Re: F-22 ADA Programming

[Maciej Sobczak <see.my.homepage@gmail.com>]

> I recall from when I was actively teaching Ada that one large DoD contractor (one of the largest) decided to use C++ instead of Ada.  The reason?   We can hire C++ programmers right out of college.  This is rather strange.  Why would they think that recent  college graduates schooled in a clunky language such as C++, where they wrote mostly toy programs using features that we cannot allow (notice the long list of proscriptions for JSF), who need to be retaught how to use the language safely -- how can they think they are getting any benefit?

Good point, but to be fair you should note that it is not possible to pick fresh Ada graduates and put them right in the middle of the safety critical project, either. Do you think that JSF is long? What about "Guidance on the Use of Ada95 in the Development of High Integrity Systems" as a comparative example?
Let's try:

JSF: 141 pages, ~60p are actual rules
Ada guide: 111 pages, ~100 seem to contain actual rules

I don't claim that the above numbers are enough to justify the language choice, but certainly the picture is more complex than what you have described. I agree that fresh graduates have to be re-taught. But this also means that you need the best graduates you can find. Now, if you have a big group of C++ graduates and a very small group of Ada ones, which group is more likely (statistically) to include those best (in terms of, say, IQ or whatever other parameter) that you actually want to employ?
The statistics says that you should be looking in the bigger group and the real life seems to confirm: personally, in my country I know a couple of very good C++ programmers and I know *ZERO* good Ada programmers. Now, where should I be looking for good software engineers?

So, you really want to target the C++ graduates. Now, whether you will spend your (and theirs) costly time teaching them Ada or re-teaching them C++ is another question, but it is not surprising that managers expect the second option to be safer.


> I used to ask they, "Why would you choose a programming language that is inherently error-prone and expect an outcome that is error-free?"

Because in a safety-critical project it is not the language alone that is responsible for the error-free outcome. The whole point of re-teaching the graduates is to move out of that "inherently error-prone" zone.

-- 
Maciej Sobczak * http://www.msobczak.com * http://www.inspirel.com