Re: Looking for PD Ada interpreter

[mfeldman@seas.gwu.edu (Michael Feldman)]

In article <1991Jun14.183916.867@lonex.radc.af.mil> vanderwerkend@lonexb.radc.af.mil (Dan Vanderwerken) writes:
>In article <31357@hydra.gatech.EDU> bw10@prism.gatech.EDU (WILLIAMS,BRYAN D.) writes:
>>I'm looking for a public-domain Ada *interpreter* that will run under
>[rest of message deleted]
>
>I keep seeing messages like this in this Ada group, and I keep seeing messages
>about how _expensive_ Ada compilers, etc. are.  A great many people out there
>are asking about PD Ada compilers/interpreters but I've seen no answers.
>
>Do any PD Ada compilers/interpreters exist?  Or are all of these people asking
>the same FAQ and not getting any answers?
>
>I don't see how any PD Ada compilers could exist.  Ada is a rather large
>language and for DoD work must be periodically certified.
                                                ^^^^^^^^^
>
As one who has complained as loudly as anyone about the high price of
compilers, I feel obliged to take the other side for a moment. Bear
with me.

First, let's get some terminology straight. The word is "validated." What
does it mean? First, that the compiler developer has tested the product
using the current ACVC (Ada Compiler Validation Capability), several
thousand programs that test conformance to the standard, and found the
compiler to be in conformance ("passed" all the relevant tests). Second,
that a DoD designee has visited the developer site and repeated the tests.
There used to be a user fee for this of a number of thousand dollars, to
defray the cost of the site visit (is the fee still in effect?)

The labor to create a validated compiler, and pay the fees for the official
validation process, makes creating a validated compiler expensive, and so
most developers feel they should collect some revenue to offset these costs. 
That is one important reason that there aren't a lot of shareware or PD Ada 
systems out there. Essentially there are none. The only approximation is the 
various Unix/VMS/DOS versions of Ada/Ed, developed by New York University with 
DoD funding. (I can supply contacts by e-mail if anyone is interested).

What good is validation? Its main virtue is a guarantee of conformance to
the standard. This means that to an extent greater than with ANY OTHER 
LANGUAGE, one can write programs that will compile and give the same
behavior under ANY compiler on ANY platform. Validation does NOT guarantee
100% portability, which given hardware differences and spots where the
standard deliberately allows implementor discretion, isn't really
achievable. 

I believe this is especially useful in the education world, where students
move from OS to OS to get experience in lots of them. Try moving a Turbo
Pascal program (even one that doesn't use PC goodies) from your PC to
Unix to see how hopeless the Pascal standard is. (The Pascal standard
is a joke.) Try moving your Turbo program to Quick Pascal, staying on
the same machine. I can tell you from personal experience that it ain't easy.
With Ada, you can REALLY move stuff around, as long as you stay away from
machine-specific goodies (graphics, say). It works, folks.

We in the Ada community expend LOTS of effort picking apart those
areas of non-portability, and sometimes lose sight of the fact that 
validation offers a far higher degree of machine independence than has
ever been achieved before. Perhaps, as C compilers mature and adhere to
the ANSI standard, programs that aren't too tricky may be as portable as
Ada programs ar But I have had a enough grief moving C programs from
SunOS to HP/UX to appreciate what validation buys us.

I still hear questions from people who used interim compilers 5 years ago
like "does compiler XYZ do tasking?" or "does compiler PQR do generics."
The answer is "if it's validated, it has to." Validation covers ALL of
the language (though some machine-specific things like the stuff in 
Chapter 13 may not be tested too thoroughly, because they are machine-
specific anyway.)

Validation does NOT buy perfection, because validation is TESTING and
we know that "testing only shows the presence of bugs, not the absence of
bugs." Validation does not buy speed, either, because correctness is not the
same as efficiency. "It's easier to make a correct program fast than a
fast program correct."

Since 1983 (my first encounter with Ada compilers), I have watched Ada 
compilers on VM, VMS, Unix, DOS, and Macintosh mature and stabilize very 
nicely over the successive versions. They are MUCH more correct, and they
are MUCH more efficient. Maturity yields correctness; competition yields
speed. Current versions of compilers typically have both.

Someone has to pay the freight for this. The prices are still higher than
for other languages on similar platforms. There's no GNU Ada (yet, though
this might very well happen...). On the other hand, if you're in an
educational institution, have you checked the prices recently? $149.00
for a DOS or Mac compiler is surely getting into the right ballpark. 

And server licenses for Unix compilers are, I think, around $1000. these days, 
which is higher than I wish it were but still kind-of affordable for a school,
considering the number of users it would have. I believe that for multiple
servers the price comes down below that. 

And I have heard (though not from primary sources) that DEC and Sun will 
provide Ada compilers essentially free to their educational hardware customers.
I believe this is also the case for the IBM mainframe compiler, if the school 
has the right arrangement with IBM. I have used this last compiler recently,
and it's quite a respectable piece of software (TeleSoft built it).
Most vendors seem to be in a "be nice to the schools" mood these days.

I hope this little discourse has at least provided some understanding of the
issues. I'll archive it and ship it by e-mail to anyone who posts the
question again.

Cheers -

Mike