.e., a painful process).

If we assume that the efforts of the Ada-9X committee to penetrate the
university community will be successful and that the tools/compilers
produced through the outreach are of high quality, then we shall see a
blossoming of Ada.  However, this is a pretty long stretch and I cannot
see short odds on it.

A long time back, I got a copy of the first productized Pascal compiler for
$25 and a 2400' tape.  This inexpensive entry hooked me on Pascal.  Of
course, CDC offered an Algol compiler for many thousands of dollars which
was actually worse (by any measure of robustness or generated code quality)
than the cheap Pascal compiler from Zurich.  Because of this and later
inexpensive compilers from UCSD, JRT and Borland, Pascal was propelled into
its place as a teaching language.  It also helped that Pascal is an
excellent teaching language containing most of the good structuring
features while remaining small enough to be easily taught in one semester.

C followed the same path.  When universities were given the choice between
DEC's very expensive operating systems and UNIX, which was available for
about $250, they picked the emasculated Multics from AT&T. Since the system
came complete with a compiler, albeit a relatively low-level language, C
found a lot of takers.  After all, the price was right.

Ada, unfortunately, does not have a compiler which is of comparable power
to either Turbo C++ or Turbo Pascal (next week we get Turbo Pascal 7.0 with
Objects!) at anywhere near the price of these fine products.  The compilers
are sold as either stripped educational versions or much higher priced ones
with which one can actually compile a program over 100 lines.  The obvious
reason for this is that the market is small and likely to remain so as long
as it is perceived as a government-only (and only DoD at that) market.

There is another parallel, which I know will gall the UNIX people.  I can
purchase, off-the-shelf, 8-10 different communications libraries in C for
DOS systems.  Yet there does not appear to be a UNIX equivalent to either
Turbo Async or the Greenleaf C++ Comm libraries.  Additionally, libraries
for UNIX historically are very expensive and carry hefty run-time licenses.
The DOS libraries are cheap and the run-times are royalty-free once they
are bound into an executable.  The reason for this distinction is that
there are around 80M DOS systems and several million UNIX systems.  Just
like the comparison between Ada and C market share.  To catch up requires
exponential growth which only the most deluded UNIX advocates believe
possible.

{Flame_Mode=Off}

Regards, Mikey (michael.hagerty@nitelog.com)

 * JABBER v1.3B1 #B042 * Brain: the only computer made of meat.