Re: ada software reuse

Re: ada software reuse

[news]

In article <1993Mar22.133141.9851@grebyn.com>, karl@grebyn.com (Karl A. Nyberg) writes:

*I know real reuse is hard.  Harder than I want to invest in at the moment.

Ada reuse?  Wouldn't that be kind of like reusing toilet paper?  Why
would you want to do that??  Shouldn't each individual Ada scandal be
unique???



-- 
Ted Holden
HTE

Re: ada software reuse

[crispen]

wdvs@uunet.uu.net  (Terminal Boredom) asks:

>Can anyone please explain how ADA supports the concept of software reuse,
>including how ada supports this concept.

Well, the ADA (Americans for Democratic Action) was (is?) an
organization which supplied ratings on Congressmen and Senators on
how closely they adhered to the "liberal Democratic" party line in
their voting behavior.  Thus, Teddy Kennedy could be expected to
have a higher ADA rating than Jesse Helms.

I'm sorry to report that ADA has, to my knowledge, never taken a
position on software reuse, though one could certainly say that
reuse is quite a conservative undertaking (but, since the SEI
thinks that reuse has nothing to do with maturity, one could argue
just as convincingly that it's radical ;-)

Ada, on the other hand, facilitates software reuse by such concepts
as packages, generics, and abstract typing and by the simple, powerful
fact that it's a standardized language without significant varieties.

I haven't got a clue as to what "ada" is.

Sorry for the smart-ass reply, but I suspect that your question
isn't susceptible to a simple, brief answer.  I presume you're
already aware of the Software Productivity Consortium's Synthesis
process for reuse, and the STARS Megaprogramming concept, as well
as the Software Engineering Institute's Structural Modeling work
and ARPA's Domain Specific Software Architecture project.  If not,
I'd suggest you check out some of the papers.  I've contributed in
a very minor way myself.

I contend (and I am not by the wildest stretch of the imagination
original in this) that reuse requires at least: (a) a mechanism for
capturing domain knowledge, (b) a process for extracting that
knowledge in the form of adapted components, and (c) an adaptable
architecture, specific to the problem domain, on which those components
can hang (though the architecture may well be an instantiation of
"eternal verity" architectural principles).  People knowledgeable about
reuse do not agree on all these points.

If one has the understanding of the mechanisms of Ada that can be
obtained from a couple of 40-hour courses (provided one is directed
toward software engineering principles in those courses) it will
become apparent how Ada facilitates each of those elements of
reuse.
+-------------------------------+--------------------------------------+
| Bob Crispen                   |   Who will babysit the babysitters?  |
| crispen@foxy.boeing.com       +--------------------------------------+
| (205) 461-3296                |Opinions expressed here are mine alone|
+-------------------------------+--------------------------------------+

ada software reuse

[Terminal Boredom]

Can anyone please explain how ADA supports the concept of software reuse,
including how ada supports this concept.

wdvs@uk.ac.bton.unix

Re: ada software reuse

[Gregory Aharonian]

>Can anyone please explain how ADA supports the concept of software reuse,
>including how ada supports this concept.

    Given that the largest amounts of source code being reused by language
is in the following order (demographics for the entire planet):
		Cobol
		C/C++
		Fortran
		Ada
		Pascal
		Lisp

    I argue that as long as a language has the basic features of a decent
syntax, modularity, and tools/syntax for argument checking, requirements
that most languages meet, then the language will support reuse.  And given
the success of "primitive" languages in supporting reuse, the structure of
a language itself is a marginal factor in promoting/discouraging reuse.

     As I have long argued to myself it seems, and completely contrary to
current DoD policy, the success or failure of a software reuse effort depends
greatly on social and economic factors such as acquistion regulations,
programmer incentives, software library construction and marketing, licensing,
information distribution technology utlization (like the Internet) and
other topics that make or break successful software reuse in other languages.

     For example, despite having the most extensive records of the location
of reusable defense software in the country and measures of the flows of these
programs, I cannot get any support from the DoD to perform studies of what
I know as it impacts the non-technical aspects of defense software reuse.
The DoD just doesn't care.  On the other hand, a company in Florida got DoD
funding to do a study to measure Ada reusability of source code by measuring
the indenting depth of Ada modules (an extreme technical approach whose total
syntactic basis makes it useless for the more important semantic aspects of
reuse).

    Thus with regards to Ada, the language itself does little to promote or
discourage reuse.  Unfortunately for Ada, the government policies involving
it are such that Ada reuse will never amount to much.

