Re: Ada Reuse libraries

Re: Ada Reuse libraries

[Gregory Aharonian]

>        Asset Source for Software Engineering Technology (ASSET).
>
>ASSET (Asset Source for Software Engineering Technology) is a software
>reuse library and reuse information exchange available to software
>developers in government, industry, and education.  ASSET is sponsored
>by ARPA's STARS (Software Technology for Adaptable, Reliable Systems)
>Program to serve as a national resource for the advancement of software
>reuse across the DoD.  The ASSET library, located in Morgantown, WV, is
>connected to INTERNET allowing world-wide access to reusable software
>assets.
>  ....[most stuff about what ASSET offers]....

    A millimicro kudo to ASSET.  After two years of funding, and 
hundreds of thousands of dollars, they finally figured out what USENET
is and broadcast an announcement on comp.lang.ada.  I heard they were
going to also list the titles to the components in their database, but
someone realized it would be real embarassing to do so.

    It's a shame they figured this out just as they overan their budgets
with extravagent and costly support services (like providing free 800
service dialup to their computers for users).  How long would that last
if they were creating ASSET with their own money?  Even quantum physicists
can't measure transitions that quick.
    It's funny, how totally unrealistic the business plans of ASSET are.
Case in point.  A few months ago, I recieved a letter from Chuck Lillie
of ASSET (with a copy cc'ed to John Foreman) soliciting contributions of
information on reusable software for ASSET's repositories (if you subtract
out the stuff they recieved from the NRL STARS Ada collection, C stuff
from the Free Software Foundation, and alot of documents, there is nothing
much left in ASSET).  
    Like most bureacratic operations, there were three pages of schemas
on how I had to describe my components for submission to ASSET.  One schema
was actually about the component, and surprise surprise, there is no field
for the cost of the component, something that wouldn't occur to a socialist
bureacrat.  I am supposed to go through tremendous effort in preparing
information for ASSET, and not be able to include information on the most
important aspect to a businessman, cost?  Since the staff at ASSET never
had to sell a component before, it wouldn't occur to them to include such
a field.  Even worse, there were other fields that software reuse business
people need to actually sell a component that were not included in ASSET's
schema.
    And from materials others have sent me describing what ASSET has,
it appears that the ASSET staff has no clue on how to locate reusable Ada
software.  As a measure, my collection of information on reusable software
not of DoD origin is larger than their entire collection.  And Rick Conn
makes these guys look silly with his volunteer effort that is orders more
effective.
    Yet meanwhile the ASSET staff and John Foreman are quoted all over the
place about how much they want to help business people who have software
reuse businesses:

"...To provide an electronic marketplace for reusable software products...."

Well guys, what you are doing helps no one.  It does not reflect the knowledge
of people who have actually run software reuse businesses, and have to close
sales to earn their money.

    Now some of you are saying, "fine Greg, we agree that ASSET is a waste
of money and the whole operation could be replaced by one college student
with a CDROM printer.  But shouldn't you try helping these guys first before
you air their dirty laundry.  At least then you would be justified in
spleening them all over the Internet".  Considering they have all of the
money, and I don't, why I should be their big brother?

    But it is a nice idea to help. I had it three years ago (near the end of
my nice phase when I was loosing all of my money).  There was a conference
at SEI on the economics of software reuse. I went there, sat on the management
session chaired by Chuck Lillie and suggested all kinds of great things
about how to make Ada reuse work.  Got a lot of compliments on my
directory (in fact, Chuck asked everyone in our group to take one of my
brochures back with them - no orders though), as well as my suggestions
on how to make software reuse work (like adding a clause to RFPs requiring
lists of reusable software).  Met John Foreman too, gave me a nice tour
of the SEI facilities - great plush rugs.
   So you would figure that when these guys got to ASSET, someone would
call and find out how I did with my own money what these guys are trying
to do with taxpayers money, especially after I phoned a few times offering
some free advice as a teaser for contract work when they started operations.

   Again, being polite and working the channels got me absolutely no
response - I guess cost-effective software reuse operations is not in
their agenda.  For a while I didn't care, since I have seen more DoD
software reuse efforts come and go when their funding was cut off, and
figured eventually Senator Byrd would be out of office and ASSET's 
funding stopped.

