|
|
By: Paul Rogers
You never forget your first time. Mine was with an IBM 1620 at California State College in Long Beach.
There aren't many of these antiques left. It takes a lot of dedication.
Front left is an IBM 1311 disk drive, the first drive with a removable "disk pack" of 6 14" platters. Capacity: 2 million characters (not bytes, characters. I'll explain later.)
Going counter clockwise, I don't recognize the next device. We didn't have one of those. It appears to be some sort of posting accounting machine, an electric typewriter with a wide carriage for accounting forms.
Next is the 1620 CPU. It came in two models, most obviously identified by the typewriter used as the console I/O device, in three memory sizes: 20,000, 40,000 or 60,000 decimal digits of core memory. Not "K", not bytes. This was a decimal machine, not binary. Memory was organized in 6 bit "positions", 4 bits for a decimal digit, a "check" bit, and a "flag" bit. Core memory of magnetized ferrite beads was inherently non-volatile! It did not have a fixed word length. It did not have addressable registers, all operands were in core memory. (Add two 100 digit numbers? Sure, one instruction.) The design was called "CADET" (can't add, doesn't even try). Arithmetic was done by table look-up. The Binary Coded Decimal Interchange Code, BCD, used by external devices, the Model 026 keypunch for example, only defined one set of upper case letters, 10 digits, some punctuation, and a few special characters. (Oh, so THAT's why these old timers keep shouting in their e-Mail!) It's amazing how little you need to make a computer if you're ingenious enough.
Next is a Model 26 keypunch machine.
And finally in the front row is a Model 1443 line printer. The characters were on a comb-like bar that went back and forth in front of the hammer solenoids, producing a very characteristic "ka-chunk, ka-chunk" sound. We didn't have a printer, but output everything to punch cards which we listed on a 407 accounting machine.
Behind the printer is the 1622 card reader-punch. 250 cards/minute reading, half that punching.
Next is a Model 83 card sorter.
Next to that appears to be a 1623 Storage Unit, which held the extra core memory for the 40,000 or 60,000 digit models. I'm jealous, we only had a 20,000 digit machine.
Finally, the 1621 Paper Tape Unit (reader-punch). We didn't have one of those either.
Programing was either in FORTRAN II, or an assembler called SPS, Symbolic Programming System. No operating system as such, things ran under control of Monitor I or Monitor II. Most PC users wouldn't believe the amount of software, or the amount of work we got done with these primitive machines.
Copyright © 1999, 2000 by Paul Rogers. All rights reserved.