I am willing to acknowledged that I was the culprit who posted the
summary of the report to this forum.  It is not necessary to refer
to the "somebody" or "someone" who said this or that. I take full
responsibility for my remarks.

That being said, I continue to believe the Ada policy should, rather
than being abandoned, be reinforced with better management.  The DoD
does not need to change its policy it needs to manage that policy more
effectively.

With regard to COTS and commercial software practices, the members of
the NRC committee may be overly optimistic about the virtue of such
products and practices.  Although the  DoD has not been a leader in
the number and variety of software products in the marketplace, and
has rarely contributed to the launching of a new software product,
it has been the leader in software quality.  And this leadership
vis a vis quality, has  been largely due to the influence of people such
as Barry Boehm and others on the NRC committee. Dr. Wasserman has
pioneered many important ideas we accept as part of good software
practice.  And Tucker completed the design of a superb version of
Ada.  I regularly quote Dr.Liskov in my object-oriented programming
classes.

Notwithstanding the emminence of the panelists, they have reached
a wrong conclusion if they are suggesting the that the DoD depart
from its "Ada is the default unless proven otherwise" policy.  The
outstanding work of Tucker Taft on Ada 95 has elevated Ada to a new
level of excellence.  This excellence is just being discovered by
DoD personnel and contractors.  Perhaps, two or threee years from
now, we could safely eliminate the policy, but this is exactly the
wrong time to do so.  The recommendations of the committee are based
on experience with Ada 83.  I believe they have failed, in spite of
their intimate knowledge of both the new Ada standard and the software
industry, to reach a correct conclusion.  Perhaps they are overly
confident in their regard for commercial software practice. Perhaps
they are excessively idealistic in their view of commerical software
practice.  I recognize that everyone on the  committee is a lot
smarter than I am, and that may be why I am having so much difficulty
understanding how they could reach the conclusions they have.

Rather than abandon the single-language policy using Ada, we should
learn to manage the policy effectively.  Once Ada is gone from the DoD, I
wonder what scapegoat will emerge to explain poor software policy
management.  I can foresee a lot of people from the COTS and commercial
sector pointing their fingers at each other as commercial software
integration fails to satisfy quality  in mission critical applications.  I
hope such failure, in the absence of Ada, will not maninfest itself as an
increase in the stock prices of body-bag manufacturers.

Richard Riehle




[log in to unmask]
AdaWorks Software Engineering
Suite 30
2555 Park Boulevard
Palo Alto, CA 94306
(415) 328-1815
FAX  328-1112