TEAM-ADA Archives

Team Ada: Ada Programming Language Advocacy


Options: Use Forum View

Use Monospaced Font
Show Text Part by Default
Show All Mail Headers

Message: [<< First] [< Prev] [Next >] [Last >>]
Topic: [<< First] [< Prev] [Next >] [Last >>]
Author: [<< First] [< Prev] [Next >] [Last >>]

Print Reply
Ken Garlington <[log in to unmask]>
Reply To:
Ken Garlington <[log in to unmask]>
Wed, 13 Nov 1996 16:51:48 +0000
text/plain (60 lines)
Brad Balfour wrote:
> There is a theme running through the above three paragraphs from Ken & Sy
> that I don't understand. Both of you seem to be under the impression that
> it is important for a compiler targeted at the "commercial 8-bit
> microcontroller" market to implement the full Ada language, to be
> validated, and then to enforce restrictions on using many features (via
> Annex H).
> It would seem to me to be much simpler to just:
> a) have the compiler process only those features that you want, and
> b) ignore formal validation of the compiler
> No one else in the commercial marketplace (e.g., automobiles, washing
> machines,etc.) worries about whether the compiler for their langauge
> (assembly or C) is validated. Why should they care if their Ada compiler is
> or not?

1. I worry about validation, because I am concerned about _military_ applications
using 8-bit controllers.

2. There is an argument for commercial applications as well, that says a compiler
that doesn't have a validation certificate (even a restricted one) may be more
difficult to market because of a perception (right or wrong) that the compiler
wasn't "good enough" to validate. This perception could be particularly difficult
to overcome in safety-critical applications, or when dealing with ISO 9000.
There is also a related argument for dual-use (military and commercial) applications.

> If you, the maker of the compiler, want to ensure it's correctness, then
> all  obtain a copy of the ACVC, delete those test that don't make sense for
> your (subset) compiler, and run the test suite. You can even provide these
> results to your customer upon request.
> As long as we are not talking about the US DoD -- and we are not -- then
> validation is not a requirement at all. Let's not let this impede our
> progress into new commercial domains.

I would think that if it's that easy to perform the tests and generate the
results, then it should not be that much _more_ difficult to get an
official-looking ceritificate that says, essentially, an independent
organization agrees that you passed those selected tests, and that those
tests were selected in some reasonable manner. Let's not let a
cumbersome validation process (if it is cumbersome) prevent vendors from
having a useful marketing tool, particularly if a non-DoD organization
(NIST) is going to be doing these validations!

> Brad
> --
> Brad Balfour                            SIGAda WWW Server
> CACI, Inc.                      
> (703) 277-6767                          and also try:
> [log in to unmask]           
> address: 3930 Pender Drive * Fairfax, VA 22030

LMTAS - "Our Brand Means Quality"
For more info, see or