Re: Desirability of C++

["Ken Garlington" <Ken.Garlington@computer.org>]

"Richard D Riehle" <laoXhai@ix.netcom.com> wrote in message
news:8eclae$afj$1@slb7.atl.mindspring.net...
> Here is the U.S. it gets pretty silly sometimes.  Last week, yet
> another senior military officer pronounced in an influential meeting,
> "Ada is dead."  Readers of this forum will undestand that it was a
> stupid thing to say, but those withing listening range will deferentially
> nod in agreement, as if this absurdity actually had any meaning

I can't speak for "senior military officers," but I know that some
Government employees lower in the hierarchy have a preference, at least for
safety-critical systems. On a very recent project, during one of the very
first briefings I gave, one of the questions was (and I believe this is a
nearly-exact phrasing), "You guys are still going to be using Ada, right?"

This has been a real problem for us, in fact, given that some of our
projects want to use DSPs to implement their safety-critical systems (which
certainly seems reasonable), but Ada doesn't have nearly as broad a
representation in that market as C, and it is quite a bit more expensive for
us to use Ada on those systems.

> Why has no one ever written a press release for ComputerWorld, or
published
> an advertisement in ComputerWorld, Government Computer Week, or other
general
> publications in the computer field announcing that their Ada compiler is
used
> for the Boeing 777 or the Channel Tunnel, or the Fokker 100, or brings you
> your nightly newscast via a communications satellite?

Well, the press releases have certainly been written. I posted one recently
on this newsgroup. It's just a question of getting them seen by people that
care.

I'm not sure that it matters anymore what Government program officers think
about software languages. Assume they wanted Ada. What would they do about
it? The acquisition initiatives prevent them from specifying requirements at
this level; even the safety requirements have been watered down
significantly (IMHO). They can indicate _outcomes_ (e.g., error rates of the
delivered software) but not the _means_ by which they're met. Couple that
with the drive to use COTS, etc. and it's tough to get excited about a
language, from their perspective. I suppose they could jawbone the
contractors (as described above), but that's about it.