>Suppose, for example, that one of you had an interesting solution to
>a problem in data structures, one that could be generic, and could
>also be used by a C or C++ programmer. The article would not be
>language specific, but the code could be in Ada with clear illustration
>of how to incorporate into, via instantiation, into a program written
>in some other language.
There are 2 problems I see with this:
1) While encouraging people to consider incorporating existing Ada code into
their new applications, it is not encouraging people to use Ada to develop their
new systems, and could be considered as a stop-gap measure.
2) A C or C++ programmer who could be encouraged to use this suggestion would
have to invest in an Ada compiler in order to do it.
>Ada's hospitality toward other language environments makes it ideal for
>demonstrating how to create wrappers for legacy software that still
>works, and linking together pieces of code from various language
I assume you are referring here to development of new Ada systems interfacing to
legacy code in e.g. C?
>Another approach is to write an article that demonstrates a solution
>to a problem and include code in Java, JGNAT, and Smalltalk. This is
>a little softer strategy and avoids the impression of being the work
>of a language bigot.
This sounds like a good idea to me. Especially if you show e.g. Java, C++ and
Ada inter-operating. I'm not so sure about Smalltalk - it still seems to me to
be a bit of an "academic" language. I think you'd need to define what your
target audience is going to be before you start and I think Java and C++
programmers are probably a good place to start.
>Still another path to achieving the goal is to craft an article that
>describes a problem and its solution giving heavy appreciation for all
>the languages suited to the solution and simply include Ada among them,
>with perhaps an almost imperceptible emphasis on Ada.
This could be combined with the above so that e.g. you have a system where Java
is used for a GUI communicating with a system that requires speed and
reliability written in Ada. This would show how easy it is (presumably using
JGNAT to avoid having to delve into e.g. JNI) to provide such a system.