Wesley Groleau wrote:
>To beat a dead horse, one more:  One paper was very well-written (from an
>writer-editor-publisher's point of view), had careful, thoughtful ranking
>and weighting of various criteria applied to various languages.  But it
>was still obvious that the weights of the criteria were (consciously or
>not) designed to make Ada the winner.

Wait a minute!  I think I've been insulted here, since I'm sure this
reference is to my paper "Guidelines for Choosing a Computer Language:
Support for the Visionary Organization."  (For those who haven't seen it, you
can find it at

Perhaps I'm being overly sensitive (pride of authorship and all that), but I
guess my first reaction to Wes' comment is -- well, what language do you
expect would come out on top most of the time if _all_ you are looking at is
language ratings?  On how many criteria would you expect C++ or Fortran or
Assembly to rate better than Ada?

However, let's take this a little farther.  If you _really_ examine the
paper, there is a lot more to it than just criteria for rating languages.
 Ada and other 3GLs aren't even considered for a project unless it is large
and very complex (> 50K LOC is a general guideline for this).  Even so, a
language rating is only the first step in the process of choosing an
appropriate language.  The next step is called "software engineering
considerations," and it takes into account, along with the language ratings,
development methods, development process, metrics and measurement,
application domains, software reuse, reengineering, development environment,
and education and training.  Once a language is (tentatively) chosen based on
these considerations, then a product selection process looks for appropriate
tools which support this language, the development process, and the project
requirements.  If tool support is lacking because of the language choice,
then that choice is revisited.

So I am unable to understand how this somehow harms the "cause" of Ada.  As I
said, what language would you expect to come out on top most of the time when
_just_ looking at language criteria?  However, after all software engineering
considerations have been done, Ada may or may not come out on top, and that
is as it should be.

Now, I am making no claims that my paper captures the ideal language
selection process.  It has already been reviewed and iterated on before
getting to its current state.  But I am sure it could still be improved upon.
 I welcome any suggestions for changes, additions, or deletions to/from the
process.  But could we be _constructive_ here?

Am I being too sensitive?  Was this paper a waste of time?  Is it the
approach that is bad?  The content?  What?


Dr. Patricia K. Lawlis  |      c. j. kemp systems, inc.
[log in to unmask]          | (937) 878-3303  voice and fax
[log in to unmask]      | (602) 460-7399  voice and fax