Posted on 02/20/2012 7:27:49 AM PST by ShadowAce
I know of many critical national defense systems running on computer made in the 1960’s through the early 1980’s. Why? Because they work. The engineers of that time knew how to create systems the run. Today’s computer programmers just love to create cool looking things and managers don’t know a thing abuout computers and don’t know how to make such systems.
The 402 pictured was an “accounting machine,” taking a stack of punched cards, doing some simple totalizing math, and printing out journals and statements.
It was “programmed” by a panel with hundreds of jack holes into which you plugged up to a couple hundred patch wires to do the appropriate column arrangements and totalizer functions. The panels were on frames which were removable, so you could have a shelf of them ready for whichever run you needed to make.
The 402 was little sister to the 407, which I helped to maintain on a couple of occasions (in 1963, mind you).
It’s a wonder they can still maintain that stuff. An IBM field office held a very large stock of exotic bits and pieces for these monsters back in the day, but that’s all long gone by now.
I don’t think programmers or codes from yesteryear are *inherently* “better”. It’s more a case, I think, of if you have a code that has years of field use under its belt, then presumably it’s been “battle tested” and had the bugs wrung out of it. For large, complex software, that’s worth a lot.
Put another way you’d rather have version 10.0 than version 1.0 of just about anything.
Not really. Those systems were working years ago and after many attempts to duplicate their functionality they still remain. Many of the lessons learned years ago have been forgotten.
Upgrading a PDP8 from its native 4K (12-bit words) of core was a big deal, involving massive add-on boxes. It could support, however clunkily, up to eight 4Kword banks, seven of them external to the CPU box.
We bought a rare non-catalog DEC 8K word add-on For our PDP8/L in 1972. It was a 10.5” high box, same size as the CPU. It cost $7000. The box could be populated, IIRC, with two more core stacks for a total add-on of 16K words.
Of course a PDP8 word, being 12 bits, was what we would call a byte and a half.
DEC persisted with word widths that were multiples of 3 bits (i.e., 12, 18, and 36) from its founding in the 50’s up through 1971, when they introduced the 16-bit PDP-11. From then on, it would be all 8-bit byte addressable machines.
So octal was the natural notation system for the early DEC machines. This artifact shows up in Unix and in the C language, which began life on the 18-bit DEC PDP7/9/15 series. It became an annoyance when Unix was ported to the PDP-11, and all subsequent architectures from DEC and others.
The next year we had gotten a Honeywell mini-computer that was still programmed with data cards. We had problems with the cards swelling up due to humidity, and the computer was considered stae-of-the-art, it allowed us to program in FORTRAN and COBOL.
Pinging you because I know how much you love that Apple IIe of yours.
...upcoming high-performance line of Amiga computers...?
I see.
You've rekindled old memories. Was a programmer on the 360s starting in 1967, coding in RPG and using a punch card compiler that was a foot high - you inserted your program in the middle. Got looked down on by the COBOL guys but we did some NEAT stuff with it. We replaced our tape drives with SEVEN MEG hard disk drives and nearly wet ourselves with random access and all that space, which was more than we would EVER use.
Think out mainframe was 16K and when we wrote programs that were too large, all we got was a "Program too big" message. You had to blindly start trimming wherever you could as you didn't know how far you had exceeded memory. One time we got an invoice to print by removing a period.
For a long time we prided ourselves on writing tight code ('cause we HAD to). Computer time was more expensive than programmers until one day we realized power had increased so much that the reverse came true. I used to get yelled at because I didn't spend hours desk-checking code (and still get compile errors) and instead just let the computer find them in seconds. Had a hard time explaining to the non-geek manager that times had changed.
Years ago some government guy said not to worry about them tracking you as they had tons of info on punched tape or regular tape that they couldn't read as nobody had the programs or knew how to change them. Hope that's still true.
If we followed the “if it ain’t broke, don’t fix it” philosophy, we’d all be driving Model T’s and working in CICS and there would be little reason for innovation.
I prefer the “if it isn’t broke, improve it” philosophy.
Ah, swelling cards. IBM made the best quality cards, but was forced by the Justice Department in the 1950s to spin off their card manufacturing business and allow competitors.
IBM still marketed the best quality cards through their captive manufacturer, though.
IN field service, we had this precision card gauge; a precisely stamped and lithographed aluminum sheet by which you could gauge the dimensions of the card, and the accuracy of the placement of all 960 punched holes.
If a customer was having problems with his punched cards and they were from a third party supplier, we Customer Engineers could hold one of his cards up on our gauge and show him the dimensional problem; however, we were not allowed to say “take a look at your crappy off-brand cards,” or the Justice Department, we were advised, would come and whack our collective pee-pee.
“Many of the lessons learned years ago have been forgotten.”
Being a retired Sys/App Assembler programmer from days gone by, can current programmers read ‘core dumps’ understand machine language and op codes etc? just askin.
OS/2 was way ahead of its time, and far superior to the version of Windows available at that time. The reason OS/2 is still in use, especially in financial institutions, is the fact that version 2 was specifically designed to act as a peer with AS400 systems.
What killed OS/2 commercially was that the hardware needed to run it really wasn't available at the time (it was a REAL 32 bit OS), and there simply weren't that many applications written for it: That was back in the day when IBM was really iron fisted when it came to licensing technology and IP.
Mark
OS2 Warp on a 386 in the basement about 1994, slick OS, but nothing native to run though. Ended up going back to Windows in 1995.
I've always found the quickest way to progress is "If it ain't broke, you're not trying hard enough."
The corollary is "Technology is a lot like violence--if it doesn't solve your problem, you're not using enough of it."
That's a new one to me.
Disclaimer: Opinions posted on Free Republic are those of the individual posters and do not necessarily represent the opinion of Free Republic or its management. All materials posted herein are protected by copyright law and the exemption for fair use of copyrighted works.