Re: Hijacking a Thread was RE: New Ada compiler for .NET

[Richard Riehle <richard@adaworks.com>]

Adrian Hoe wrote:

> Could the use of Ada to develop weapon systems the main hatred against
> Ada? I think this should provoke some deep thinking among CLA and I
> definitely is expecting some bashing.... Oh.. oh..

I do recall, from the late late Eighties and early Nineties, that at least

one prominent SF Bay Area company noted for its real-time systems
support, refused to publicly acknowledge any association with Ada
because its founders considered Ada the programming language
for "killing and maiming."    These guys were U.C. Berkeley grads.

That being said, many other dual-use ideas and tools originally
developed under DoD auspices have not suffered that fate, perhaps
because they were never directly advertised as originating with
the DoD.

It can be conjectured that the DoD mandate was, in part responsible
for the resistance to Ada, but not because it was designed for weapon
systems.   Rather,  major compiler vendors (RR Software and Meridian
excluded) treated Ada as a DoD feeding trough and charged such
outrageous licensing fees that few commercial companies could even
consider using it.   Others, such as Tandem Computers, created what
they called "checkbox" compilers, never intending that anyone would
ever use the compiler.  In the case of Tandem, Ada could have been
a strategic advantage, but they never interfaced their compiler with
the rest of their operating system and development environments.

The whole idea of validating a compiler turned out to be incorrectly
conceived.   Validation required nothing more than that the compiler
conform to the standard, but there was little incentive for the compiler
to be integrated into the rest of the tools, resources, and capabilities
of the targeted platform.   Such checkbox compilers became a source
of frustration for programmers, most of whom discovered they needed
to create their own libraries to interface to their expected platform.

Many people recognized this when Ada 95 came along, but that
recognition was too late to reinvigorate Ada's reputation.  Then
the DoD withdrew its support for Ada at the very point in its
development when good tools began to appear.   This was a little
like "grabbing defeat from the jaws of victory."    On the other
hand, if the DoD had continued with the mandate,  the same compiler
publishers would still be seeking to feed at the trough of government
money and continue to charge huge fees for the use of the technology.

There has to be some happy medium between completely free, a la
GNAT, and a reasonably priced compiler that comes complete with
tools, direct support for a given platform, etc., versus the feeding
frenzy inspired by any government mandate, Ada or otherwise.  It
is no accident that so many compiler publishers have abandoned
or de-escalated their support for Ada now that they do not have
a captive customer.

I believe Ada can stand on its own.  The work of ACT, along with
many others has helped to democratize the language.   However,
ACT is not the only game in town and others continue to create
serious compilers for the rigorous demands of real-time, safety-critical
software.  DDC-I and Irvine Compiler Corporation come to mind
as examples of these kinds of responsible commercial companies
who continue to be strong supporters of Ada.  I apologize if I
have overlooked anyone else, but the fact is that none of those
companies does much to promote Ada beyond their traditional
customer base.  Oh, yes, Aonix made an attempt before it
was purchased by an a bunch of ignoramouses who failed to
see the value of their Ada capability until too late.

Richard Riehle