8 ms·

ARX, Arthur and RISC OS (2012)

58 points by mr_golyadkin 6y ago

codeulike 6y ago
The development of the first ARM chip is just such an amazing story. I just wanted to say a bit about why Acorn decided to do this whole RISC thing at all, Acorn was a manufacturer of home computers, why on earth would you be crazy enough to design your own chip? Sophie did an analysis of the compiler output for a lot of common compilers as part of the process of trying to choose which chip we should use next, we were looking at things like the 68000 and the National Semiconductor 16032, or 32016 as it got renamed. I don't know if you remember, a second processor was done with that 32016 and Nat Semi had produced a poster that said "32016 is not a late 16 bit machine it's an early 32 bit machine", a year later we crossed out 'early' and wrote 'late' on it. This analysis showed that all of the complicated instructions and all of complicated mem-copies, the compilers just don't generate them. They generate Load/Store, add, subtract, and/or, compare, branch, that's about it really, the rest of them, unless you're writing in assembler, you just don't use them, so they're a complete waste. Now, the aforementioned, Arthur Norman, who had written the LISP interpreter for the BBC Micro, and was in and out of Acorn the whole time, had come up with a design for a machine with 3 instructions called the SKI machine. Whose instructions were called 'S', 'K' and 'I' and it ran LISP in hardware and he built one of these things, it never worked, it was so big and covered in wire-wrap it was always broken. The idea was to prove that you really didn't need very many instructions for a general purpose machine. ... Tony Thompson, who had done that implementation of BASIC 64, he got to do the core of the operating system, so the real guts of Arthur itself. That brought it out of reset, powered it up, organised all the memory and so forth. We had a really interesting time with part of that, we had one of the machines that we just could not get this thing to boot reliably. You could boot it, turn it off, reboot it and sometimes it would work and sometimes it wouldn't. It turned out, it would boot fine if you left it long enough, but if you didn't turn it off for very long then it didn't reset properly, and this was because the fan on the board was still spinning and the back EMF on the fan was enough power to keep the ARM running. And that's why you've got an ARM in your phone today, it would take no power to keep a 3 micron ARM with 25000 transistors would run for 30 seconds off the energy stored in the fan.
fanf2 6y ago
Arthur Norman’s SKI Machine was interesting: rather than being LISP (as mentioned in the talk) it was a combinator reduction machine https://en.wikipedia.org/wiki/Combinatory_logic https://en.wikipedia.org/wiki/Combinatory_logic - SK combinators are the way Miranda does lazy evaluation, so the SKI Machine was basically Miranda in hardware. Miranda was a forerunner of Haskell https://en.wikipedia.org/wiki/Miranda_(programming_language) https://en.wikipedia.org/wiki/Miranda_(programming_language)
kiwijamo 6y ago
A shout out to those involved in Acorn, RISC OS, etc around the late 1980s and throughout the 1990's. Around 1992 or so, our family bought our first computer (when I was around 5 years old) which was an Acorn A3000 running (if I recall correctly) RISC OS 2 on ROM. The school I went to also had Acorn computers (including the RISC PC toward the latter half of the 1990’s). Nowadays I consider myself extraordinarily privileged to have experienced RISC OS on Acorn hardware in its heyday. I see so many things in modern computing inspired by RISC OS. The taskbar in modern era Windows and Mac OS is remarkably similar to RISC OS 2 which was released in 1988. RISC OS was also an early pioneer of anti-aliasing fonts and had it years before it appeared in Windows. However, sadly, there are many things now lost to history. A machine running RISC OS 2 booted in literally <3 seconds. It was basically just a console screen with “RISC OS 2048K” (or something along those lines) and a second or two later you were at the desktop. It is near impossible to find any modern computing device that boots as quick as the A3000 running RISC OS did. RISC OS also had a wonderful file-save dialog which was simply the file icon—which you were to simply drag and drop directly to the folder you wanted to save the file to! Given you usually had the right folder open anyway, it was usually a case of just dragging to your already open folder. So much more sensible than having to search for a folder in a small explorer window. Thanks to all involved. Acorn and their RISC OS sparked a lifelong interest in computers. Thank you!
- AnIdiotOnTheNet 6y ago
  I guess because I'm in a bit of a mood thanks to arguing with Linux Desktop evangelists lately, I'd like to point out that Rox Filer implemented exactly this file saving functionality (and many other good ideas RiscOS) on Linux. Then they built Rox Desktop around it. It was great. There were AppDirs, context-menu-only, spacial file management, drag-n-drop to save, etc. Naturally, the project's reward for trying to improve the Linux Desktop experience was to have all its ideas ignored by the wider Linux community as it faded into obscurity. If anybody was ever wondering why all these people who have gripes about the Linux Desktop experience never seem receptive to your pleas of "contribute! you can make it better!", this is why.
  - twic 6y ago
jeffrallen 6y ago
The whole thing is fine, but this passage is GOLD: > We had a really interesting time with part of that, we had one of the machines that we just could not get this thing to boot reliably. You could boot it, turn it off, reboot it and sometimes it would work and sometimes it wouldn't. It turned out, it would boot fine if you left it long enough, but if you didn't turn it off for very long then it didn't reset properly, and this was because the fan on the board was still spinning and the back EMF on the fan was enough power to keep the ARM running. And that's why you've got an ARM in your phone today, it would take no power to keep a 3 micron ARM with 25000 transistors would run for 30 seconds off the energy stored in the fan.
- m0xte 6y ago
  There's quite a bit of power in that sort of stuff. If you move the steppers on my 3d printer while it's not plugged in it'll light the backlight on the LCD screen very brightly!
  - jeffrallen 6y ago
    I double dog dare you to put your tongue on it!
    - m0xte 6y ago
      Hahaha no way :)
quattrofan 6y ago
Fabulous read, and a lot of it really made me laugh especially the F0 keyboard bit.

