Re: F22 Raptor in slashdot.org

[Richard Riehle <richard@adaworks.com>]

Robert Love wrote:

> In <MrNoSpam-4A654D.12114802082003@news-server.bigpond.net.au> Dale
> Stanbrough  wrote:
> > There's been a mention of the F22 Raptor over on slashdot.org.
> > If anyone knows about the programming that has gone into this
> > it might be worthwhile posting some follow up info.
>
> I was going to ask about this.  If I remember, the new code is in C++
> and the legacy stuff is Ada.  I don't know percentages, or if there will
> be a public report on what software is causing the reboots to be needed.

1) JSF is scheduled to be predominantly C++ and C.  Ada will be used
    for some systems.

2) Raptor uses a lot of Ada as well as other languages.

3) While Ada may be better suited to the  development of military and
     commercial avionics and other safety-critical software, the development

     is still carried out by fallible humans.   A  good tool will help, but
the
     old adage, "A fool with a tool is still a fool" continues to hold true.

4) The complexity of both JSF and Raptor exceeds anything previously
     attempted.  Large, complex projects that follow the "grand design"
     process are likely to have  a huge number of defects simply because
     of their using "grand design."

5) While I believe Ada scales up better than C++ in this kind of system,
there
     are ways that skilled developers can use a limited version of C++ to
achieve
     a modesly satisfactory result.   Unfortunately, I have little
confidence that
     those developers will limit themselves to only those aspects of C++
that
     are safe.   The temptation to do otherwise is too great.

6)  When a system written in Ada does fail, we must be honest enough to
admit
      that it was created in Ada.  All the excuses we make sound like
excuses. I
      have seen excellent projects writtten in Ada -- other Ada projects
that
     would, in the words of a mythical specials forces colonel,
     "make a billy goat puke."

7)  Finally, I do believe that, all other things being equal, a programming
team has
     a greater probability of success using Ada than C++ or other
alternatives.

Richard Riehle