Re: A primeval C compiler

Robert Harley (Robert.Harley@inria.fr)
Thu, 12 Aug 1999 16:22:51 +0200 (MET DST)


This message from comp.arch seemed so much like what Rohit writes when
he's in bit-spew mode that I just had to forward it here. Rohit, are
you writing under a pseudonym these days?

Topic: The effect of computerisation on productivity, and a mostly
valid historical comparison...

------------------------------------------------------------------------------
Ketil Z Malde <ketil@ii.uib.no> wrote:
>The point was that the PC revolution is touted as increasing
>productivity all over the board, and I suspect that this is in many
>areas a myth.

Superficially, that is probably true. But it reminds me of an
editorial piece I read in a (unremembered) magazine 10-15 years
or so ago, which discussed exactly this claim that computers
would or would not increase productivity in the work place.

Ten/fifteen years ago there was a great deal less to crow about
regarding any productivity increase! Networking was not exactly
ubiquitous, and a truly useful computer was still mini-sized
rather than micro. The article pointed out that it did indeed
seem to be that computers have caused more time to be wasted
fooling with them for little or no gain rather than leading to
increased productivity. Instead of less paperwork or less file
cabinet space, computers caused everyone and their assistant to
generate piles upon piles of evermore useless paper!

But then it went on to compare that to the evolution of the
electric motor and how it changed the processes by which we do
business in not just a decade or two, but over a century. The
suggestion was that computers will take a similar path.

Just prior to electric motors, almost all manufacturing plants
had a coal fired steam boiler/engine to produce power.
Efficient use meant a vertical building, with a belt driven
power distribution system between floors turning shafts that
distributed power horizontally on each floor. High energy tools
and processes were located closer to the belt on each floor, and
lower on the vertical system. The system was tried and true,
and well understood. It was cost effective suffer the
inefficiencies of moving material between floors of a building
designed for efficient distribution of power from the steam
plant on the bottom floor.

The introduction of electric motors initially resulted in the
removal of the steam plant and installation of a single large
electric motor to replace it. It was perhaps cleaner, but did
not result in any significant productivity increase. And of
course most new plants were designed based on past history, and
followed exactly the same plan. And it took *decades* before it
became widely understood that the vertical power distribution
system was 1) no longer needed, and 2) less productive than the
use of many smaller electric motors distributed throughout a
building designed for easier movement of material rather than
power distribution.

But even that, over a period of decades as newer buildings
encouraged different business processes, was not the real win
from electric motors. Eventually they were made smaller and
smaller, and instead of many motors located in widely
distributed points around a plant, the motors began to appear
embedded in each and every tool. And *that* was when business
(because it was easy to do, or more correctly it was hard to
avoid) dramatically changed the processes being used, and
formulated more productive processes to take advantage of the
"new" technology. It took approximately 60 years!

Now think about the move from centralized mainframe systems
to departmental mini-computers, to desktop microcomputers.
And the next step is a toaster with a micro more capable than
yesterday's departmental mini and connected to the Internet
to boot.

That will *force* even the most reluctant businesses to revamp
all of their processes. And finally, decades after computers
first were predicted to be capable of increased productivity,
the promise will finally be fulfilled.

Floyd

-- 
Floyd L. Davidson                          floyd@barrow.com
Ukpeagvik (Barrow, Alaska)
------------------------------------------------------------------------------