Psion: the last computer
Secrets of the Sony we never had
’My brain was coming out of my ears. Literally’ — Nick Healey
Nick Healey was a lead designer on the Series 3 range of PDAs, head of the design team for the Psion Series 5 applicatins & UI, and head of design at Symbian.
By the time of the early 1990s, Psion had survived a few cycles of redundancy, so everyone who was left there really knew their stuff, which was handy as the company didn’t really do much management, training, or strategy.
As for hiring policy, the plan was always just to recruit big brains. Psion did the single most important thing that management should do, and did it very well: find smart people, make them happy so they don’t leave, and set them free to use their talents. Do that and everything else takes care of itself, I reckon. It’s really not rocket science, as long as you can find middle managers and indeed senior managers who can accept life in such an environment. If you get this right, you don’t have to bother reading books on management, which is another bonus.
I’d worked on the specs for the apps and the user interface for the Series 3 machines We were pretty proud of those machines. There aren’t many tech gadgets from 12 years ago that I can imagine using nowadays, although they are bringing back the Casio LCD TV experience on your mobile phone, only worse and more expensive than before. Even today, the 3MX does contacts, diary and note-taking really well. It’s a great PDA as long as you don’t want mobile email or web, and who’d want them unless you had a wide screen and a keyboard you can type on. Come to think of it — like a Series 5. Or a 3MX! I’m just converting from using a Revo to a Series 3MX at the moment, by the way.
But no one could say how long the Series 3a would sell for. It was an uncomfortable time for us at Psion — it was possible that the Newton would take over the market, or perhaps the next HP palmtop, or something rumoured to be coming from Microsoft.
Organising the new Organiser
So Protea kicked off properly. David Wood, Bill Batchelor and myself were the three development managers in the software team, and on the Protea project itself, that became known as the Series 5. Bill was the project manager.
I ran the specification team, and David had the job of creating and managing the application and UI-level Software Architecture — and teaching all the new coders. Having long ago forsaken coding myself, I was the only designer in a sea of coders. My software specification team had the job of what the software should and shouldn’t do, and the UI to do it.
We moved into new offices and September 5, 1994, 12 new coders arrived, doubling the size of the software team. It was the day after Colly’s [Myers] 40th birthday bash: “Happy birthday, see what you can do with these guys.”
Unfortunately, I was delayed by designing and project managing the PsiWin connectivity software, a thousand apologies for which, well for the reliability anyway, and by the time I got to the Protea project the software engineers had already written lots of stuff.
They’d implemented Windows-style dialogs, for example — a 2D-layout that had enormous flexibility, but it was completely unusable from the keyboard. They agreed to throw it all away and write an “up-and-down” dialog layout you could navigate with just the arrow keys, after a month of gentle persuasion and personal threats.
Arguing the toss
It was a very argumentative culture, just because of the number of experienced, argumentative people. Healthy, I think, if a bit scary.
Colly’s influence was all-encompassing, and it could be scary as a newbie when Colly himself will come to your desk and tell you what a **** you are. You soon learnt that this was just part of the “training” at Psion. And that you could give it back to him.
The culture was to argue as long as you liked with anyone you wanted to. I had an argument with Colly once, with me saying that impossible timescales are worse than having no timescales, to him it was just the opposite. Like it really matters. But it lasted from about 9pm till they threw us out of the pub at half eleven, and it continued on the street outside the pub till a quarter to one, when some understandably miffed woman leant out of her bedroom window and impolitely invited us to go home. Sadly, neither of us had given an inch in four hours. Colly had a heart attack when he got home — nothing life-threatening, and he forgave me.
Designing for simplicity
The overall goal was to make a machine that would wow the PDA world, which at the time was pretty much the Series 3, but an aim was also to interfere with the laptop market a little. So — an uber-Series3, a jack of even more trades. And then the Palm Pilot arrived mid-project, and everyone at Psion laughed, because it hardly did anything.
With the Palm Pilot, we learned the hard way about functionality for normal people [It was the Palm that taught us “the hard way” and showed us that in fact we weren’t really getting to normal people at Psion]. It turned out that what reviewers called “user-friendly” — only about one in ten of the population could actually handle. It’s still the same nowadays. I did run user tests on the first Series 5 prototypes, actually, in my own time, not that I had any, to the sound of a stable door being bolted, but normal people weren’t formally part of the development process. They’re “rule one” nowadays, for me, with every client I work with.
On Protea we had a “Thought Police”, which was a committee of three, David and Bill and myself, and if you had an idea for a big change to the spec, you only needed the approval of two. It allowed a huge range of new ideas and fixes to go through, without us getting bogged down in red tape. Come to think of it I’m honoured in the Series 3MX easter egg as “Thought Police” too. Clearly, I was enjoying myself.
“With the Palm Pilot, we learned the hard way about functionality for normal people.”
Products ship when “pressure to ship” exceeds “desire to improve”. Protea, like every Psion project in the 90s, took twice as long as its original schedule, as we always attempted too much, and ending up losing things late in the project which would have been much better lost earlier. A fair few of those things turned up in the 5MX.
And also, we treated what were really “best case” targets as deadlines, and of course hurt ourselves by rushing things to try to achieve the unachievable. So “desire to improve” stayed pretty high, as more and more things went wrong, while “pressure to ship” slowly mounted to the same high level.
