RE: Ada The Best Language?

["Robert C. Leif, Ph.D." <rleif@rleif.com>]

From: Bob Leif
To: Richard Riehle et al.

Although, I believe that Ada is the best language and agree with most of the
arguments expressed by the Ada proponents in this thread, I do not know that
Ada is the best language. In fact, given the paucity of cited experimental
evidence, no one knows in the scientific or engineering sense which is the
"Best Language". Since no precise, objective means exists to quantify
software and the number of experiments to determine the quality of a major
software manufacturing tool, the programming language, is very small;
software development is neither a science nor an engineering discipline.
Presently, software language training probably belongs in the humanities as
a branch of linguistics. Parenthetically, this would probable greatly
improve the readability of the software.

Given the amount of money, human effort, and suffering involved in the
creation and use of software, this lack of scientific knowledge concerning
software development is particularly deplorable. I used the word suffering
to describe the effect of defects on the users of software. It probably is
also pertinent to those who pay for the development of software. This
miserable state is to be propounded by attempts to create new expensive
weaponry, such as a missile defense system. The institutions of higher
learning, the US National Science foundation and the US Department of
Defense share the responsibility for this situation.

A pharmaceutical company needs to invest about half a billion US dollars to
determine the safety and efficacy of a new drug. The total yearly cost of
software in the US and the rest of the developed world most likely is in the
range of 50 to 500 billion US dollars. This is 100 to 1,000 times the cost
of testing a new drug. We need much less discussion and much more data. My
opinion is that the services of experts in clinical drug trials will be
required for meaningful studies. Drug trials are not my field. However, I
know enough to state that our present software technology, at best with a
little charity, is equivalent to the results of a phase 1 clinical drug
trial. There are 3 phases of a drug trial. Phase 1 only establishes the
therapeutic level; it does not provide any acceptable evidence concerning
efficacy.

-----Original Message-----
From: comp.lang.ada-admin@ada.eu.org
[mailto:comp.lang.ada-admin@ada.eu.org]On Behalf Of Lao Xiao Hai
Sent: Saturday, July 21, 2001 11:20 AM
To: comp.lang.ada@ada.eu.org
Subject: Re: Ada The Best Language?




codesavvy wrote:

> > Does C++ offer built in and portable multitasking, protected objects,
> > real-time performance, built-in portable interrupt interfacing,
> > distributed programming?
>
> No but do you have any data that measures how much more productive an
> Ada developer is using these features?  BTW I like these features in
> Ada 95 a lot.  I doubt, however, that producitivity is increased
> significantly by having them available.

I thought Robert might respond to this since he has often dealt with this
issue so succinctly and with such clarity.   I'll give it a shot, echoing
Robert's most cogent points.

Ada is  intended to be more readable than writeable.  Consequently, simple
measures of productivity involving SLOC per man-month, etc. seldom have
a direct relationship to the lifecycle metrics of an Ada project.   For
software
with a long lifespan, the more readable the code, the more easily it is to
understand it.   Since some estimates put the ratio of source code at 20
percent
for development and 80 percent for maintenance,  this becomes important.
Moreover, that 80 percent has sometimes been anecdotally divided into its
own percentages where 80 percent of the maintenance is trying to figure out
what the code does, two percent making the changes, and eight percent doing
the testing.   These are WAG's, of course, but reflect yet another of Ada's
goals:  enhance the understandability of the code so future programmers can
be more productive during maintenance.

Another important goal of Ada is to have a language design that makes it
possible for a compiler to catch as many errors as possible as early in the
development process as possible.   This particular feature of the language
is one reason why many of us choose it over some alternative such as C++.
Some of us, not you obviously, regard Ada as a productivity tool that helps
us with quality issues early in the software lifecycle.   Productivity
without
quality is of dubious value when creating serious software.   This is not to
impugn the quality of your software.   Rather, some of us feel we are more
effective using the crutch of a tool such as Ada that keeps us from making
the kinds of mistakes we might make in a less rigorously defined language
such as C++.    Be gentle with us. Were we more intelligent, more careful,
more skillful, we might not need the kind of protection afforded by Ada.
Meanwhile, those who are better qualified to deal with the mysteries of
C++ are free to pursue their craft without the guarantees provided by Ada.
So be it.   Go forth and debug.

Richard Riehle
richard@adaworks.com
http://www.adaworks.com