Newsgroups: cmu.cs.opinion
From: jmount+@GS6.SP.CS.CMU.EDU (John Mount)
Subject: Real software engineering
Message-ID: <>
Date: Mon, 06 Apr 92 16:04:47 -0400
Organization: School of Computer Science, Carnegie Mellon
Originator: jmount@GS6.SP.CS.CMU.EDU
Lines: 36

From this month's American Scientist:

  ... in Latvia, a 975-foot-long steel-arch bridge was being tested in
  accordance with Soviet regulations that called for strength tests of
  bridges every five years.  Engineers had ordered 14 trucks filled with
  gravel to be driven onto the bridge, and it collapsed just as the last
  truck drove on, apparently too quickly to be stopped when the danger
  signs appeared.  Ten people were killed, including the engineer in
  charge of the test.  He was standing beneath the bridge in accordance
  with "an old Russian engineering custom" that is supposed to show the
  engineer's confidence and responsibility.  (I have head of Eastern
  European engineers standing with their entire families beneath bridges
  the engineers designed.  Frank Lloyd Wright is said to have knocked
  out the falsework himself from beneath the great cantilever of
  Fallingwater when workmen refused to perform the act that they
  believed would lead to certain disaster.)

A few "old CS" customs like this and I'm sure we would all find the
time to take a quality approach to software engineering.  CS students
would fight to learn more theory ("are *really* sure halts in 23*n^3
time?") and formal verification methods ("are you *positive* it never
leaves the critical region with the semaphore in its pocket?").  You
wouldn't hear any more "SDI software could work" crap.  You would hear
things in the halls like: "sure it takes a little longer use compiler
X, but it has a much stricter type system."  Robotics instute faculty
standing under Amber legs saying "no no I'm perfeclty safe, it knows 
not to step on faculty."

Shrink-wrap fitness for purpose disclaimers are killing this field!!

--- It is kind of strange being in CS theory, given computers really do exist.
John Mount:               (412)268-6247
School of Computer Science, Carnegie Mellon University, 
5000 Forbes Av., Pittsburgh PA 15213-3890