2007 Interview with Stuart Goodwin
You've been quoted in the past as saying that the first Acorn Microcomputer had its roots in an automated cow feeder that you designed in 1977. Did your cow feeder contribute to you becoming involved with Acorn, or were you already on their radar by that point?
Yes, the cow feeder lead directly to the System 1 - but it didn't, of itself, cause Acorn to find me. Hermann simply went to the Cambridge University Processor Group and asked for recommendations (who knew what?). I think it was probably John Gibbons who referred Hermann to me - as the person who knew most about CMOS. And that came about because the previous summer (before the cow feeder) I'd worked for ICI Fibres Research centre, where I'd introduced them to CMOS because I'd bought the RCA catalogue and read it cover to cover - all the datasheets, all the application notes. Of course, I'd built nothing, then, but when it came to new logic circuits for equipment in the research centre, I moved them straight from their DTL logic modules to CMOS logic - I built them a wrap detector (spotted when the thread wrapped around the machine and broke) and a drop counter (watched drops falling: they were clear, so it was a lens effect from an IR emitter to a photodiode). So I had practical experience in CMOS - I was building a digital clock (I started in school and kept elaborating it - its got two time of day counters, a memory, an alarm, a down counter: its still in use by my bed!) - and Hermann wanted someone capable of building "an electronic pocket book" (we might call it a PDA now) which clearly needed CMOS. So that was me.
At the point that you were designing the System 1, did you have any particular brief in terms of what the computer would be used for (or was it more a case of showing people what such a machine could be used for)?
Not really. Again, what haappened was that Hermann wanted me to design his pocket diary, so I drew up a paper design and described it to him and said what it could do. And he asked if it would work, so I showed him the designs for the cow feeder and for my own computer. Will they work? Of course! Your pocket book is a simplification of this part of my own computer design. Will all that work? Yes, of course! So he challenged me to build it. We modified the design somewhat to suit the parts he and Science of Cambridge already had (e.g the IO combo device rather than ordering the matching IO for the 6502 that I already had bought - a beauty in white ceramic with gold pins) and built it over that summer. Of course, it worked. (At this stage, that was just the base CPU board and keypad and calculator display - the cassette interface was Steve Furber's design for the cassette interface of his home machine!)
There was no other brief: the System 1 was an adaption of part of the computer I was building for myself. I cared what my machine would be able to, Hermann simply cared about the fact it could be built and that I could explain why it was built so and why the 6502 was better than the SC/MP that Science of Cambridge used. When Steve started to join in, it was much the same.
How would you describe the evolution from the System 1 to the first home Acorn machines? Was it a straightforward progression, or were any major direction changes made along the way?
A progression in a strange set of ways: the parts of Steve's machine and my designs combined, in an order decided by Hermann with a commonality introduced by Chris Turner (I'd simply used eurocard sized veroboard and verowire to build it by hand without PCBs, Chris decided to go the whole hog and adopt the eurocard, DIN indirect connectors and racking standard - I'd used direct connectors (and I think Steve had, too)). The memory board designs were mostly from Steve's machine, the 80 character soft font VDU mostly from mine (though I don't recall if I was a fan of the 6845 or if that was Steve). Eclectic selection of components - we used Motorola, Intel, MOSTEK, National as we saw fit! (though nothing from Signetics who'd made the microprocessor in Steve's home computer...)
I've read that Chris Curry particularly pushed to take Acorn into the home computing market, and that some of the engineers wanted to steer well clear. What was your opinion at the time on the direction the company was taking?
I don't think that's correct. The disagreement was over technology rather than direction. We simply didn't like the chips in the Atom, in particular the 6847 which was, after all, an NTSC only video chip. The techies eventually won the argument - the BBC Microcomputer was made with our choice of chips!
You're widely credited with developing BBC Basic, which is still widely used today. How much of your own stamp did you put on it, and did it feel like a significant achievement at the time?
BBC BASIC is a compromise between my advanced interpreter of the day and the BBC's desire to keep the language "standard". Most of the time the advanced features managed to get included unchanged, so overall I was happy (the only significant loss I can remember is over labels vs line numbers - the BBC committee insisted on line numbers).
Yes, it felt significant at the time: a fast BASIC interpreter which was convenient to use (in ROM on all machines) with many advanced features.
The Atom, BBC Models A and B and then the Electron all launched within 3 and a half years of one another. Was each machine radically different from a development perspective, or was the bulk of the non-cosmetic work behind the scenes effectively done once the Atom had hit the shelves?
Large amount of common software really made it possible. A/B/Electron really quite similar to each other ("only" a new main logic chip to design). The Proton which became the BBC Microcomputer had been in development before the Atom conceptually - it inherited many of the parts of the Acorn System range (and all the software and some of the new hardware was prototyped there - for example, I designed the MODE 0 character set using the System range's 80 character soft font display), though the impetus to start it as a real project only came after the Atom was financially successful. And the project only gained real momentum with the impending visit of the BBC people to see a Proton prototype - which didn't exist at the time: we had a week to build it...
How close was the prototype Proton to the finished models? Was it closer in spec to the Model A or B?
