14 Sep 15
When I was at school, in about 1970, one of the maths masters asked a group of us if we would like to go and see a computer. How strange that sounds now! The computer was housed in the basement of a Government building – it took up most of the floor area and consisted of large cabinets with things whirring about inside, together with sundry peripherals such as a printer and a device for reading punched cards. Our maths master was a bit of an amateur programmer and was allowed to use the machine occasionally along with some other approved people. After all, it was probably the only computer in town.
In 1973 I started work in the same building and there were plenty of people who still talked about the recent arrival of the computer, so I would guess it was installed in the late 1960's. My question is, exactly how powerful was that machine compared to modern technology? I'm betting that any smartphone would beat it hands down and possibly even my TV would outperform it. It was viewed as a miracle at the time, but how good was it?
As you might guess, I am not a very scientific horse.
14 Sep 15
Originally posted by C J HorseThe late '60s early '70s was the era of the mini-computer. The PDP-11 [1] is exactly that era, although the machine you saw was a mainframe. In those days they were still talking about megaflops [2] in hushed voices. A smartphone can get to a gigaflop. So in terms of pure speed they are over a thousand times more powerful.
When I was at school, in about 1970, one of the maths masters asked a group of us if we would like to go and see a computer. How strange that sounds now! The computer was housed in the basement of a Government building – it took up most of the floor area and consisted of large cabinets with things whirring about inside, together with sundry peripherals suc ...[text shortened]... racle at the time, but how good was it?
As you might guess, I am not a very scientific horse.
[1] https://en.wikipedia.org/wiki/PDP-11
but see also DEC vms VAX, which is a fairer comparison with modern machines as it had a 32 bit architecture.
[2] flop = floating point operation per second, in other words a 1 gigaflop processor can execute 1 billion basic arithmetic operations (add 0.5 and 1.5 for example) every second.
14 Sep 15
Originally posted by DeepThoughtI actually worked on the PDP-11 and although it may have had a 32 bit CPU, it was still in the megaflop area, nowhere near a gigaflop. It's funny. ATT, I was working at Goddard Space Flight Center, and we were being trained on Apollo tracking and timing but they asked me to spend 2 hours a day on these old mainframes, it seems they had to show usage, machine time, to justify to the government they were being used enough to keep them. I was introduced to a very early version, you had to enter digital numbers one switch at a time, say 16 switches side by side, 1000100010001000 say, then hit a switch that was 'enter', then do that again and again a jillion times to get an active program, and then you could put that program into a punched card reader so at least you wouldn't have to go through all that rot twice🙂 But even that was an advance over my first job at Goddard. My first job there was "programming' an analog computer, there was this connector board with 1728 pins that connected to it's mating cradle, and the interconnections defined a program. I was just a technician with the job of wiring the program designed by an engineer and the program I was wiring was designed to analyze the stability of a spinning satellite with weights of some mass distributed around the inside perifery, would it spin stably in orbit. So to change a configuration to 4 weights around the inside from 3, say, meant a rewiring of the analog rack! We then, using an engineer I worked with by the name of Horst Slinglofff, we got the analog computer to be able to be controlled by one of the PDP11's so at least you could do the same thing with punched cards to change configurations. Man, were those glory days🙂 Horst had to work like the devil to get even 50 Kilobytes per second of data transfer from the main computer over a coax cable🙂 Boy, were we ahead of our time🙂
The late '60s early '70s was the era of the mini-computer. The PDP-11 [1] is exactly that era, although the machine you saw was a mainframe. In those days they were still talking about megaflops [2] in hushed voices. A smartphone can get to a gigaflop. So in terms of pure speed they are over a thousand times more powerful.
[1] https://en.wikipedia ...[text shortened]... or can execute 1 billion basic arithmetic operations (add 0.5 and 1.5 for example) every second.
15 Sep 15
Originally posted by sonhouseWhen my Dad was working for ICI in the 60's they had a main-frame at another site. So he'd write a program in FORTRAN, cycle down to the post office and send it by telegram to the operation. It was compiled by hand by some technicians and they'd send the bug report by the next morning. This would leave my Dad two hours to debug the program and cycle to the telegraph office to send the revision. The revised program's bug report would inevitably arrive the following morning...
I actually worked on the PDP-11 and although it may have had a 32 bit CPU, it was still in the megaflop area, nowhere near a gigaflop. It's funny. ATT, I was working at Goddard Space Flight Center, and we were being trained on Apollo tracking and timing but they asked me to spend 2 hours a day on these old mainframes, it seems they had to show usage, machin ...[text shortened]... ond of data transfer from the main computer over a coax cable🙂 Boy, were we ahead of our time🙂
25 Sep 15
Originally posted by DeepThoughtNow most bugs are sniffed out by software.
When my Dad was working for ICI in the 60's they had a main-frame at another site. So he'd write a program in FORTRAN, cycle down to the post office and send it by telegram to the operation. It was compiled by hand by some technicians and they'd send the bug report by the next morning. This would leave my Dad two hours to debug the program and cycle t ...[text shortened]... the revision. The revised program's bug report would inevitably arrive the following morning...
Originally posted by sonhouseI would put it differently. Many bugs are prevented by better programming languages, compilers and IDEs. Bugs do still occur. They are just not the type that can be easily prevented automatically.
Now most bugs are sniffed out by software.
I would say that bugs are prevented not 'sniffed out'. There are very few commonly used tools that actually track down bugs. I certainly have never used one.