Re: why Ada not used much in commercial server side development

[David Emery <emery@mitre.org>]

In general, I've found that the money invested in a quality
Ada binding (strongly typed, modular, etc) is easily returned
in avoiding/detecting errors at the point of the interface. 
This is true both due to the fact that the interface is both
between parts of the program and also different programmers
(e.g. the style arguments on this newsgroup), and also between
possibly two design styles.  

On one project, 8 weeks total, 5 programmers (40 staff-weeks), 1
staff-week spent doing the binding, we figure that we saved 2 or 3
staff-weeks worth of debugging.  As we implemented about 1/7 of the
total system for this prototype, we extrapolated another 14 
staff-weeks of savings due to the binding.  In reality, after a 
while you'd stop making mistakes, but I could project easily a
5-1 return on investment for the binding over the total development
of the project.
               dave