Model A and B are essentially the same, just different in memory and peripherals. They're thus both aspects of Proton. Proton as originally conceived did not have the SAA5050 character generator (MODE 7) and teletext capabilities. It did have the second processor - the original concept of using multiple processors for the design was what unified the techies at Acorn, since it meant everyone could have the processor they wanted. (The BBC Microcomputer being the IO processor of a total Proton system with whichever language processor you wanted.)
How much input did representatives from the BBC have into the computer? You've mentioned the insistence on line numbers in BBC Basic, but were there any other significant inclusions (or exclusions) that they wanted in the finished models?
The whole teletext side of it (as noted above, but also including a full teletext off-air adaptor and the ability to download programmes from TV broadcasts). They placed some big desires on the precision of our PAL encoder which we met by making a digital one. I guess Acorn got the contract partly because what we were building anyway was close to what they wanted and they saw that our engineering was good enough to make it happen.
What did you think about the BBC Computer Literacy Project link-up? A blessing or a curse for the techies?
I don't think we really cared much about it. We were sceptical that millions of people would learn to actually program, but hopeful that people would learn to use the machines.
Were you surprised that the BBC Model A wasn't more successful? Was there always going to be a kit to upgrade Model A to B?
No, we always thought Model B was what people needed. With disc drives. After all, that was effectively what we had (System 3 or 5 machines with 80 character displays with discs and later on hard drives).
What did you think of the BBC Model B at the time, and - looking back - is there anything you'd change about it?
The SAA5050 was a pain - expensive and different from the rest of the machine - but it did give a lot more memory for programs! The reliability of early VIDPROCs was poor which was a shame (in the early years (before ARM) we never got to make a custom chip on a proven set of design tools...). Most of what we wanted (in an IO processor) was there (that's why it cost so much!).
How closely did Acorn look at other home computers in the marketplace, and how do you think the Acorn range compared to rival machines?
We didn't really have time to look at other machines (there was a WHOLE LOT OF HARD WORK to make this stuff - we were mostly working 12 hours a day, 7 days a week!) though we knew the rough hardware specs of most of them and many of the staff had prior experience of older models. We were fairly determined to make things that were better - more powerful, more flexible - though I don't think we really got that right until ARM (project A started in 1983...). Mostly we were building what we wanted for ourselves. As we built language processors, most of the development machinery moved from System range to Model B plus 6502 second processors and that reinforced our desire to make things capable: we were using them all the time ourselves.
The perception I've always had of the BBC Model B is that parents would buy the computer as an educational tool because it's what the kids were using at school, and then the kids would only actually use it to play games. Was this perception shared by you and others at Acorn, and were you surprised to find that a games industry built up around the Beeb?
I don't think I really thought about it. It was nice to see complicated games programmed on the machine.
Did you ever play games on the Acorn machines? If so, do any particular titles stick out?
I never had the time, patience, reactions, inclination... Frak! sticks to my mind. But the rest of the company were playing Revs or Elite. I wrote a patience game (which later on became !Patience and swept the company!). I played Tertis (not a misspelling) on the ARM a bit.
Did you have any involvement with the Acorn Electron? With the benefit of hindsight, do you think the Electron was a) necessary and b) a success?
I designed the case size (out of a cardboard paper hankerchief box!) - we didn't like what the first industrial designer came up with and needed a "rebuttal". Otherwise I didn't have that much to do with it after the basic parameters for the SoC were set - that was all Steve's problem.
I'm sure a cheaper entry into the Acorn range was necessary.
I'm sure the Electron wasn't a success: it should have been a year earlier.
The BBC Model B+ is often referred to as a mere stopgap between the Model B and the Master range. Is this fair?
B+ and its shadow memory were a necessary development to make the Master. Maybe one should think of all machines as a stopgap until something better comes along? Maybe the Master was a stopgap for the Archimedes?
By the time the Master and Master Compact launched, the Atari ST and the Commodore Amiga - both 16-bit machines - were already on the market. In spite of the steps forward the Master offered, was there a concern that Acorn was falling behind in the marketplace?
No: we had ARM - we knew what was coming with Archimedes.
How did Acorn's financial problems and production issues of the mid-80s affect you and your work? What's your take on how these problems came about?
Electron was late. And there was the "Christmas that wasn't" - when all the disposable income seemed to go on something other than a home computer (CD players in particular).
Hermann mostly kept these problems away from the staff. Which made the redundancies harder to understand.
Was the ARM project ever in jeopardy as a result of these problems?
What was it like to work at Acorn during the heyday of the Model B?
"Never a dull moment, frequently an exasperating one"
Atom, Proton, Electron, Master and Archimedes. Who came up with the names, and what did you think of them?
Many of the project code names were mine. Atom, Proton, Neutron, ARM. And, of course, the famous Fred and Jim (Chris Turner added Sheila later). It was surprising when project names became something people watched. System 1 was "Hawk" until it became System 1. I don't think any of the techies had much of an opinion on names - we knew that dull ones may be needed so changing Proton into BBC Microcomputer was fine. We knew we had to call things something before they had proper names (and even had to call things something before they were even well defined things). Sometimes (Project A - ARM and Project B - Master) we had dull project names...
Interview by Stuart Goodwin, Autumn 2007.