PEN-L feedback on Java post

Louis Proyect lnp3 at
Mon Apr 9 18:07:46 MDT 2001

I wrote something along your lines, Louis, in my Class Warfare book:

Michael Cusumano's study of Japanese programming presents perhaps the most
influential case in favor of the Japanese model. Based on a sample of 20
U.S. and 11 Japanese firms, he finds that Japanese programmers produced 60
to 70 percent more lines of code. Defects per line of delivered code in the
Japanese sample remained 1/2 to 1/4 of the U.S. projects (Cusumano 1991, p.
458). So far, so good. But Cusumano's title conveys an additional message
about the Japanese system: "Japan's Software Factories: A Challenge to U.S.
Management•. Cusumano's software factories are revolutionary in the same
sense that Henry Ford's assembly line was: Both depended on the
interchangeability of parts. Of course, Japanese programs are not actually
interchangeable, but they are built up of chunks of programming components,
which are designed to be suitable for use in other programs. Cusumano
reports that Toshiba's software factory with 2,300 personnel delivered
systems in the mid-1980s containing on average nearly 50 percent reused
code (ibid, p. 218). According to Cusumano, Hitachi is the leader in this
area. In 1969, Hitachi first established a software facility labeled and
managed as a factory, calling the facility the Software Works or, more
literally, the Software Factory (Cusumano 1991, p. 161). Faced with a
shortage of skilled programmers, Hitachi settled on the model of factory
division of labor for programming (Cusumano 1991, pp. 161 and 173). By
1988, Hitachi ranked highest among Japan's computer vendors in customer
satisfaction with hardware, overall price performance, and maintenance
services for both hardware and software, as well as fixing software defects
promptly (ibid, p. 162).

--Michael Perelman:


Louis's discussion of Java reminded me of the effect of (Peter? Philip?)
Ramus's 16th century "reform" of Aristotlean dialectcs, as chronicled by
Walter Ong. In a nutshell, Ramus made dialectic more "teachable" by
reforming it into a vast hierarchy of dichotomies. The resulting knowledge
may be crap but it can be tested to make sure it's the _right_ crap and not
just any old crap.

Tom Walker (604) 947-2213

Louis Proyect
Marxism mailing list:

More information about the Marxism mailing list