    And to illustrate how screwed up software reuse is in the Defense 
community, I will tell you a little story.  Recently a group of engineers
at a local defense contractor was looking for some data fusion software
for a million dollar project they were doing for the DoD.  While they had
sources inside the company, I offered to visit their facilities and make
a presentation of sources of sensor fusion software they could acquire and
reuse.  I figure I could have saved them about $40,000, since I have great
sources for this kind of software.
     Anyways, they seemed interested, and asked their managers if they
could arrange for me to make a presentation.  Their managers said NO, for
one reason - there was no charge number for them to charge spending a hour
listening to me talk about software reuse for their project.  It's bad
enough that defense regulations make it almost impossible for me to sell
reusable defense software to defense contractors, but to not even be able
to get into the front door to make a pitch because there are NO charge
numbers is lunacy.

    Meanwhile the DoD is funding four software reuse centers and two software
reuse working groups who no one evers hear from, who don't talk to anyone
but each other, and are staffed by people with no prior experience in 
running software reuse businesses.  They are so apathetic to their jobs they
haven't even contacted me to try and incorporate my extensive records of
defense software into their databases.

    I got into Ada reuse because I thought, of all of the languages available
for reuse, Ada has the most going for it, and I still believe this to be true.
Unfortunately, never has a language been assoicated with so many fucked up
policies and documented waste of tens of millions of dollars with regards to
reuse.   My company, Infotrans, Ada Libraries Limited, Karl Grebyn, and all
of the others who got burnt big time trying to Ada reuse as a business know
the reality all to well.

    A piece of advice.  Never ask a group of socialists what is inherently
a socioeconomic free market question.

Greg Aharonian
Source Translation & Optimization

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

ada software reuse

[Karl A. Nyberg]

In article <SRCTRAN.93Mar19103844@world.std.com> srctran@world.std.com (Gregory Aharonian) writes:

>My company, Infotrans, Ada Libraries Limited, Karl Grebyn, and all
>of the others who got burnt big time trying to Ada reuse as a business know
>the reality all to well.

After ten years in the Ada business, I'm glad to have achieved such name
recognition.  Perhaps in the next ten years people will stop using my
company's name as my last name.  I'm sure that people like Jean Alsys and
Ken Telesoft never had that problem. :-)

What Grebyn Corporation does can't REALLY be classed (oops, tagged? :-)) as
Ada reuse.  Our efforts have simply been redistribution.  We take the
various redistributable (i.e., not including stuff like CAMP, ACEC) major
Ada software repositories - STARS (and yes, there is some useful stuff
there), SIMTEL, AJPO and other packages (aflex/ayacc, paradise, Ada/Ed, GEF,
etc.)  and duplicate them for folks who can't pick them up via anonymous FTP
or whatever.  I'll freely admit that there's currently very little value
added, and nothing that I would construe as reuse.  Internally we use some
of this stuff, but not on a large scale or in a formal approach.

And we haven't gotten burned in the distribution business (except by people
who complain that the service we provide is somehow unethical or by the
folks in the various services who place such restrictions on redistribution
to preclude this kind of service - perhaps to attempt to ensure their
revenue?).  It hasn't done the kind of business that book publishing or
consulting has, but we haven't sunk "big time" money or effort into it.

I've "lost" more money with "Ada : Sources and Resources" and "Ada Monthly"
than I ever did with tape redistribution.  And I'm convinced enough that it
is viable as a product line (not a full-fledged business) that I'm doing
some reorganization of the various repositories and software to produce a CD
ROM.

I know real reuse is hard.  Harder than I want to invest in at the moment.

-- Karl --
-- 
Karl A. Nyberg				karl@grebyn.com
Post Office Box 497			Grebyn Corporation
Vienna, VA 22183-0497 USA		+1-703-281-2194

Re: ada software reuse

[Richard A. O'Keefe]

In article <SRCTRAN.93Mar19103844@world.std.com>, srctran@world.std.com (Gregory Aharonian) writes:
>     I argue that as long as a language has the basic features of a decent
> syntax, modularity, and tools/syntax for argument checking, requirements
> that most languages meet, then the language will support reuse.  And given
> the success of "primitive" languages in supporting reuse, the structure of
> a language itself is a marginal factor in promoting/discouraging reuse.

Basically, there are two key steps in software reuse.
    1.	FINDING some software to reuse.
    2.	ADAPTING it to your environment.
Of existing well-known languages, Ada and Standard ML are streets ahead
of the competition when it comes to step 2.

This leaves step 1.  One of the few languages to address this issue
is Common Lisp.  (See section 25.2 of Common Lisp, the Language, 2nd ed.)
By all accounts this is the hardest step, and it isn't clear that any
programming language has a special advantage here.