It was a deeply stressful way to live. At one point it felt like my brain was coming out of my ears — I mean, literally, for days, weeks, my head felt like it was exploding. I had to have an MRI scan. They found nothing, as they say.
It was all particularly depressing for me because in 1994 to 1995, I had written up a load of thoughts for the other senior managers on the reasons why all our projects ended up taking twice the desired timescale. Like using best-case targets as deadlines, and so on. I guess I just didn’t sell it well enough.
We had one fun Protea meeting in February 1996, with the “deadline” then October/November 1996, in which the 15 or so managers on the project, from Colly down, were invited to give their best estimate as to what the real ship date would be. Everyone reckoned October, November or December, bar one guy who said February 1997 — take a bow, Geert Bollen — and me. I said a few words as to why I felt it had all the same problems in front of it as all previous projects, and that it would therefore take twice its current schedule, which came to the last week of May 1997.
Colly tore my head off for daring to say such heresy, but in the end I was out by a week, and we shipped the ROM the first week of June 1997. And again the week after. It gives me no pleasure to say this. It did almost make my brain explode.
Indeed, if you looked at every previous moment that we’d rescheduled the project, and just doubled the timescale we’d estimated each time, then every single doubled estimate came out at May 1997. Hardly Bill’s fault, I hasten to add — the company was asking him to try to achieve each impossible deadline. He was well aware of all this.
By the way I heard a couple of years ago from one who was there that the “Zen of Palm” was to a fair extent reverse-engineered after the success of the Pilot, that in fact they actually dropped a lot of functionality from it, late on, through time pressure. That’s pretty funny, if true.
“Psion did the single most important thing that management should do: find smart people, make them happy so they don’t leave, and set them free to use their talents.”
But it’s a hard thing to manage. Symbian, I gather, are currently trying to reduce the number of experienced and expensive employees and they have a micromanaged production-line environment there. It’s ironic that the move towards this was born out of a desire to hit software targets better than in the old Psion days. The best intentions, and all that. The timescales are more predictable, but at what cost?
A new nightmare at Symbian
A year after Protea shipped, Psion Software overnight became Symbian, and all the software people bar about one moved over. Suddenly, Psion Computers had no software design experience. I did have the offer of going to Psion Computers and being their design person for the software side, and in retrospect I so wish I had done.
Psion were a bit unsure about speccing out the software in their products by themselves. Like they did with the Revo: they just chucked the Series 5 software onto a much smaller screen — without a backlight too — and so a lot of the text is smaller than it ought to be.
So I found myself at Symbian as its lead software designer. It was a nightmare.
It was set up with a schedule to develop four new phone-centric versions of the EPOC (now renamed SymbianOS) UI in 18 months, which was laughably unachievable even without having to find and train enormous numbers of new employees. And before four new corporate overlords each began demanding every possible second of your time so that the other three wouldn’t get it.
And being the only senior software design person was an especially unjoyful experience. On one of the four platforms, we had to combine the spec of our previous Psion applications and user interface, with the specification of a new device Nokia had been developing. So there’s me locked in a room with about 15 Nokia designers, who, it slowly transpires, view Symbian as a software house they have just bought in order to deliver their product.
I explain to them the thinking behind the various things in the Psion UI — why it just saves changes automatically, why a second tap means “open”, why it has infoprints [informational messages] in the corner, why ToDos appear on the main Diary screen, all that kind of stuff that we’d worked out over the years, and they listen and then they say “yes, but we want you to do this”. So we didn’t exactly get on well in those meetings.
And after a couple of days we realised there was an even bigger underlying issue, which was that we each put entirely different types of stuff in our specs because they did entirely different jobs as part of entirely different development processes. So I went back to the board and asked the rather obvious question, “So what’s Symbian’s development process?” and got the answer: “We haven’t got one yet. Just do your best.” It was great fun, the Symbian deal.
A few months down the line I’m in a meeting with one of the other Symbian shareholders about one of their phones, and I find myself selling a largely fictional plan for how we could allegedly deliver their software in the given timeframe, and their senior guy is tearing me to shreds. Now this is a guy who was involved in the due diligence on the formation of Symbian, and approved all these ludicrous delivery dates. I opened my mouth to tell him exactly where he could stick his dates, and after several seconds with my mouth open, I decided to quit. It had been seven months of hell.
Then I really wished I’d gone to Psion Computers!
What really happened to Psion?
In 1998, Psion lost all of its software expertise to Symbian, which proceeded to slowly borg it into bureaucracy. But Psion had absolutely fantastic product engineers, and grew a new software team. It also had a fantastic brand in this country, and in many others, and it had a history of designing groovy new things too.
But even after this, I’m told they turned down internal proposals to develop some really exciting new stuff like an iPod, and a TomTom, years before Apple and TomTom did it.
They even pulled out of some straightforward projects — like pulling the Bluetooth Revo at the last minute. My best guess is that they got distracted by their ownership of the stake in Symbian — as you well might, if you’re thinking you might be about to own the mobile phone world for free.
Since leaving Symbian in 1999, Nick Healey’s UI consultancy Slash Design has designed smartphones, PDAs, PC-phone integration software, Linux suites, mobile apps, internet radios, pointing devices, and a restaurant.
Sponsored: Benefits from the lessons learned in HPC