The Economist says: US Military software in 8th place

[srctran@world.std.com (Gregory Aharonian)]

    The January 23rd edition of the Economist has an article on software
engineering with an interesting set of statistics.  The article discusses
the measuring of software complexity and productivity.  The metric discussed
is Albrecht's Function Points, which is based on measuring the inputs, the
outputs, inquiries, files and interfaces, each of the five with a different
weight.  Function Points are popular in the MIS world, especially with Cobol
programs, though they are now being applied to other languages with TI and
Unisys (ironic) offering commercial products.

    The article ends by reporting on results of analyzing 1000's of US and
foreign software projects by Caper Jones at Software Productivity Research
(whose software cost estimating tools are popular with the DoD), with the
following table presented:

		TABLE OF SOFTWARE ENGINEERING PRODUCTIVITY (1991)

	MIS SOFTWARE		SYSTEMS SOFTWARE	MILITARY SOFTWARE

1	America			Japan			France
2	France			America			Israel
3	Britain			Germany			South Korea
4	Canada			France			Britain
5	Switzerland		Britain			Germany
6	Germany			India			Sweden
7	Japan			Taiwan			Italy
8	Norway			South Korea		AMERICA
9	Sweden			Holland			Brazil
10	India			Sweden			Egypt

    Even accounting for a sampling problem in the survey, this is a truly
embarassing performance for the US military software community (I mean
being between Italy and Brazil only counts in lambada competitions).
    How many more outside critiques is it going to take for the DoD to
realize its software policies and initiatives are giving America a lousy
return for our tax dollars?  You might not believe us clowns, but groups like
the GAO and the Economist staff cannot be so easily ignored.  With four DoD
software reuse efforts, and at least three DoD CASE efforts (STARS, KBSA and
one from DARPA I can't find a name for), with tens of millions of dollars
being spent annually over the last five years, how did America end up in
8th place for military software productivity?  The waste of duplicative
efforts is only tolerable if at least one succeeds.  This table seems to
suggest not.

    And since most of the non-mandated world associates Ada with military
software, those in management (many of whom read the Economist) are going
to assume, rightly or wrongly, that Ada is not worth getting into, and that
Ada 9X is more of the same.  If this table was not discussed at last week's
Ada conference, it only goes to show the gross indifference of the Mandated
world to anything else.

    But hey, you see right through me and realize that I made this up just
to be harassing, that I bribed the Economist staff to publish material
detrimental to the pride of US military software development, that I go
around spreading flouride into public water systems.  Yes, yes its me.

    How about for one day we have an open day on comp.lang.ada, where
everyone can post anything without fear of losing job or contract.  If its
anything like the private email I get, it will be a blast (and truly make
the stuff Ted and I post look amateurish).


Greg Aharonian
Source Translation & Optimization

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