Re: "Forced to Use Ada"

[rjh@cs.purdue.EDU (Bob Hathaway)]

In article <4624@hubcap.UUCP>, ofut@hubcap.UUCP (A. Jeff Offutt) writes:
> From article <6125@medusa.cs.purdue.edu>, by rjh@cs.purdue.EDU (Bob Hathaway):
> > ...  I'm glad to see someone (the government) advocating modern
> > programming languages along with software engineering; the more support
> > the better.
> 
> C'mon, be careful what you say.  As a scientist/engineer, I do not want
> anybody *mandating* the use of technology without clear *technical* proof
> that that is in every case the best solution.
> 
> Whether you are an Ada fan, a Modula-II fan, a C fan or even open-minded,
> it is very clear that the scientific community is not convinced that
> Ada (or any language) is the best technology for all applications.  Or
> for any application.
^^^^^^^^^^^^^^^^^^^^^^
> -- 
> Jeff Offutt

Its not the language which is important its the facilities it provides.  
Does any of the above languages provide all of the necessary and desirable
constructs to provide well designed software and a method for validating
correct compilers?  What other language provides concurrency, dynamic
exception handling, generics, reasonable encapsulation constructs, Adts,
complete control structures, variable number of parameters with defaults,
etc; there are some languages which don't even provide the basics
for well designed software such as a reasonable type system or dynamic
memory management (ie. ForTran) or strong typechecking and choice of 
parameter modes (C), etc.  Ada was designed to standardize software and it
could replace almost any language with exceptions being rare.  For
scientific software, Ada provides Adts to model vectors, matrices, and other
mathematical objects and operators can be overloaded to provide high quality
code.  Ada's methodolgy is an excellent base for well designed software; as
a scientist/engineer what doesn't Ada provide for you?  Even if other
reasonably complete languages could be standardized and validated I wouldn't
want to learn X number of languages to read other programmers code,
the fewer languages the better.  While research languages will continue to
explore new ideas, it will be several years before another language replaces
Ada.  I can't think of any application Ada is poorly suited to with its
emphasis on well engineered software, can you justify your underlined claim
above?

Bob Hathaway
rjh@purdue.edu