Tuesday, July 15, 2008

I used to hate computers:

A few years after we owned a PC, and I was actually working with IBM PCs, my wife noticed something remarkable: I didn't hate them. I had been expressing my hatred of computers -- for their unreliability -- for years, all the way through the minicomputer era, from PDP-8's to every flavor of Mini I came across.

After she pointed this remarkable fact out to me, I realized what it meant. PCs in the 1980's were not great machines, but they worked ever so much more reliably than all the mini computers I'd ever dealt with. And by the way, if you're old enough to have used mini computers in the 1970's you may not understand why I found them so unreliable, but I'll explain.

In the 1970's I worked for three companies that developed software on many configurations of mini computers. We set up the desired configurations by swapping boards. There were times that I moved these foot-square boards between computers daily. Every printed circuit board was full of genuine, old-fashioned wires. (I remember the first time, years later, that I saw a printed circuit board with no wires at all, just pristine printed circuits and chips. I thought it was a miracle.) These boards slid in and out of narrow slots. The manufacturers did not expect you to move their boards around all the time, they just didn't engineer for it.

When you moved boards, they snagged wires. Or their wires snagged other boards.

There's a great irony here. Those old computers had such simple operating systems that it was child's play to configure a machine after you changed its boards. Often you did nothing at all, you just ran software that knew which boards where on the computer. But snagging those wires made the computers behave undependably. I never wanted to entrust my source code to any of them.
Post a Comment