Re: What about Ada?

["Howard W. LUDWIG" <Howard_W_Ludwig@ccmail.orl.mmc.com>]

Daniel P Hudson wrote:
> 
> dewar@cs.nyu.edu (Robert Dewar) wrote:
> >Daniel said
> 
> >" Hmm, Not sure since they just developed another standard for Ada, whats
> > that like 10 now or something?"
> 
> >Daniel, I guess you are not very familiar with Ada! There have been
> >precisely TWO standards for Ada, the original Ada, now referred to as
> >Ada 83, and the current Ada standard, referred to as Ada 95. These are
> >the only two standards, they appeared in 1983 and 1995 respectively.
> >Unlike the situation with Pascal, the ANSI standard is identical to
> >the ISO standard.
> 
>  That covers ISO, now try the DoD's standards that came before ANSI/ISO
>  stuck their nose in it. Back when ACVC was responcible for determining
>  whether or not a implementation was conforming, actually I think they still
>  do this today maybe, although I doubt anyone checks to see whether it
>  has thier OK or not. There aren't actually 10, maybe 5 or 6 different
>  standards for Ada from its original implementation by the Dod up to
>  today's ISO Ada 9X' [95' isn't it?].

I believe there are several items being confused here:

First, the Ada language itself has been through three standardized versions:
	MIL-STD-1815, dated 1980-12-10;
	ANSI/MIL-STD-1815A, dated 1983-01-22 (the motivation for the nomenclature
		Ada 83), also known (with identical definition) as ISO 8652:1987 and
		FIPS 119;
	ANSI/ISO/IEC 8652:1995 (the motivation for the nomenclature Ada 95), also known
		(with identical definition) as FIPS 119-1.
Thus, in terms of broader-based standardization (such as ANSI and ISO, not just the USA
Department of Defense), there have been two versions of the language as Robert Dewar
has pointed out.  There was only one MIL-STD version prior to those; it had been
submitted to ANSI for adoption, but, for various reasons, the MIL-STD was updated with
the newer (1983) version adopted by ANSI.  (Note:  the "A" following the "1815" denotes
the first revision of the MIL-STD; with the changes in procurement policy, the DOD will
use the ANSI/ISO standard rather than MIL-STD, so the MIL-STD-1815 will be regarded as
superseded and not revised to follow ANSI/ISO.  This demonstrates that there was only
one pre-ANSI/ISO definition of Ada, namely MIL-STD-1815.)

Third, just to make sure there is no confusion on this point, there was a sequence
of requirements documents (Strawman .. Steelman) to specify what the (at that time)
new language would need to do, but none of them constituted a definition of the
language itself.

Second, in order to carry the Ada certification mark, an Ada compiler must complete
successfully and officially the Ada validation process, meaning to pass the requisite
tests in the Ada Compiler Validation Capability (ACVC) suite.  For Ada 83, the suites
were named 1.n, where n got up to 11.  For Ada 95, they are named 2.n.  Different
suite versions do NOT constitute different definitions of the language.  They simply
correspond to our growth in experience with what sorts of tests need to be run to
provide more confidence in the validation of compilers against the stable standard.

Howard W. LUDWIG

Working, but not speaking, for Lockheed Martin Electronics & Missiles Co.