>Maybe quality is not as important than other goals in reality?
These days, I would place my bet on:
a) At the programmer level, a great many psychological processes
actually conspire against quality. For example, most programmers
are still unaware of the Myers curve, and therefore allocate their
testing efforts inappropriately. Another: the psychological tendency to
overestimate
conjunctive events makes us underestimate the need for testing
("My code is 90% failure free, your code is 90% failure free, and Sam's
code is 90% failure free -- this project must be in pretty good shape!"
No, actually the math says that this project has less than a 75% chance
of not producing failures). That's just two of an array of psychological
forces conspiring against quality in the very act of programming.
b) There is a Nash equilibrium that prevents the widespread use of honest
estimation methods (hence, DeMarco's 1980 prediction that quantitative
software estimation would be the norm by the year 2000 could not have been
more wrong). That, in turn, is a primary force in making quality
the sacrifice of choice in software projects everywhere. The Nash equilibrium
requires people to commit to the fallacious date, and the feature set,
and (usually) the budget, so the only thing left to surrender is quality.
For another example of an industry-wide Nash equilibrium, see Wall
Street investment analysts, where even after millions of dollars of fines
for their blatant and self-serving over-optimism, the WSJ (5/7/2002)
reports that analysts still insisted on a positive recommendation for
97.5% of all stocks.
Everything to do with programming is psychology, except for floating
point math :-).
----------------------------------------------------------------------
PPIG Discuss List ([EMAIL PROTECTED])
Discuss admin: http://limitlessmail.net/mailman/listinfo/discuss
Announce admin: http://limitlessmail.net/mailman/listinfo/announce
PPIG Discuss archive: http://www.mail-archive.com/discuss%40ppig.org/
