Re: Why don't large companies use Ada?

Brian J. Zimbelman (bjz_at_innsol.com) writes:
>I think one major reason has been omitted from the discussion. A major factor
>in what language I develop in is what language the company/department already
>has invested in. Ada developers are rare. C/C++ developers are all over the
>place. Therefore, most of the time, my customers want the product in C or
>C++. The price of software is not how much it costs to develop it, but how
>much it costs to maintain it. If I can't find a developer who knows the
>language it was developed in, then I chose the wrong language no matter how
>cheap the initial development was.

I'd like to contest this. Not the conclusion in the latter part of the text, but the presumption in the first part. Sure, C programmers come 13 by the dozen, but good C++ programmers? OK, I only have the perspective the Swedish market, so I might be very wrong, but here the situation is that if you know C++ well, you have lots of jobs to choose from. Or can make a good living as a consultant.

My point is that just because you know C, doesn't mean that you have less to learn to become a good C++ programmer, than someone whose never seen C before. In fact, if you have experience in Eiffel, Simula, Smalltalk or any other O-O language, you are probably better fitted than the guy who has been hacking C for 10 years, but cannot spell to imheritance, eh I mean inheritance.

From this follows that if you have a bunch of C programmers (or Cobol programmers, an even more common phenonemom), you wish to re-educate, Ada or C++ is not much difference. It is a considerable investment in both cases.

Of course, in real life managers think it is much easier to upgrade the C programmer to a C++ programmer than an Ada one, and goes on shooting himself in the foot.

Then again, I have a feeling that everyone who say that they are using C++ are really doing it. They might use it as a C compiler with strict type checking, which undeniably is a great leap forward. But sometimes the C++ compiler is too slow, so they don't even use it...

No matter what, another poster was optimistic and hoped that Ada 9x will become popular. How much I like Ada, I fear that it is too late. Ada 83 has a few weaknesses, not O-O for instance, but still superiour to C++ if you ask me. But such considerations has never had much importance  in the software industry. Fortran, Cobol, C and now C++, that just how the story goes. (Although I must hasten to add that in comparison with the first three, C++ is an enormous improvment.)

-- 
Erland Sommarskog, sommar_at_enea.se, Stockholm
Pour qui est-ce qui vous croyez que je parle?