Jaundiced Eye: #2

This is what I was thinking. This is what I thought was worth recording. This feature is irregular. The future is unknown. Feedback is welcome.

Notes on Norman

Professor Donald Norman, former Apple Fellow (University of California) [1]
Why is Everything so Complicated in the Digital Age?
Fourth Innovation Lecture & Inaugural Lecture for
  The Digital World Research Centre [2], University of Surrey
1999-03-03, 5.30 pm

Notes expanded to full article, Tuesday 23rd March 1999.

After an introduction by the Vice-Chancellor of the University of Surrey -- he's calling himself the Chief Executive these days in a case of what I can only describe as Business Envy, but he still possesses the measured tones and pomp that you've come to expect from reading any Terry Pratchett novel -- Don Norman took the podium and began to speak, quietly. He made the following points, presented here in roughly chronological order.

Norman says he's for for-profit education, selling courses to companies for employee lifelong learning, and that he's involved with companies doing that. This received a large number of nods of approbation among the audience; the University is running more and more short courses. I swear, the UK education system is going to short-term US hell, where you learn nothing much at all in your twenties but lifelong learning will supposedly make up for it, while the needs of industry reign paramount. If Norman underwent for-profit education himself and doing his PhD bankrupted himself and his family as a result, well and good; the thought of people pulling up the drawbridge behind them and destroying the educational systems that they benefited from and that moulded them is a trend that I have always found particularly worrisome. Come to think of it, the tendency for privately-educated UK politicans -- the Chief Executives of UK plc -- to destroy the educational systems that they didn't pass through is probably rather worse.

He trotted out Moore's law, which really needs retiring since it's past its sell-by date. Gordon Moore originally noted and extrapolated on the increasing number of transistors on Intel's CISC processors, and that observation has been generalised in increasingly suspicious ways ever since [3]. Moore's law is about as credible and reliable as Bode's law [4] these days. It may even be artificially maintained by market expectations. If it's a law, I'm an orangoutan. Norman used Moore to ask what you would do with a computer a thousand times more powerful in fifteen years' time. The answer will be 'wait for Microsoft Word to finish printing your document to the Linux spooler', I expect.

He contrasted specialist and generalist devices, saying that good specialist devices suited to a clear purpose are what's needed; tools that humans can understand. Since Nokia, one of the companies sponsoring the Digital World Initiative, now make cellular phones you can play games on, I think Norman's point is being completely ignored in practice here. I don't know anyone who fully understands their cellular phone, much less anyone worried by this failing.

Technical development isn't the problem; Norman thinks we have more tools and abilities than we know what to do with, and he's right there. Social implications, and the use of technology for clear communication, are the problems. It's not what we've got, but how it gets implemented and used. There is a lack of good conceptual models for human/computer interfaces; building your own mental models of behaviour of digital devices from the limited non-redundant information presented, on a video recorder LCD, say, is tricky, and design criteria often favour form and minimization over understandability and usability.

Norman illustrated this using salt and pepper shakers. He asked how you would know which was which (convention only goes so far), and then segued smoothly into a discussion of seemingly identical controls in that old favourite, the nuclear power plant control room. Improvements in said control rooms and in cockpits were cited - Norman claimed there weren't any crashes on commercially-scheduled flights last year, which caused me to do a double-take (SwissAir? Far eastern flights?), but I presume that he was talking about internal US flights.

Speaking of Norman's US perspective, I had to do some mental gear-shifting when he started talking about how hard early phonographs were to use, too. I believe he meant record players. And, speaking of gear-shifting: when he talked about how easy a car was to drive, he contrasted the complex car with the simple unicycle, which is far more basic in construction but is much harder to master.

This comparison gave pause for thought. 'Easy to drive' may be true for US automatics, on wide, straight US roads, but here in the UK we have driving lessons, narrow roads, and gears. Oh, and roundabouts; you know, traffic circles. Nothing much happens if you fall off a unicycle, whereas cars do annoying things like explode if you steer them into concrete walls at speed. 'Easy' includes weak penalties when things go wrong, and then there's cognitive load. That's why driving a car is so tiring, whereas once you've learned to balance on a unicycle the skill becomes innate. This is a trite example that doesn't hold up to close inspection, in my view.

