Re: Does Ada95 beat FORTRAN?!?

[marukka@delphi.com]

Thomas Koenig <ig25@fg70.rz.uni-karlsruhe.de> writes:
 
>>I was interested in replacing a lot of old Fortran-77 code with Ada 
>>code. Does anyone feel that Ada95 is
>>better than FORTRAN?
 
I'd like to add my nickle's worth.  But first, my qualifications: 18 years
of scientific software developement for large DoD contractors. 15 years of
FORTRAN, 10 years of Ada, and way too much Mil-Std 2167A.  Enough of this.
 
Depending on what your applications are, Ada may have advantages over FORTRAN.
It also has disadvantages.  Generally, FORTRAN compilers work well on all
systems, but Ada development tends to be harder on Unix systems than on VAX/VMS.
Don't be fooled by slogans.  It is just as easy to write spaghetti code in
Ada as in (I've seen many examples personally.)  Also, when translating
code into Ada, it is often tempting to over type.  Too many user-defined types
in a strongly typed language leads to innumerable headaches.  Furthermore, if
you try to translate FORTRAN-77 code into Ada, that's just what you'll get:
 code written in Ada.  It won't be either pretty or efficient.
 
You don't mention what platforms your code resides on.  That could be an
important factor.  You also neglected to mention what kind of applications
you are running.  Anything that involves string processing, message routing,
or the like is probably better off in Ada.  If you are tracking satellites,
processing radar images, or guessing weather, FORTRAN is probably the better
choice (especially since, in the latter case, you will probably be using
a high-performance computer like a Cray-90, a CM-5, or even an IBM-3090.)
The high performance computers tend to like FORTRANs better than Ada.
 
			Mark Von Hendy
  Sr. Scientific Programmer/Analyst (but what's a job title anyway? :)