Re: Ada in Iraq

["John R. Strohm" <strohm@airmail.net>]

"Russ" <18k11tm001@sneakemail.com> wrote in message
news:bebbba07.0304241712.698fbc85@posting.google.com...
> 18k11tm001@sneakemail.com (Russ) wrote in message
news:<bebbba07.0304230951.59468757@posting.google.com>...
>
> > I work for a US government lab that does extensive work in air traffic
> > management (ATM). We will have major input into the future of ATM in
> > the US. As some of you may recall, I have a problem with some of Ada's
> > syntax, but I am nevertheless convinced that Ada is fundamentally
> > solid and is the right choice for the kind of software we develop and
> > will develop in the future.
> >
> > Unfortunately, however, none of my colleagues is even willing to
> > consider using Ada. It's considered a "non-starter." C, C++, and Java
> > are the only languages considered. Ada wasn't even on the radar screen
> > until I brought it up. Also, I get the impression that professors who
> > study safety-critical systems are mostly using Java. "Real-time Java"
> > is gaining major momentum, and I fear that it could crush Ada.
> >
> > I have the privledge of working with top experts in the world on
> > "revolutionary" new ATM system architectures. I ocassionally try to
> > sell them on Ada, but my efforts are starting to become a sort of
> > joke. Some of them think I am "obsessed" with Ada, and none of them
> > seems to think the choice of language is of fundamental importance
> > anyway. It's an "implementation detail" as far as most of the managers
> > are concerned.
>
> Let me follow up on my earlier post. Here is an excerpt from an email
> I recently received from a very competent and productive software
> engineer who works down the hall from me and who has great influence
> over our choice of language:
>
> Your continued obsession with Ada for purely academic reasons is a
> seemingly naive approach to real software development.  I have not
> seen any legitimate justification for switching to Ada.  Which feature
> of Ada critical to our **** development cannot be achieved with
> C/C++/Java?  There are several reasons why Ada is not practical.
> First, there is no in-house large-scale Ada application development
> experience among the software developers or civil servants.  We can't
> wait around while 40 developers come up speed. Furthermore, i'll be
> blunt and say there was "little" in-house C++  experience when the
> **** was redesigned and we're still paying the price for that
> inexperience.  Second, the FAA does not use Ada for the rest of its
> FFP software development.  The bottomline is that in today's world, if
> you want to draw from the largest pool of talent, you better be
> programming in C, C++ or Java.  Call it inertia if you want. It
> doesn't matter really.  Ultimately, any truly critical Ada feature
> will eventually be added to C++ or Java.

I fear that your Ada proselytizing efforts, and more to the point your ATM
project, are already doomed.  The decision, at least in this guy's mind, has
already been made.

Boeing faced EXACTLY those problems on the 777 project.  In particular, the
brake controller subcontractor had NO Ada experience, and they screamed
bloody murder.  After Boeing told them "If you don't want to do this project
our way, we can certainly find another subcontractor who will", they sat
down and learned Ada.  Result: even though they had to start late, and even
though they had to train people from scratch, and even though they had to
throw away their earlier efforts, they STILL came in on schedule and under
budget.

General Dynamics faced EXACTLY those problems with F-16C/D development,
although they were using JOVIAL J73.  They had some limited experience with
JOVIAL J3B on F-16A/B, but nowhere near enough to be at critical mass.  They
also had to contract out development of the JOVIAL J73 compilers: there were
none in existence at the time the project started.