Self-defeating DoD software reuse policies

Self-defeating DoD software reuse policies

[Gregory Aharonian]

Self-Defeating Reuse Policies

    The November 22 issue of Defense Electronics has an interview with Ralph
Crafts, mainly on the better need to treat software in the DoD, and to get
Ada accepted.  He cites one incident that reflects stupidity at its highest.

---

DE: You've also commented that you are concerned about lack of accessibility
to the DoD software repository.  Has this improved any in your opinion?

CRAFTS:  It's worse now than ever.  They had a program in the Army known as
RAPID that has now been elevated to a DoD-wide reuse activity (DISA).  I
was contacted by a professor at a major university who wanted to get some
existing Ada code to use for teaching Ada software engineering.  He called
me about who he could contact.  I mentioned RAPID.  He said they thought of
RAPID too, but couldn't sign the agreement.  I didn't know what he was
talking about.  He said you have to sign a nondisclosure statement before
they give you access to the repository.
    This is government-funded software that's not classified.  It should be
in the public domain.  There are statutes in the law that require any
government-funded software be made available to the private sector, unless
it's classified.  I asked him to fax me a copy of the nondisclosure statement.
The first item you have to agree to is that the software you get from the
library will only be used for government purposes.  You are prohibited from
reverse engineering it, from copying it or distributing it.  It's absurd
and ludicrous.  How can you promote reuse with that type of behavior?

----------------------------------

Now I haven't paid much attention to RAPID/DISA, since I knew from the outset
that their plans were based on very faulty models of the economic costs of
the social and management aspects of running a reusable software business.
However, even I didn't guess that they would come up with such self-defeating
policy, though I kind of suspected it when they decided not to make the
software available over the Internet, either anonymous ftp-style, or by
using the comp.sources.XXX approach, or using the NETLIB/STATLIB approach
on the Internet (which if you are involved with DoD reuse and don't know
about is a good sign to quit before you are caught).

If what Ralph Crafts has observed is still policy, it reflects either a very
socialist, bureaucratic power controlling operation (the DoD has always been
uncomfortable with the near anarchy of the Internet, despite its tremendous
successes in information exchange and software reuse), or the DoD's attempt
to run an operation (DISA) based on a faulty economic model (i.e. don't let
people reverse engineer and modify modules, because you have to spend time
and money testing and validating endless variations of similar modules, and
then figure a way to help someone choose between fifteen versions of a
Fast Fourier Transform algorithm in Ada).  The DoD has neither the time,
expertise or money to deal with a practical software reuse center, and so
impose policies that will wreck any chance of the library from being useful
to the general public.

On top of which, if a software reuse center is such a viable, sustainable
economic concept, why does the DoD have to start such a business?  Why not
leave it to the private sector?  After all if the DoD's success with the
reuse center is as successful as its attempts to fund development of
Ada compiler tool sets (which people seemed to have conveniently forgotten),
then the DISA will eventually die, and this time the private sector won't
be there to save the situation, becuase everyone knows that Defense
contractors don't like to buy third-party reusable software (and I can
count the companies out of business that assumed otherwise) because of DoD
policies that the DoD refuses to change or examine.

Of course, this is related to the Don't Transfer Research Project Agency's
inability to produce a directory of all of the software that it has funded.
I recently received in the mail from Institute of New Generation Computer
Technology (ICOT - their Fifth Generation effort) a catalog of all of the
software they developed and are now making available to the public.
There are seventy one systems available - expert shells, knowledge
representation, robot design, constraint logic programming, theorem proving,
natural language analysis and translation, logic simulators, circuit
designers and CASE-based reasoning tools.  170 Megabytes of software are
available representing millions of lines of code.  Their catalog is in
Japanese and English, and the software is available over the Internet and
by tape.  They are very helpful when contacted, and place no restrictions
on what you can do with the software.

(In fact this reminds me of one DTRPA stupidity story.  A few years ago I
came across a natural language interface shell to databases funded by
DTRPA and developed by Unisys.  I called up the contact person at Unisys
to get a copy of the code to play with.  She informed me that Unisys was
allowed to put in a few percent of its own money ( >5% of total project cost),
which being proprietary could not be released, which rendered the rest of
the code unreusable. She suggested I write letters all over Unisys and DTRPA.
Imagine my tax dollars completely locked up because of this oversight.  This
is not the way to foster software reuse).

Why, in twenty years (ten of which in the Reuse Era), has the DoD been
unable to do they same thing with all of its software (and much larger
budgets), especially when some jerk in Boston has almost been able to it
out of his own wallet?

The DoD's understanding of the economic, social, aspects of reuse are
minimal, and any hopes that Ada will be more successful by the existence
of DoD reuse centers is very wishful thinking.

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

Re: Self-defeating DoD software reuse policies

[Ralph Johnson]

By and large I agree with what Greg says.  However,

srctran@world.std.com (Gregory Aharonian) writes:

>On top of which, if a software reuse center is such a viable, sustainable
>economic concept, why does the DoD have to start such a business?  Why not
>leave it to the private sector?  

Developing a reuse marketplace seems to be a chicken-and-the-egg
problem, and the argument is that it needs a jump-start.  I'm not
convinced this is true, but it is a reasonable argument.  Of course,
as Greg says, it is not clear that DOD is doing it right.

>be there to save the situation, becuase everyone knows that Defense
>contractors don't like to buy third-party reusable software (and I can
>count the companies out of business that assumed otherwise) because of DoD
>policies that the DoD refuses to change or examine.

If the DOD *really* wants to make reuse a reality, it is entirely
within its power.  It should drop cost-plus software development,
let companies make horrendous profits if they can figure out a way
to develop it efficiently, and let a few companies figure out how
to take advantage of reuse to become filthy rich.  I bet it would
take less than ten years.

>Why, in twenty years (ten of which in the Reuse Era), has the DoD been
>unable to do they same thing with all of its software (and much larger
>budgets), especially when some jerk in Boston has almost been able to it
>out of his own wallet?

This is unfair.  "Crazy genius", perhaps.  Richard Stallman is one of
the best programmers in the world.  One of the many mistakes that the
DOD makes is that writing reusable software is the same as writing any
other kind of software.  This is not true; it is harder, and it takes
more skilled people.  If you've seen the rates that Stallman charges
for consulting, you'd have a better idea of how he has managed to pull
it off.  And make no mistake, the Free Software Foundation is a great
reuse success story, in spite of the fact that the software industry
is not going the way Stallman would like it to go.

Ralph Johnson -- University of Illinois at Urnana-Champaign

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