NetNews Usenet Archive 1993 #3/NN_1993_3.iso/spool/comp/software/5511

Internet Message Format | 1993-01-28 | 3.8 KB

Path: sparky!uunet!olivea!spool.mu.edu!howland.reston.ans.net!usc!elroy.jpl.nasa.gov!decwrl!netcomsv!netcom.com!mcgregor
From: [e-mail] (Scott L. McGregor)
Newsgroups: comp.software-eng
Subject: Re: Why is the Software Process NOT working
Message-ID: <[e-mail]>
Date: 28 Jan 93 07:34:32 GMT
References: <[e-mail]> <[e-mail]>
Organization: Prescient Software, Inc.
Lines: 63


Scott McLoughlin's comments reminded me of one other point worth
making. A primary change that is taking place in Software is the move
from a primarily Contract based model to (shrinkwrap) product model.
Twenty years ago, nearly all software was contractual--either
contracted with outside custom software houses, or between users and
their captive MIS departments.  In some countries, most notably Japan,
this is still true.  But it has been changing. In the late 70s and
early 80s you started to see some Mainframe applications products.
Applications REALLY took off with word processing, spreadsheets and
PCs.

McLoughlin notes that it has been reported that half of the software
built for Fortune 500 and Military & Gov't sectors is not used.  It is
useful to note that this is primarily CONTRACT software.  It is worth
noting that conformance to specs that turn out to be insufficient to
ensure usefulness is not limited to SOFTWARE. Other CONTRACT
situations tend to have this occur more frequently. This is an
inherent risk in any contract situation as there is always the
possibility that the contract will be flawed but the contractor will
meet the letter of the requirements and be paid.

McLoufhlin noted that it would be absurd if such a high percentage of
AUTO FACTORIES created useless product. This is true, but it is worth
noting that few automobiles are custom (contract) made any more.  Most
are mass produced PRODUCTS.  If they miss the market they just aren't
purchased and the company fails, unlike the contract situation where
they are paid for because they are what was asked for even if what was
asked for doesn't turn out to be suitable. If we consider the PRODUCT
SOFTWARE industry, we won't see that high percentage of useless
products, because in that environment this imperils the company. (We do
still see it more than we see in autos, in part because the software
product industry is young and people are carrying over inappropriate
models from contract development experiences--as a result Version 1.0
products have a poor reputation in general, and applications software
companies have been more likely to fail than product companies in
other industries).

We need to remember that most of today's SWE theory was tested and refined
primarily in past contract situations, especially military and government
contract situations. Some of this may need adjustment for the growing
commercial shrinkwrap applications software products businesses,
because survival and success criteria for the two industries are
somewhat different. I think some of the discussions of SEI, CMM etc. here
reflect this growing awareness of the differing assumptions of
contract vs. product software development strategies.  Further
education on this topic will be especially important for Mil/Gov't
contractors undergoing "conversion" to civilian product businesses,
and for software producers in countries like Japan. Several
consultants who post here, including ourselves, are able to help
people with this sort of transition which can indeed be more difficult
than it may appear on the surface.

--
--
Scott L. McGregor        [e-mail]
President            tel: [phone]
Prescient Software, Inc.    fax: [phone]
3494 Yuba Avenue
San Jose, CA 95117-2967

Prescient Software sells Merge Ahead, the tool for Merging Text or Code and
offers consulting  & training in project management and design for usability.