Conference turris::languages

38.0. "Ada subsets" by FRIDAY::HAGERMAN () Wed Oct 17 1984 16:03

Here's an interesting approach to the Ada 'subset' idea.  This is 
condensed from the October 1, 1984 issue of Electronics Week.

'Ada software gets off ground as jet flies under its control'

Irvine, Calif. -- Ada, the Defense Department's valiant attempt at a
standardized programming language, has literally gotten off the ground.
An Air Force F-15 jet fighter was successfully flown last month under the
control of an on-board digital flight computer with programs written
in Ada.  It was the first such mission-critical test for Ada software.

The DOD wants to make Ada its standard language for imbedded systems, but
development of the language has been slow because of the difficulty of
standardizing something as volatile as a software language [?].  The
software compiler being used in the Ada test has not yet been validated as
meeting the standardization requirements.

Its creator, Irvine Computer Science Corp, Irvine, Calif., has concentrated
on the compiler's performance, rather than its compliance with DOD standards.
Code generated by the compiler for the flight test had to execute
fast enough to allow the processor to make 80 system checks per second.

...

The Flight Dynamics Laboratory [part of the Air Force's Wright Aeronautical
Laboratories] conducted the flight tests jointly with McDonnell Aircraft Co.

...

There are no speed requirements in the validation suite.

...

"We expect our compiler to validate eventually," Eilers [president of
Irvine Computer] says.  "It's arguable which is easier, to focus on 
performance first or to concentrate on meeting those 2000 [validation] tests.
The irony of validation is that if you spend all your time on those
tests, you can forget about performance.  The other compilers would not
yet be capable of flying an airplane."

...

Validation, he [John Johnen, manager of software marketing for Zilog,
whose z8002 is among eight processors for which the compiler can
generate code] adds, would not be difficult to achieve, because
all of the functions that are used in the flight system have met
the validation tests.[?]

end of article.

My question is, will this language become known as ICVIAS (Irvine
Computer's Version of the International Ada Subset)?

Instead of ICVIAS this should be ICOVIAS: Irvine Computer's OWN Version
of the International Ada Subset.  (It's easier to pronounce, also.)

Andy V