|
On
While possible, the idea that university or hobby software can be better than
software developed by a multi-billion dollar corporations doesn't jump out
as a likely scenario. Interoperability is God, and failing to provide it
is a fine reason for a software project to fail! Let’s see. I have worked for multi-billion dollar
companies (Bell Labs when it was part of AT&T, General Motors and
Westinghouse). And I have worked
for Universities ( At Bell Labs I worked on software
administration methodologies for No. 5ESS.
At the time I left there were about 200 software developers working on
various components of the “switch” (call processing, OS, database
and many others). My colleagues
used to complain that we had to spend almost all our times in meetings. The language used was C and development
occurred across many machines (Vaxen) and had to be
integrated. Standards were
necessary and became a great point of contention. So did tools. Sometime someone would say, “You
know if you could just get the 10 right people in a room together you could get
this developed much more quickly.” In the meantime we were hiring masters level developers at a very high rate. The technical staff would say to the
management, “Haven’t you read ‘The Mythical Man Month’”
and the bosses would answer, “Yes, keep hiring.” I left there to go to CMU’s Robotics
Institute. This was about
1980. Brian Reid was a graduate
student who had developed the Scribe Document System alone or with a very small
team. Emacs
had been developed by one guy (rms) at MIT and James
Gosling converted it for use on Unix. Rick Rashid and a few graduate students
developed the kernel of the Mach operating system. Now, none of these projects were on the
same scale as the ESS software, where the target was a distributed processor
with extremely high demands for real-time reliability but they still seemed to
yield very impressive products. It seems to me like there are two
different kinds of software projects and they have different needs in terms of
types of resources and approaches.
It is interesting, however, that Rick Rashid
has been VP for research at Microsoft for many years. I assume he still is. Frank |
============================================================ FRIAM Applied Complexity Group listserv Meets Fridays 9a-11:30 at cafe at St. John's College lectures, archives, unsubscribe, maps at http://www.friam.org
