From: Rick DeNatale Date: 2007-07-21T09:09:39+09:00 Subject: Re: Taking the bull by its horns [was background] On 7/20/07, Morton Goldberg wrote: > That's pretty complete. I've never read a history of IBM -- I was > going entirely from my memory of working with early IBM computers and > the oral tradition transmitted me by the many IBM employees I have > known over the years. BTW, you omitted the 1130, which I remember as > small scientific/engineering machine roughly contemporary with the > 1620. But we seem to have determined my memory isn't to be trusted. > So where do you place the 1130? Well I'm working somewhat from memory, but the 1130 was later (1965) than the 1620 (1959), the 1130 was introduced after the first 360s, and used similar electronics packaging called SLT for solid logic technology which was somewhere between discrete transistors (as the 1620 used) and ICs. The architectures were completely different, the 1130 was a 16-bit binary word machine while the 1630 was decimal using 6-bit BCD codes. Numbers were represented by varying length strings of these BCD codes addressed by the low order digit, and one of the extra bits (called the flag bit) used to mark the high order digit, and if set on the low order digit that the number was negative. I know a little bit about the 1620 since it was the first computer I encountered. The main design goal of the 1620 was to be low priced (for the day) that's something it did have in common with the 1130. The first model of the 1620 had no hardware for addition, instead it used a lookup table to add numbers right to left digit by digit. The IBM internal code name was Cadet, which is purported to have really meant Can't Add, Doesn't Even Try. When I interviewed with IBM right out of college, one of the guys I talked to was named Jack Gibson, who's most known for the Gibson Mix which was a benchmark used to compare processor speeds. Jack was quite tickled when I told him that I'd cut my teeth on the 1620. He'd been on the project management team on the 1620 having come from a similar job on Stretch. He said that when an engineer on Stretch wanted approval to add a gate to the machine (gate here means a large swinging rack holding circuit cards), and estimated a cost of say 10s of thousands of dollars, the request was granted without much thought. But on the 1620, if an engineer came in and said he needed to add some pull-up resistors to a few circuits, and that it would cost, say 10s of dollars, he was told to go back to the drawing board. -- Rick DeNatale My blog on Ruby http://talklikeaduck.denhaven2.com/