    But lately with this letter from Chuckie making absolutely no
business sense, with extravagant customer support services for customers
paying nothing, with fancy brochures describing consulting services that
compete directly with me and others, with no attempts to use technologies
like CDROMs to really help promote software reuse inside the DoD, all the
while seeing ASSET praised to the hilt in the press and STARS Newsletters
(though admittedly no one takes the newsletter seriously), it just got to
be too much.  I tried to be nice, and didn't even get a form letter of
rejection.

    But you might counter, "Greg they are learning, albeit slowly.  For
example, the RIGaMORTIS reuse interoperability group which ASSET helps
run has now put a cost field into their schema for components. Doesn't
that show good business sense?"
    Close, but no cigar (tobacco or candy).  For if RIGaMORTIS had sought
the advice of real software reuse businessmen, they would have found out
that a one component field for component cost is as useless as no fields
for cost.  For example, I have about 5000 technical reports in my database
that include source code listings (I just received from an Air Force unit
a nice computer aided tutorial simulating air to air combat - quite nice,
written in Ada with 3D graphics on a PC - which is a tremendous feat if
you actually try doing this).

    Now suppose someone wanted me to supply them with this package of
components.  What would the cost be?  Well it depends.  I could charge
them for the cost of the technical report itself (about $20 to $50
depending on length), plus any amount less than what it would cost them
to recreate the program from scratch (about $60,000 in this case for a
few man years of effort). I could charge them to type in the code and
make sure it is compilable - either a fixed fee or per-line-of-code
fee.  I can charge them to translate it to another language (usually the
requests are for C/C++).  I could charge them to enhance the combat
simulations algorithms by using other packages in my database.  Then
I could alter my charges depending on the size of the company. I usually
give small companies a break, since I know what it is like to stretch
a dollar, while big companies, who will be saving lots of money by using
my services, I tend to charge some amount between what it actuall costs
me and what it would cost them to do it themselves - everyone is happy
and the technology gets transferred.

   So if I were to translate this pricing policy into schema fields, 
I would probably need about two dozen fields to capture the art of
selling and transferring reusable software, and even then, I am not sure
if it would be possible to do so, because you can't factor in the
negotiations that arise when actually trying to SELL components that
are domain specific.

   But does this occur to any of the socialist bureacrats at ASSET and
RIGaMORTIS?  No, because none of them are reuse businessmen, even while
that is the fundamental nature of what they are doing.  Does it occur to
them to ask the few of us who are software reuse businessmen?  No, no,
no ( Nanette, probably their wife's names).  Does it occur to RIGaMORTIS
to post an announcement to comp.lang.ada, comp.sw.components, or
comp.software-eng soliciting advice?  No know nose.  Is there current
17 field schema any use for reuse businessmen? GNO, GNO, GNO.

     All other Ada policies aside, for which I am an amateur analyst,
ASSET is a waste of money that has no idea and no interest in the business
aspects of software reuse.  They have not approached and rebuffed not only
those few of us doing reuse as a business, but also those few reuse
efforts inside the government that have been successful for many years
(including a few inside the DoD that predate ASSET/DSRO/RAPID/MOUSE).
When anyone from ASSET talks about their promoting of the business aspects
of software reuse, they are lying, or at best naive.

     And since IBM and SAIC, under John Foreman at DARPA, are overseeing
this mess, I assume they approve of these activities.

     I do not make up this waste, I just report.  I may be obnoxious in
my words (okay I am obnoxious in my words), but isn't the waste itself,
and its tolerance, just as obnoxious and not more so?

    But credit where credit is due, so minimicrokudos to ASSET for finally
figuring out what USENET is.  Now, can we spell C - D - R - O - M ???


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

Re: Ada Reuse libraries

[MILLS,JOHN M.]

In article <SRCTRAN.93Jun18192807@world.std.com> srctran@world.std.com (Gregory
 Aharonian) writes:
 >    It's a shame they figured this out just as they overan their budgets
 >with extravagent and costly support services (like providing free 800
 >service dialup to their computers for users).  How long would that last
 >if they were creating ASSET with their own money?  Even quantum physicists
 >can't measure transitions that quick.

[ pages of self-serving bloviation eleded ]

Well Greg, you finally made my K file. Sayonara.

--jmm--

-- 
John M. Mills, SRE; Georgia Tech/GTRI/TSDL, Atlanta, GA 30332
uucp: ...!{decvax,hplabs,ncar,purdue,rutgers}!gatech!prism!jm59
Internet: john.m.mills@gtri.gatech.edu
EBENE Chocolat Noir 72% de Cacao - WEISS - 42000 St.Etienne - very fine