Discussion Forum—A Way with Words, a fun radio show and podcast about language

I discovered a while ago that young people think "booting up" a computer comes from "booting" it—which is correct—and that "booting" it refers, obviously, to kicking it to get it started. This is a very reasonable theory, but it's wrong. Some of us were alive when we spoke of not booting but "bootstrapping" the computer.

Computers need electronic signals in their magnetic memory in order to work. How to get those signals into memory? Back in the misty dawn of the computer age, computer technicians did it by flipping hardware switches, bit by bit. That wasn't the whole operating system, of course; they flipped just enough settings that the computer would be able to read further programming in from paper tape, which would allow them to read in a regular program from paper or maybe electronic tape. (Hard disks came later.) But as soon as you powered down the computer, all the RAM would lose its charge and hence its contents and you'd have to do it all over again.

Later they came up with ROM, Read-Only Memory; you couldn't write new contents to it, but on the other hand it wouldn't lose its contents with the power off either. Someone had to get its contents into that chip, but after that you could use it as often as you wanted to turn on your computer.

I'm not that old; by the time I got out of college and into the business, they'd come up with with EPROM, Erasable Programmable ROM. The company I worked for could take a chip to the local college, pry off the protective back and borrow a device that would irradiate the chip with strong ultra-violet light for a while, I think about half an hour. Then we could take it back and plug it into a programmer, a device that would write each byte dozens of times per second, hundreds of times per byte, to get it to hold its value. After that it acted like permanent ROM, until we erased it again. I once made a minor change to the programming on such a chip, which I suppose makes me a (really old) systems programmer.

The instructions on that chip would read in fuller loading instructions from a ROM chip or three. Those loading instructions would pull the full OS from a paper tape or even (wonder of wonders!) an 8-inch diskette. The OS would then be prepared to accept primitive commands, load programs and execute programs, etc. Basically the same thing happens today; it's just that our computers have more and faster of all the same items, barring the paper tape of course. But the computer, when you power it up, has hardwired into it the ability to read a very small piece of ROM that reads fuller load instructions from more ROM, which chooses whether to read a basic part of the OS from either the CD drive or from the 0 cylinder of your hard drive, which loads up the rest of the OS from the CD or HD....

Now, here's the point: When you're used to setting the starting instructions by flipping toggle switches, the notion of a computer that can start itself is magic. Once they figured out the basic idea, everything else followed from that. But this new process, that had the computer starting itself, was like lifting yourself by your own bootstraps.

And that's where the term "bootstrapping" came from, and why we speak of "booting up" the computer nowadays.