58 points by mr_golyadkin 6y ago

The development of the first ARM chip is just such an amazing story. I just wanted to say a bit about why Acorn decided to do this whole RISC thing at all, Acorn was a manufacturer of home computers, why on earth would you be crazy enough to design your own chip? Sophie did an analysis of the compiler output for a lot of common compilers as part of the process of trying to choose which chip we should use next, we were looking at things like the 68000 and the National Semiconductor 16032, or 32016 as it got renamed. I don't know if you remember, a second processor was done with that 32016 and Nat Semi had produced a poster that said "32016 is not a late 16 bit machine it's an early 32 bit machine", a year later we crossed out 'early' and wrote 'late' on it. This analysis showed that all of the complicated instructions and all of complicated mem-copies, the compilers just don't generate them. They generate Load/Store, add, subtract, and/or, compare, branch, that's about it really, the rest of them, unless you're writing in assembler, you just don't use them, so they're a complete waste. Now, the aforementioned, Arthur Norman, who had written the LISP interpreter for the BBC Micro, and was in and out of Acorn the whole time, had come up with a design for a machine with 3 instructions called the SKI machine. Whose instructions were called 'S', 'K' and 'I' and it ran LISP in hardware and he built one of these things, it never worked, it was so big and covered in wire-wrap it was always broken. The idea was to prove that you really didn't need very many instructions for a general purpose machine. ... Tony Thompson, who had done that implementation of BASIC 64, he got to do the core of the operating system, so the real guts of Arthur itself. That brought it out of reset, powered it up, organised all the memory and so forth. We had a really interesting time with part of that, we had one of the machines that we just could not get this thing to boot reliably. You could boot it, turn it off, reboot it and sometimes it would work and sometimes it wouldn't. It turned out, it would boot fine if you left it long enough, but if you didn't turn it off for very long then it didn't reset properly, and this was because the fan on the board was still spinning and the back EMF on the fan was enough power to keep the ARM running. And that's why you've got an ARM in your phone today, it would take no power to keep a 3 micron ARM with 25000 transistors would run for 30 seconds off the energy stored in the fan.

