Remberance of things past - computer division

Miscellaneous and Personal Stuff I Must Share
21

My first brush with computers was a Fortran IV class in college. We had to produce stacks of keypunch cards to feed the goddamn thing, and I sucked at it. What the hell is a “do loop”? :slight_smile:

Then HP came out with a calculator that could do trig functions and everybody was in awe.

I didn’t buy (couldn’t afford) my first computer until about 1986, which was a Tandy 8086. It had a 5.25" floppy and no memory to speak of. I had pirated floppies of WordPerfect and dBase, but they were of little use to me. Mainly I used the thing to fuss around with learning how to tweak the settings. I remember that Tandy offered a 10 MB external drive, but I couldn’t afford to buy one. From there I moved through the 286/386/486/Pentium series. Along the way, I bought an Atari console with Pong, Tank, and Space Invaders. High times.

22

All you little kiddies who remember waaay back to the days of the first PC’s – Timesharing computer systems were already old-hat by then.

I learned Fortran II on an IBM 1620 – Yeah, puched cards galore.

When I got to college (1969), they had a CDC-6400 – a relatively modern (then) super-computer of the day. The engineering department had two Univac SS-90 computers – archaic even then – that were donated by a lumber company up north, for a tax write-off no doubt, when they upgraded to IBM 360’s.

The Univac SS-90 had a grand total of 5000 words (of 10 decimal digits each) of memory, stored on a rotating magnetic drum. Considering 2 digits approx. equal to one modern byte, that’s about 50K bytes. No mass storage device at all – for that, we had punched cards.

The machine stood about 6.5 feet tall and maybe 10 feet wide. The back-plane was entirely hand-wired. It had hundreds of cards plugged into slots, each card having four to six little solid state devices on it. (Meaning, each being one flip-flop. ETA: Each one, a little white disk about the size of an Alka-Seltzer tablet.) The logic of the computer instructions was implemented largely in the way these were connected together with those wires on the back place.

We had a thick stack of the logic diagram blue-prints on a big easel. And we had cartons and cartons of spare parts. And there were several dozen empty slots in the machine. We had some engineering types who were all into that – we designed some new computer instructions we thought would be neat to have, and they designed the logic circuitry to make it so. Then they plugged in a few extra cards and wired it all up. We called that “hands-IN programming”.

One day, one of the guys brought in his antique Teletype, which used 5-bit Baudot code. We ran a wire across the floor, from the sign bit of the X-register to a power amplifier, and from there to the electromagnet that drives the teletype. With some carefully timed programming, I wrote a subroutine to output one character by toggling the sign bit of the X-register for just the right number of machine cycles, for each bit, to match what the Teletype needed to see. Later, we did similarly for the input – I think we connected that wire to the Card Punch Ready signal, since we didn’t really use the card punch that much.

ETA: Visiting the Bay Area? Check out the Computer History Museum in Mountain View.

23

The Computer Museum in Mountain View had a working copy of the Difference Engine (made not long ago, of course) on loan. I saw a demo. It was awesome.

Ada Augusta is my pinup girl. :slight_smile:

24

I did the stats on my senior honors thesis in 1982 using punch cards! I feel so old!!!

25

I’ve always regretted that I never got to actually program a 1620. I learned what assembly language was from 1620 books in my public library. Cooper Union had one, which was tempting, but when I visited MIT there were IBM 1130s around every corner, so I was hooked.

My last post was made before I read this. Do visit - not only is it cool, but it is free. I am so volunteering there when I retire. I learned to program the PDP-1 as an undergrad - the one Spacewar was developed on - so I’m sure I can pick it up again after 40 years or so.

26

Speaking of pinups of Ada. Here is a nice one. Very safe for work.

27

I saved all summer for my first floppy drive, ~$250 for a 5.25 inch drive. The discs were about $5 each, giving me all of 720k of storage (if you cut out the tab allowing you to make the disk double-sided).

A month or so ago, I bought a 750Gb external HD for $75.

Had I wanted 750gigs at 1982 prices I would have needed to spend $5,500,000. :eek:

28

It’s there still. I took my brother there when he was in town, just about a month ago, and we saw the demo.

I want to know if it can compute the value of .999… once and for all. :slight_smile:

29

NOPs. These were empty instructions you would sprinkle along the length of your program. It stood for “no operation”. If you wanted to insert a line later, you only had to shift things to the nearest NOP rather than shifting every line to the end (on systems that did not support adding a line inside a program but only at the end).

30

That’s why they invented assemblers.

31

Burned in my memory forever.

SET BLASTER=A220 I5 D1 T3 P330 H6 E620

32

Burned in my memory

4900796

(1620 programmers will know.)

33

Yeah, I remember those days. How many thousands of dollars did that 286 cost?

Back then your computer could easily cost more then your car.

34

Things were simpler back in the day.

I learned Fortran II from the IBM 1620 Fortran Reference Manual – a softcover book maybe 50 pages long, half of which described the Fortran II language, and the other half was full of other technical details (like how to save compiled subroutines on the disk).

I took it home and read the entire language half of the manual (about 30 pages) in one sitting, and I knew Fortran II.

Today, try buying a good beginner’s teach-yourself book. As I write this, I have before me a teach-yourself book on JavaScript and DOM scripting. 760 pages.

Visual Basic – about the same.

C++ and Java books – easily over 1000 pages each.

PHP – 610 pages.

And on and on. Who lives long enough to read all that these days?

35

This is the reason I wrote my first assembler. But you didn’t shift anything - you inserted a jmp at the nop to jump to the inserted code, and then returned - since there was seldom only one instruction you wanted to insert.

36

Real men learned their languages from books by Dan McCracken. Who I am happy to say I met once.

37

But it could play a mean game of blackjack. :wink:

Your story reminds me of a question, in fact: When did the modern hexadecimal notation come into use? Did IBM introduce it alongside eight-bit bytes with the System/360?

(No, I’m not going to reminisce here. I will link you to the SIMH Project, which contains emulators for a number of the truly old computers mentioned in this thread so far, and a lot of others besides.)

38

Those were the days when Real Programming students listened to lectures by Donald Knuth (if you were at Stanford) or Butler Lampson (if you were at Berkeley). Of course, MIT and CMU had their luminaries too.

39

My first program was in Fortran on the 1620 at San Jose State. (I was in high school, but a teacher was a graduate student at SJS and let us use his account.) My second program was on the same machine but, using the multiple-punch key on the IBM keypunch (I guess an IBM 029, which would have been brand-new at the time), I wrote the program in absolute machine code. The only output device that computer had (besides lightbulbs) was a card punch.

Were you following me around? :smiley: My college had CDC-6400’s. We used to run our programs on graveyard shift, to get fast two-hour turnaround times.

By the mid-1970’s I worked for a company that ran full-page ads for The World’s Fastest NMOS Ram. Those chips had 768 bits of memory each – no, “bits” is not a typo for “kbits.” I remember the S.J. Mercury-News sending a photographer once, when we had (a record-setting) 8 megabytes assembled into a single large cabinet.

I once programmed a controller with 32 bytes of RAM – again, not a misprint. It was only the Chinese Remainder Algorithm required by some ECC that posed trouble; fortunately I also had a 128-byte FIFO at my disposal, and I could use that for the algorithm’s necessary temporaries.

40

real men learned programming from programs. you looked at some program listings and realized the syntax and conventions. you looked at books later for details when you got stuck.