Alsys, how insignificant is 18,000?

Alsys, how insignificant is 18,000?

[Gregory Aharonian]

    Recently I read some very nice promotional literature from Alsys, which
I have to admit does the best job of marketing Ada (for example, the just got
a mention in the May 1993 issue of Computer Design).

    However, in their literature, they include one information that is quite
interesting.  At one point, they state:

	"We have the largest customer base in the Ada industry"
	 with over 18,000 installations worldwide.

Great for Alsys, but what does that say about Ada?  For example, Microsoft
has sold over 25,000 development kits for its new Windows NT operating
system, C/C++ kits for a four million line C++ operating system.  Given
that the equipment and environment needed for Windows NT is as costly as
an Ada development project (minus the excessive documentation costs), and
given that Windows NT is only a small part of the C/C++ world, one can
conclude that Ada use outside the Mandated world is insignificant, a
message reinforced by the actions of Ada compiler vendors to refocus on
C++.

	Either someone qualified better take responsibility for the health
of Ada, or the Ada Mandate should be dropped.  A language that is
"obscure" and "eclipsed" should not be the foundation of national security.

Greg Aharonian
-- 
**************************************************************************
Greg Aharonian
 Source Translation & Optimiztion
 P.O. Box 404, Belmont, MA 02178

Re: Alsys, how insignificant is 18,000?

[dog.ee.lbl.gov!pasteur!agate!linus!linus.mitre.org!linus!sdl]

In article <SRCTRAN.93May13125019@world.std.com> srctran@world.std.com (Gregory
 Aharonian) writes:

>     Recently I read some very nice promotional literature from Alsys, which
> I have to admit does the best job of marketing Ada (for example, the just got
> a mention in the May 1993 issue of Computer Design).
> 
>     However, in their literature, they include one information that is quite
> interesting.  At one point, they state:
> 
> 	"We have the largest customer base in the Ada industry"
> 	 with over 18,000 installations worldwide.
> 
> Great for Alsys, but what does that say about Ada?  For example, Microsoft
> has sold over 25,000 development kits for its new Windows NT operating
> system, C/C++ kits for a four million line C++ operating system.  Given
> that the equipment and environment needed for Windows NT is as costly as
> an Ada development project (minus the excessive documentation costs), and
> given that Windows NT is only a small part of the C/C++ world, one can
> conclude that Ada use outside the Mandated world is insignificant, a
> message reinforced by the actions of Ada compiler vendors to refocus on
> C++.

I'm troubled by attempting to make engineering/technological decisions
based on popularity polls.  When I started in the software business, I
believed that software development was a skill requiring much
training, and therefore wasn't for everybody.  Then along came the IBM
and other PC's with their BASIC interpreters, and this mass delusion
started that anyone who could program anything was indeed a programmer
(or even a software engineer).  This delusion, of course, has no basis
in either logic or fact.

According to your logic, MRI scans, organ transplants, and other
modern advanced medical technologies would be "dead" because they are
clearly more expensive, and far less "popular," than first aid kits
and Band Aids.



--
Steven Litvintchouk
MITRE Corporation
202 Burlington Road
Bedford, MA  01730-1420

Fone:  (617)271-7753
ARPA:  sdl@mitre.org
UUCP:  linus!sdl

Re: Alsys, how insignificant is 18,000?

[Gregory Aharonian]

>According to your logic, MRI scans, organ transplants, and other
>modern advanced medical technologies would be "dead" because they are
>clearly more expensive, and far less "popular," than first aid kits
>and Band Aids.

   No, that's not my logic.  Take MRI scans - they can provide information
unattainable with first aid kits and Band Aids (unless your first aid kit
contains one of Star Trek's tricorders).  The cost of an equivalent number
of band aids and tongue depressors and needles and string to open up
your brain and get the same information that an MRI scan can give is orders
more expensive than an MRI scan cost (and much more deadly).

    Thus an MRI scan is more cost effective for the goal - trying to get
specific detailed information about the interior of the body.  

    Your analogy also suffers because there is a far wider differential
in capabilities between band aids and MRI, than there is between Ada and
C++ (or anything else) compilers.  People are choosing C++ technology
over Ada, with their own money (and fault in your analogy), because for
90% of their missions, C++ is more cost effective than Ada (given the
supply of programmers, tools, libraries, books that are available).

    About the only arena in which this discussion is more difficult,
large scale programming projects, for both the military and the civilian
sector, Ada may be more cost effective than C++.  However since there is
no independently validated economic models and data sets to address this
question, and since there are growing defections inside the DoD itself
to use C/C++ for very large projects, the existence of the Mandate is
questionable, and a potential threat to national security.

    So my original point still sticks, that outside the Mandated world,
where people are free to spend their own money, Ada is dead.  And until
someone can come up with some better demographic data than the stuff I
have been posting, (something maybe MITRE can pester ESD to fund them 
to find), and some better explanation of why all of the Ada compiler
companies are refocusing on C++, no one will believe any DoD claims about
Ada's "success".

Greg
-- 
**************************************************************************
Greg Aharonian
 Source Translation & Optimiztion
 P.O. Box 404, Belmont, MA 02178

Re: Alsys, how insignificant is 18,000?

[David Emery]

>People are choosing C++ technology over Ada, with their own money (and
>fault in your analogy), because for 90% of their missions, C++ is more
>cost effective than Ada (given the supply of programmers, tools,
>libraries, books that are available).

I'd like to see some studies demonstrating:
	1.  People make these decisions based on cost-effectiveness,
	    or any other rational process, and
	2.  C++ _is_ more cost-effective than Ada, C, FORTRAN, BLISS,
	    or any other language.