fanf2 6y ago

Arthur Norman’s SKI Machine was interesting: rather than being LISP (as mentioned in the talk) it was a combinator reduction machine https://en.wikipedia.org/wiki/Combinatory_logic https://en.wikipedia.org/wiki/Combinatory_logic - SK combinators are the way Miranda does lazy evaluation, so the SKI Machine was basically Miranda in hardware. Miranda was a forerunner of Haskell https://en.wikipedia.org/wiki/Miranda_(programming_language) https://en.wikipedia.org/wiki/Miranda_(programming_language)

kiwijamo 6y ago

A shout out to those involved in Acorn, RISC OS, etc around the late 1980s and throughout the 1990's. Around 1992 or so, our family bought our first computer (when I was around 5 years old) which was an Acorn A3000 running (if I recall correctly) RISC OS 2 on ROM. The school I went to also had Acorn computers (including the RISC PC toward the latter half of the 1990’s). Nowadays I consider myself extraordinarily privileged to have experienced RISC OS on Acorn hardware in its heyday. I see so many things in modern computing inspired by RISC OS. The taskbar in modern era Windows and Mac OS is remarkably similar to RISC OS 2 which was released in 1988. RISC OS was also an early pioneer of anti-aliasing fonts and had it years before it appeared in Windows. However, sadly, there are many things now lost to history. A machine running RISC OS 2 booted in literally <3 seconds. It was basically just a console screen with “RISC OS 2048K” (or something along those lines) and a second or two later you were at the desktop. It is near impossible to find any modern computing device that boots as quick as the A3000 running RISC OS did. RISC OS also had a wonderful file-save dialog which was simply the file icon—which you were to simply drag and drop directly to the folder you wanted to save the file to! Given you usually had the right folder open anyway, it was usually a case of just dragging to your already open folder. So much more sensible than having to search for a folder in a small explorer window. Thanks to all involved. Acorn and their RISC OS sparked a lifelong interest in computers. Thank you!

AnIdiotOnTheNet 6y ago
I guess because I'm in a bit of a mood thanks to arguing with Linux Desktop evangelists lately, I'd like to point out that Rox Filer implemented exactly this file saving functionality (and many other good ideas RiscOS) on Linux. Then they built Rox Desktop around it. It was great. There were AppDirs, context-menu-only, spacial file management, drag-n-drop to save, etc. Naturally, the project's reward for trying to improve the Linux Desktop experience was to have all its ideas ignored by the wider Linux community as it faded into obscurity. If anybody was ever wondering why all these people who have gripes about the Linux Desktop experience never seem receptive to your pleas of "contribute! you can make it better!", this is why.
- twic 6y ago

jeffrallen 6y ago

The whole thing is fine, but this passage is GOLD: > We had a really interesting time with part of that, we had one of the machines that we just could not get this thing to boot reliably. You could boot it, turn it off, reboot it and sometimes it would work and sometimes it wouldn't. It turned out, it would boot fine if you left it long enough, but if you didn't turn it off for very long then it didn't reset properly, and this was because the fan on the board was still spinning and the back EMF on the fan was enough power to keep the ARM running. And that's why you've got an ARM in your phone today, it would take no power to keep a 3 micron ARM with 25000 transistors would run for 30 seconds off the energy stored in the fan.

m0xte 6y ago
There's quite a bit of power in that sort of stuff. If you move the steppers on my 3d printer while it's not plugged in it'll light the backlight on the LCD screen very brightly!
- jeffrallen 6y ago
  I double dog dare you to put your tongue on it!
  - m0xte 6y ago
    Hahaha no way :)

quattrofan 6y ago

Fabulous read, and a lot of it really made me laugh especially the F0 keyboard bit.