ted@grebyn.com (Ted Holden) (03/10/90)
I assume that denizens of the C/UNIX world are mostly remote
from the cares of DOD and that is as it should be. I assume also
that few in the C/UNIX world know, or care enough about Ada to
quickly be able to engage in effective debate with "software
engineers" such as the wolfe man... that they probably think the
future of computing will involve choices between C++, Smalltalk,
Eiffel; again, that is how it should be, but I doubt that's how it
WILL be. I believe that the choice will be between C++ and Ada
and, that if America is to have a future in computer software, that
C++ must win this battle.
I believe that continued use of Ada and the theory of
"software engineering" which travels along with it will eventually
lead to total paralysis within DOD, NASA, FAA, and other government
agencies as well as within a few big banks and large private
organizations which are unfortunate enough to entrust major
projects to it. I see Ada as a larger threat than communism at
this point in time.
Bill Wolfe (the wolfe man) has been so kind as to point out
several of the remaining problems with C language lately, and we
all know the nature and seriousness of these problems: some
feature which gets used once every five years might possibly
confuse somebody, or some dummy might actually get confused by the
deeper meaning of "break" every three years or so...
The problems with Ada, however, tend to be more grandiose than
that, e.g.:
o Major project utterly fails to perform adequately and is
cancelled, millions wasted.
o Major project many months behind schedule (e.g.
STASNFINS, space telescope, WISS etc.)
o Project cannot be done in Ada, hence gets done in
combinations of Ada and fifty other things, portability
and maintainability out the window.
o User falls asleep (or dies of old age) waiting for Ada
compiles, lit cigarette burns down building.
o Airplanes fall out of sky and kill people (Airbus).
o User's ears fall off....
Ada has more problems than can be easily counted. The two BIG
ones, however, are roughly as follows:
1: It is presently being forced to compete with C and UNIX,
hence the American software community, by and large, will
not move to Ada. This says that Ada users will continue
to pay five to twenty-five times the going rate for
everything they ever do, and reinvent every wheel, all at
taxpayers expense.
2: Grandiose delusions. Ada was meant to cover the two
extremes of computer science (embedded systems and
mainframe database applications) and everything else in
between, without allowing the poor victims to use
anything else but Ada. This makes for a gigantic,
unuseable monstrosity which NOBODY can reasonably use for
anything. C and C++ actually can handle the embedded
systems and the Database systems and everything in
between by making the language a minimal, easily
understood, and efficient core and making everything else
into libraries. If something needs fixing, you fix a
library; you don't have to be fixing or modifying the
language every third day. Meanwhile, the language itself
doesn't involve a clunk factor which prevents users from
actually doing their work.
I intend to post a number of articles in the near future
describing some of the real grief which Ada sufferers have been
experiencing for the last few years, and which hopefully will give
some of the C/C++/UNIX community a slightly better idea of how to
conduct debate with the likes of the wolfe man and his buddies.
Ted Holden
HTE