Context: Sophie was one of the main architects of the original ARM processor
Fairbairn: So was there sort of a major insight or breakthrough or whatever in doing the processor, or was it just sort of you read the papers, you sort of absorbed things and sort of went through the process? Were there any sort of major aha moments, or points that--
Wilson: There was nothing aha about it. It was just, "You can do this stuff. You just get on and do it." As I say, it was more the reverse. We expected to find a roadblock. We expected to find why National Semiconductor employed all those people in the building in Israel and why they found it so difficult. And we just never did. We also-- we thought it was going to be much harder than it really turned out to be, so we put a lot of effort up front into verification and modeling.
I particularly enjoyed this anecdote (talking about the prototype 6502 based BBC Computer, not the ARM here):
Wilson: ...Hermann determined that if there was no logical reason why it didn't work, then it must be the fault of the in-circuit emulator, and that what they should do is take the in-circuit emulator out, put a native processor in it, and surely it would work then, because there was no logical reason it shouldn't. And annoyingly, he was completely right. It did work.
There's a sequence in the BBC drama "Micro Men" which appears to dramatise this, albeit with an external clock wire as the culprit. Worth watching, it's brilliant.
I watched it this evening (stopped watching the other night as it was so good I wanted to watch the whole thing with my wife) and it was very entertaining.
Armstrong's portrayal of Sinclair seemed rather cruel, but then I don't know how accurate it might have been.
The ZX81 and ZX Spectrum were a big part of my childhood, so I had a big case of Clive-worship as a kid. Watching a clip from this I now feel slightly attacked :) but I'm settling down with a very British cup of tea to watch it tonight.
Sophie was THE main architect of the original ARM processor. She created the entire instruction set de novo and wrote the first code for it, including an OS.
There are 318 at the moment and they post several new ones each week. But they still have quite a back log so it is not surprising that an interview from 2012 hasn't been posted yet.
This list (currently 1337 items) includes videos and the transcripts:
I have now read the transcript and it has "end of tape 1" and "start of tape 2" which makes me think it might be an audio only interview. In that case it probably won't show up on Youtube.
Douglas Fairbairn, who did the interview, was the founder of VLSI Technology Inc which was the first fab to make ARMs and, together with Acorn and Apple, set up the ARM company.
In an interview I did with Prof Andy Hopper he mentioned that a CAD group was created for a networking company which got merged with Acorn. That led them to have a chip design capability that they were trying to figure out what to do with it. This is a slightly different viewpoint of the same story.
And of course ARM helped to fuel Apple's extraordinary success in the 2000s-2010s with the iPod, iPhone, and iPad ...and non-success in the 1990s with the Newton and eMate.
It really is an extraordinary record, especially for someone with no academic background in EE. She literally just sort-of picked up digital logic design intuitively, more or less all at once. And then wrote a BASIC and helped design a game-changing super-efficient processor (etc), because why not?
And no one at school noticed she might have unusual talents.
I wish we could have an interview that tries to get some insight into how this kind of mind operates.
Fairbairn: So was there sort of a major insight or breakthrough or whatever in doing the processor, or was it just sort of you read the papers, you sort of absorbed things and sort of went through the process? Were there any sort of major aha moments, or points that--
Wilson: There was nothing aha about it. It was just, "You can do this stuff. You just get on and do it." As I say, it was more the reverse. We expected to find a roadblock. We expected to find why National Semiconductor employed all those people in the building in Israel and why they found it so difficult. And we just never did. We also-- we thought it was going to be much harder than it really turned out to be, so we put a lot of effort up front into verification and modeling.