My belief is that choice of language is dictated more by perceived
popularity than any technical factor.

				dave

Re: Alsys, how insignificant is 18,000?

[Gregory Aharonian]

>I'd like to see some studies demonstrating:
>	1.  People make these decisions based on cost-effectiveness,
>	    or any other rational process, and
>	2.  C++ _is_ more cost-effective than Ada, C, FORTRAN, BLISS,
>	    or any other language.
>
>My belief is that choice of language is dictated more by perceived
>popularity than any technical factor.

I agree completely.  I would like to see such studies that are based
on full microeconomic models of life-cycle software development cost,
and the rational choice theory models of how people choose to do
things.

But for all other languages, most people, including myself, probably
woudln't care.  Things are done within budgets, based on past
experiences, and hopefully we learn from project to project.

It's only with Ada, whose use is forced by federal legislation, that
higher standards and justifications should be imposed.  And one of
the few studies, on which the legislation was somewhat justified, is
the original Mosemann studies, which privately are held to be a
joke, even by people from within the Mandated world.

So I agree with you, lets seriously collect the facts, before decisions
are made affecting national security.  To date, they haven't been.

Greg Aharonian
-- 
**************************************************************************
 Greg Aharonian
 Source Translation & Optimization
 P.O. Box 404, Belmont, MA 02178

Re: Alsys, how insignificant is 18,000?

[Christopher J. Henrich]

In <EMERY.93May14115044@goldfinger.mitre.org> emery@goldfinger.mitre.org (David
 Emery) writes:

>>People are choosing C++ technology over Ada, with their own money (and

>I'd like to see some studies demonstrating:
>	1.  People make these decisions based on cost-effectiveness,
>	    or any other rational process, and
>	2.  C++ _is_ more cost-effective than Ada, C, FORTRAN, BLISS,
>	    or any other language.

>My belief is that choice of language is dictated more by perceived
>popularity than any technical factor.

Quite likely.  And therefore, the Ada community has to make a lot
more noise in the marketplace than we have been.

It's clear that the marketplace does not always choose the
technically superior alternative.  Nor does it always choose
the most "cost effective" choice.  This is especially so, when the
issue of cost-effectiveness is unclear.  So Ada could well fail,
for reasons that don't look at all like "the real issue" to most
of the readers of cmp.lang.ada.

Regards,
Chris Henrich

Re: Alsys, how insignificant is 18,000?

[agate!howland.reston.ans.net!torn!utnut!utzoo!censor!geac!lethe!uunorth!e]

emery@goldfinger.mitre.org (David Emery) writes:

>>People are choosing C++ technology over Ada, with their own money (and
>>fault in your analogy), because for 90% of their missions, C++ is more
>>cost effective than Ada (given the supply of programmers, tools,
>>libraries, books that are available).

>I'd like to see some studies demonstrating:
>	1.  People make these decisions based on cost-effectiveness,
>	    or any other rational process, and
>	2.  C++ _is_ more cost-effective than Ada, C, FORTRAN, BLISS,
>	    or any other language.

>My belief is that choice of language is dictated more by perceived
>popularity than any technical factor.

I have been casting about for years to see if somebody has done a study
on what effects language structure has on programmer effectiveness and
reliability. I have yet to find anybody that knows of a single study in
that field.

Most of the arguments that I have seen to date about the cost effectiveness
of languages tend to center on the costs and availability of educational
material and development tools. Some times arguments will flow around
the long term costs of maintenance but in most cases these arguments are 
just for sticking to a single language as opposed to having a free for all.
 
I know that in the past when the engineering team that I was part of had
to move from a rigid structured language to C, we found that our productivity
and code reliability went down the toilet. It seems to me that people are
really focusing on the wrong features in a language as being important.
The real question should be how does this feature interact with human 
nature. I personally believe that language features that provide easy to
use access to bad coding style are the real bain of programming languages.
In C and C++ it is far too easy to break down to using dangerous language
features in Ada it is somewhat harder to do. But Ada should not be considered
the last word in language design. The things that apeal to me in terms of
neat language features may in reality not be the best choices for the users
of that language. In some ways a language is just a user interface for
programmers to get access to the underlying system and maybe some of the 
same types of research that goes into determining how users intract with
GUIs should be applied to programming languages.

-- 
Alvin Starr                   ||   voice: (416)513-6717
Eyepoint Inc.                 ||   fax:   (416)513-6718
alvin@eyepoint.com            ||

Re: Alsys, how insignificant is 18,000?

[Robert Kitzberger]

alvin@eyepoint.com (Alvin Starr) writes:

> The things that apeal to me in terms of
>neat language features may in reality not be the best choices for the users
>of that language. 

You make a good point.  As engineers we should be dispationate and
logical about chosing tools to solve problems.  An awareness of the
available tools, and a willingness to learn them and their apprpriate
uses is critical to maintain high productivity and quality.

But of course, we are (passionate) people as well as (dispassionate)
engineers ('engineering machines', as Kilgore Trout would put it) so
we often end up using the wrong tools for the job because of inertia,
laziness, or eagerness to play with a new toy.  I've heard, first-hand,
of an engineering department that purposefully used a new tool for
each new project, just so that they could have exposure to a multitude
of languages and play with new tools.  Ah, the fat old days of 
software development...

	.Bob.




--
------------------------------------------------------------------------
Bob Kitzberger                          Internet:   rlk@rational.com
Rational, Grass Valley, CA              CompuServe: 70743,1550
type Opinion is private;