Psion: the last computer
Secrets of the Sony we never had
“It’s amazing what was thrown away.” — Mark Gretton
During 17 years at Psion, Mark Gretton designed the hardware architecture for both the Series 3 and Series 5 organisers. He designed numerous ASICs along the way, and also took a management role that led to the formation of Symbian. In Mark's words:.
In the early days David Potter was very involved in the company and he was prepared to take big risks. He bought into the idea and believed in it himself.
Gradually he lost that passion in the business, and decisions were made by people who didn’t understand the consumer electronics business, and didn’t have the stomach for it.
He let it go away.
There was so much waste, so many ideas, and so many talented people. It’s amazing what was achieved — and then thrown away.
How I joined Psion
I joined Psion to do cross compiling in 1986. We were in the last stages of launching the Organiser II. The whole system was emulated on a VAX in C. This code was used as an executable spec for people who hand compiled the machine code for this tiny little washing machine processor. It was the only way to get it into a small enough space. It was incredibly labour intensive but some very bright people took an almost religious pride in the quality of their cross compiling.
I then started work on a 16-bit architecture with a completely blank sheet of paper. It wasn’t clear what the first product would be, it was simply an architecture for 16-bit handheld computing. The first product turned out to be a massive project — the MC, a revolutionary new solid-state laptop computer.
“It was a bridge too far… we couldn’t get away with that now.”
The MC project was completely mad; it was insane. This was an entirely new laptop, a new architecture four new ASIC chips, a new OS, a new GUI, and a new suite of business applications, all in a new product — and we thought we could do it with just 30 people. Looking back I can’t work out why we thought we could. Colly Myers had a saying at the time: “Don’t look down!” That was the trick.
We launched anyway, even though the hardware was dodgy and the software was dodgy. It wasn’t reliable. You couldn’t get away with that nowadays.
But it was bristling with innovations. Flash memory was new, so we invented a serial bus, and created the first solid state disks. We invented a new touchpad. We invented all sorts of new power management techniques, and a whole GUI was drawn up.
It was a bridge too far everywhere you looked, and it didn’t work out. But it let us hone that architecture.
Success out of failure
We then said let’s do a small QWERTY based product on this 16-bit architecture — that was the birth of the Series 3 — my particular project. We’d designed ASICs before — I designed three for the MC. So I started a project to put 3 ASICs and the CPU on one chip.
It’s standard practice now to make an ASP [application specific processor, or “system on a chip” — ed], but it was absolutely unheard of at the time. I spent a good few man years of my life on this project — by this time I had a CAD workstation (as opposed to a mainframe for ASIC simulation) putting an 8086 processor and all the gubbins, such as LCD controller, on that one chip. It was actually about nine months I was chained to that workstation. That chip made the 3a possible, and by then we had a stable, robust platform.
The 3a was the first time everything came together and it was a really great product — and it sold very well. So the massive undertaking of a whole new architecture was finally paying off — it went on to survive for over 10 years.
We can’t do it all alone
As sales of the 3a were taking off, more long term questions were asked, such as “do we need a new 32-bit architecture?” This lead to a relatively small team starting development on our 32-bit platform (Epoc32 as it was called then).
Again, I started on the hardware. Again I looked at lots of CPUs, but in the end chose ARM. Again, this was not the obvious, mainstream choice it would be now. Then, ARM was unheard off outside of Acorn and ARM (Newton). But technically it was strong in terms of MIPS per watt and die area: people still counted [transistor] gates quite carefully in those days.
That was Eiger, an ARM710 based ASP or SoC. In parallel, a complete new 32-bit OS was developed, a mammoth task. This went on to become the Symbian OS which has shipped in over 100 million smart-phones. In fact, it was such a massive task that it soon became clear the expected revenues from PDA sales alone could not justify this level of R&D investment.
I worked with a couple of others including Colly Myers to prepare something for the board that basically said, “we can’t compete with the hardware business alone, we should consider a licensing model, and restructuring the company around that”.
There was an expression going around at the time: “Is Psion going to die like IBM, or is Psion going to die like Apple?” Would we open up and license our technology, or — and remember Apple was really, really suffering at the time — would we die like a closed proprietary system? So we decided to die like IBM!
(Now, at TomTom I can say we’ve survived like Apple!)
So the company split up into four [Psion Computer, Psion Software, Psion Industrial, and Psion Dacom] and I joined the board of what was then Psion Software, which went on to become Symbian with Colly, Bill, and David Wood, who’s now a senior VP at Symbian. For the next year, I spent a lot of time in Tampere in Finland at Nokia, doing pre-sales, selling the idea. There wasn’t a lot to show them — prototype Series 5s and development boards for the OS — but the premise was “Look how great our 16-bit stuff is!” And Nokia went and benchmarked it and saw that yes, it really was highly efficient and highly reliable.
Back in London we were having a terrible time ramping up production of the Series 5 at our own factory in Greenford. Everything was going to shit. David Potter pulled me aside one day and said, “we need some good guys in the PDA business” and put me back into the hardware business. So I left Psion Software and joined the Computer side, running the engineering side. After the restructuring and formation of Psion Software most of the brains and software guys had been stripped from the PDA business, leaving it as more of a shipping and production entity — it had no control over its destiny and was totally wrapped up in its production problems. I desperately tried to get Psion out of the manufacturing business — there was no shortage of people in Asia willing to help us — but David Potter was against this. Again, it was one of his big strategic errors.
Then in 1998, Symbian was announced. Everyone thought they’d take over the world, while I thought ,”Shit — I’ve made a bad career move [moving back to the PDA business].”
