Re: F-22 ADA Programming

[Georg Bauhaus <bauhaus@futureapps.invalid>]

On 30.10.14 23:37, Maciej Sobczak wrote:
> I am not aware of any engineering industry where purely technical arguments are the only ones that are taken into account. It would be naive to expect that IT should be any different.

To an engineer, it is obvious that no terribly broken tools should ever
be used in engineering, including bridge building, except in an
emergency.  So obvious that engineers and mechanics just never face a
discussion like this: They do use a reasonably good soldering iron even
when there is a good candle, bellows and skillful use at hand.

Such curiosities would be admired, and requested to be understood,
but they would never replace the soldering iron.

C++, as an example of a tool different from C and used in engineering,
has got soooo much to address the shortcomings of its foundation (C).
But it keeps it, and not just as a curiosity. Why? By analogy, any engineer
would by now have considered replacing the foundation, or at least part of it.

But here in programming, we keep C as the foundation. We keep our
backwards compatibility, our pride, our demonstrable consistency of the
C standard, our group pressure, our dreams of always doing things
correctly, our stereotypes, the ubiquity of C (self-fulfilling),
etc. All this happens even in the presence of refactoring IDEs,
automatic proof systems, or static analysis (really redefining C implicitly),
This evidence would definitely allow saying Good Bye to int and *p.
So I don't believe the continued use of error prone foundations can
be explained with the help of just the pricing-related attributes of
economy.

On the other hand, the history of Ada pricing seem strikingly obvious:
The 4+ digits/year of a good tool have been in anyone's face (including
compilers for use in teaching), while the ROI would appear in an uncertain
future. How could this have been compatible with gaining market share?