Fri, 8 Dec 2000 09:58:50 -0600
Dan Baysinger sez:
> I curious as to Team Ada's take on the "Is Ada Dead or Alive Within the
> Weapons System World?" article in the December 2000 CrossTalk (The
> Journal of Defense Software Engineering). Please, note that the first
> three (somewhat disheartening) figures were developed using data from
> The article can be viewed at:
About the only thing I find to quarrel with is that in
their discussion of compilers, they failed to mention
the special nature of gnat and gcc.
What the authors fail to say is that you and I can go
out and count C and C++, compiler vendors all day
long, but our sums are largely meaningless. Here's
(1) You've got MSVC++ and you've got gcc/g++, and
everything else is down in the noise. Alas, I've got a
favorite (Jacob Navia's LCC-Win32) down in the noise,
but wishing isn't going to change the facts.
(2) The count is confounded by the huge number of
adaptations of gcc that appear under various brand
names. I wonder how many different C compilers there
actually are these days.
(3) The reason we made our count in the first place
was to see if the language is going to stay around. But
the number of compiler vendors is no longer a valid
metric because of the special nature of gcc and gnat
and the Free Software movement/business model.
Let's take a C example to show why they're special.
Suppose Microsoft is broken up and none of the Baby
Bills wants to be in the compiler business. Suppose
that at the same time Jacob Navia and all the other people
building their own compilers go off to live in a commune
to focus on peace, love, and making a little money for a
change. Further suppose a giant recession wipes out
every last company that sells a C compiler.
Now suppose your daughter wants to write or modify a
C program. Is she out of luck? Not at all. I probably
don't even have to tell you the URL where you can go
to get gcc: you could recite it from memory.
With a common code base, the number of compiler
vendors is meaningless. gcc and its kin (g77, Objective
C, gcj) will never die until the last person who knows
how to write a compiler and is interested in the language
In short, the metric of how many companies are
selling a compiler for a particular programming
language is about as relevant to us as a count of
If gnat can get into the main gcc tree -- I gather it's
very close and that one difficulty recently has been in
the politics of determining what the main gcc tree is --
Ada will live forever.
As it stands now, I've had to put in a warning on our
company's open source repository page saying that
gnat and the main distribution of gcc don't necessarily
play well together and you need to know what you're
doing when you install both. It pained me as an Ada
fan to write that, but to remain silent about that issue
would have been irresponsible.
So if you want to know what to do in practical terms to
ensure that the language we love stays around, that's
it right there.
[log in to unmask]
Speaking for myself, not my company