Let’s do a smartphone!
It was a tough time. The PDA business was not performing and it had lost its technological core and control over its destiny. Symbian were not interested in our business as they were focused on smartphones and Nokia. I survived on the management board but only just and we went through a lot of MDs.
At one level things looked up, we now had a mature and stable 32-bit OS and a cost effective platform and at last, following the Revo introduction, high quality outsourced manufacturing.
I was in charge of the Psion Computers roadmap and presented countless new products. These included a colour version of the Series 5MX, a road warrior 5MX with a built in modem from Dacom, a Blackberry style device, like the Revo, with wireless push email, called Project Blade. There was also a cost-reduced Bluetooth enabled Revo on the roadmap.
But the commonly held belief then was that PDAs were dead and everything would be connected: ie, a smartphone or Nokia Communicator-like device. At the same time David Potter recruited a new group CEO, David Levin. He was convinced we could not be successful with a smartphone without a powerful partner from the wireless world, and so we signed with Motorola to do Odin.
There was a feeling of ’what’s the point?’ Nokia’s going to kill us anyway — which was totally self-fulfilling and which I don’t subscribe to, then or now”
I was deeply against the Motorola contract — for me it was the final nail in the coffin.
Even if doing a smartphone was a good idea, Psion could have done it on its own. At the time we were working with Inventec and other companies that had their own cellular technologies. It was an investment, but compared with some of the stuff we’d taken on in the past it was nothing.
The iceberg looms
David Levin, the new CEO had no feel for consumer electronics. So against the advice of most of the Psion Computers’ management team, including me, we signed the deal with Motorola.
Motorola were notorious for shafting other companies or indeed other parts of the group, it was a highly political organisation. Either the semiconductor people were shafting the handset people, or the other way round.
Now the argument was that Motorola had the sales channels, that was the rationale. But the phone channel was the network operators and they didn’t understand smartphones; arguably, they still don’t. But PDAs were sold through the IT channel. We could have sold Odin through the IT channel quite effectively.
I was in Germany in 2001 when Margaret Rice-Jones, Psion Computer MD, took the call from Motorola — and her face went white. She told me Motorola had cancelled Odin, and that we were on our own. I knew that was the end of our business. We had no other roadmap, no plan B.
It had been an awful two years.
We were getting to the stage where we had programming resources at last, a stable platform, and were able to start turning out some product at last. There was great spirit there.
Yes, PDA sales peaked in 2003 and it wasn’t for lack of trying that I tried to revive the business.
A third revival?
Even by itself, the Bluetooth Revo which was ready for launch [cancelled 11 July 2001] would have continued to be a profitable niche, I’m quite convinced of that.
The Revo made even smaller would definitely have made a good wireless email terminal and there’s no reason that couldn’t have gone on to enjoy sales. Almost all the things were there: you need good keyboard, good battery life, you just need to put in a GPRS modem. The sales channel didn’t have to be the operator. Where there was a bit of work needed was on the server side; but there were smart people in the company and we could have cracked that. We’d just lost the ambition.
Following 2001, I worked for two years in the wilderness doing push email software, Again, there was no stomach for taking this to the consumer so it got given away to Visto.
And Visto killed that as well — David [Potter] just wanted shot of it and gave it away, the people and everything.
In the end, fear triumphed.
During the whole Symbian hype, when the Psion share price was high and the Dacom side of the business was doing well, a group was set up called The Incubator Group.
People wrote business plans for new ideas. One was Blade, the push email system. One was a Hard Disc based music player, that Ken [McAlpine] wrote (before Apple launched the iPod). And another was a DAB-based receiver, a PC peripheral, that Ken proposed. That was the only one to make it to market. Even navigation devices were discussed.
The WaveFinder radio was typical Psion: it wasn’t a great product, it had lots of flaws — but it was completely revolutionary in its time — and DAB radio went on to became a valid market.
“They got too rich probably, and too involved in the City.”
Psion did this more than once. They create a market, but you’ve got to keep at it. They made a huge technical investment with huge innovation but lost the will to keep at it. The senior management board lost interest in the technology, and they weren’t engaged in it personally.
Why? They got too rich probably, and too involved in the City — that’s another world that didn’t understand Psion.
The great stuff was done out of a can-do spirit that was gradually lost in the latter years; the split didn’t help. It wasn’t fatal in itself. Psion died by a succession of blows, people with less and less ambition — and more fear.
This fear became self-fulfilling. The mentality was that everyone will get us: “Nokia will get us, Palm will get us, or the Japanese will get us”… in the end we got ourselves!
It was OK when were in a niche — that was fine. But if you’re successful in a niche and sell some stuff, then naturally you get competition. To me this isn’t a problem — it validates the market.
The idea at Psion was that this was a bad thing and you had to move on. It’s in a marked contrast to TomTom. We’re in a real market, meeting a real market need — and we’ve sold an order of magnitude more than Psion ever did. And of course we have over 300 competitors — and still hold on to 50 per cent market share! If a niche ever grew so it wasn’t ever a niche, Psion abandoned it.
There was this fear of competition that was right in the heart of David Potter.
After leaving Psion, Mark joined a small Dutch PDA software company TomTom in 2003 run by former Psion Computer MD Harold Goddijn. As chief technology officer, Gretton assembled a team of former Psion engineers to create its first hardware products. In fiscal year 2006, TomTom reported revenues of €1.36bn making a net profit of €222m.
Sponsored: Becoming a Pragmatic Security Leader