Re: A farewell to Ada

[rsd@sei.cmu.edu (Richard S D'Ippolito)]

In article <14036@grebyn.com> ted@grebyn.com (Ted Holden) writes:

>From: Richard S D'Ippolito, Software Engineering Institute, Pittsburgh, PA
> 
>>>     Does Ada work any better for large scale systems?  Another article
>>>in the same Journal of Electronics Defense issue describing use of Ada on
>>>the 1,246,000 line Army AFATDS system claims that:
>>>
>>>      "Ninety percent of the software requirements were met with no major
>>>      software problems."
>>>
>>>as if this were good.  The man is claiming that he had major language-
>>>related problems with 124,600 lines of code out of 1,246,000.
> 
>>how can you expect to be taken seriously?  Do you really believe the math
>>and premises behind your statement?  Wow!

You replied to me:

>The journal should be in your library or in that of Carnegie Mellon.  I'm
>sure as hell not making any of this stuff up.  Get serious.  I'm quoting
>one of your own journals which was necessarily written in such a way as
>to paint Ada in as FAVORABLE a light as possible, and the most favorable
>thing the one gentleman could say was that he ONLY had major problems
>with 10% of his code.  Pretty sad.

(1) Your logic is faulty: The quoted statement does not contain the
information that you can support your statement with, there being
no-problem, minor-problem, major-problem and we-threw-the-requirement-out
categories.  You obviously need it to mean that 10% were met with major
problems in order to blame it on the language, because you could have
assumed that the 10% weren't met at all!

(2) You made an incorrect assumption about software:  It is unlikely that
10% of the requirements are met in 10% of the code lines.  For example, can
you identify the lines that are reliable, modular, and modifiable, to take
three requirements.

(3) You want us to think that the major problems were solely the fault of
the chosen implementation language, without giving any evidence for it.  I
wasn't aware that the process of requirements specification had reached such
a degree of perfection that all software requirements are correct, clear,
consise, non-redundant, non-conflicting, operable, and implementable.
Perhaps you could point me to the method.


Now, to take your points separately:

>The journal should be in your library or in that of Carnegie Mellon.  

Why _should_ it be?  Do these libraries have obligations to carry it?

>I'm sure as hell not making any of this stuff up.  

Beg pardon -- who accused you of that?

>Get serious.

Perhaps you mean to imply here that anybody who doesn't accept your premises
and conclusions without evidence is just doing it to irritate you.

>I'm quoting one of your own journals...

Neither I nor the SEI publish, own, nor control any such trade journals.
Surely you knew that, but let hyperbole get in the way of reason.

>...which was necessarily written in such a way as to paint Ada in as
>FAVORABLE a light as possible,...

Why did they wait so long?  C'mon, Ted, just express your opinions on
articles without resorting to psychological diagnosis of the writers and
editors.

>...and the most favorable thing the one gentleman could say was that he
ONLY had major problems with 10% of his code.

If you are going to emphasize a word, at least place it properly before you
make it stand out.  Do mean "...he had ONLY major problems with...", or do
you mean "...he had major problems with ONLY 10%..."?  If the latter, see
point (3) above.

>Pretty sad.

We agree totally here on the conclusion, but probabaly not on the antecedent.
There is room for serious discussion on the design of Ada; please give your
comments to the 9X committee.  Please consider attending the next TRI-Ada
conference and picking up a copy of the preceedings from the one just held
here in Pittsburgh.


Rich
-- 
When you can measure what you are speaking about, and express it in numbers,
you know something about it.
Lord Kelvin						      rsd@sei.cmu.edu
-----------------------------------------------------------------------------