Norman claimed that the best technology isn't perceived as technology, and isn't sold upon the strength of its technology. He illustrated this by waving his wrist and talking about watches as an item of jewellery, branded separate from the technology and how well they actually kept time. People will pay more for watches that actually keep measurably worse time. He mentioned SMH, the owners of Swatch and other brands, and how they'd targetted different market segments. He didn't discuss Swatch's bizarre Internet Time concept, although I imagine Negroponte would [5], given a similar speaking situation.

Oddly, there was no mention of Apple's iMac as an attempt to move personal computing into a similar mature-product post-technical-features product-differentiation phase where people judge a computer solely by the colour it comes in. (By all accounts, lime iMacs are the least popular, and special promotions to clear putrid green stock are underway.) In fact, Don Norman, former Apple Fellow, didn't mention Apple at all. Well, they're just another middling computer company with badly-designed hardware and overly complex impossible-to-use unstable software that is also going to hell these days; being associated with Apple is probably just embarrassing, for any number of possible reasons.

Human-centred design is needed, using cognitive engineering (Norman never used the word 'ergonomics', which appears to have fallen out of favour). 75% of accidents are attributed to human error, and there's a tendency to blame the human as the cause. An example tailored to the satellite-focused Surrey audience was given: a satellite launch goes off-course because digits are transposed, and has to be blown up. Don mentioned a Russian launch [6], but the fabled Mariner 1 launch sequence also sprang to mind [7] -- it's a human error, so the human gets disciplined, even though transposition errors and the like are unavoidable human behaviour and should be expected and compensated for.

If the same result occurred due to noise in the circuit, you'd redesign the circuit to prevent the effect of the noise. Norman made the point that process engineering should be designed around the human as an invariant: in the intellectual age work should be adapted to humans, rather than humans adopted to work (as occurred in the mechanical age). Though if you replace 'work' with 'education' that reads a bit worryingly; humanity is just getting soft.

Electronic books were mentioned. Norman is involved with one company, and they're doing lots of user testing, figuring out good user-centred product design, and doing all the things that Norman advocates. In the meantime, three other companies are marketing arguably-worse products, trying to find a business model that actually works and can make money.

There's a gap between these two approaches that needs to be bridged, which is where Norman believes that cross-discipline cognitive engineers -- a social science background, but a practical implementation-driven bent -- come in, to produce decent, well-designed, usable products the first time around.

This is a great idea -- in theory. Unfortunately, all the social scientists I've ever met do social sciences because they don't care for or can't cope with engineering or the hard sciences.

Norman's own background is a masters in engineering followed by a jump into social psychology for his PhD, so I think he's preaching to the wrong crowd. He should be talking at engineers who have started finding people more interesting and easier to work with than boxes, if you ask me -- but then, as one of those, my own viewpoint is equally biased. I'm with Heinlein on this: an engineer can always learn the fuzzy stuff and pick up some culture, but the reverse is much harder and far, far less likely.

Three to five people is an effective number for a group; the Mythical Man Month got a mention. Universities were criticised for always favouring theory over implementation; people doing implementations get sneered at as the lowest of the low. Yup, Norman's got an engineering background and a chip on his shoulder about it, all right. Bet he laughs at Dilbert, too.

Social sciences have problems, not least of which is physics envy and the desire or need to stick 'science' in their subject titles as justification so that people take them seriously. Social scientists need to take their understanding of people and behaviour and communicate that to different disciplines - engineers, designers, and (given the DWRC sponsors) the four cellphone manufacturers whose representatives were present. Cognitive engineering, an angle on the problem Norman's spent most of his career evangelising as a new field, got a few more mentions. Ergonomics didn't.

There was an emphasis on mature viewpoints rather than that of a 22-year-old designer enthralled by adding neat features for the sake of it -- but then Norman is looking a bit greyhaired himself these days.

It was an interesting if abstract talk -- there were no obvious visual aids apart from the salt and pepper cellars and the wristwatch. However, the lights dimmed halfway through at 6pm and then rose again, which turned out to be due to 'smart lighting' in the new lecture building that senses movement, and everyone sitting still at six.

My first thoughts: wouldn't infrared detection of body heat work better? Or of noise? And what sort of time-out sans hysteresis just happens to dim on the hour exactly? Norman's comment: 'Why is it called smart?'

A sleeping audience wouldn't need any light, and hey, it's a university lecture hall after all. I think Norman should spend some time in France, experiencing light cut-offs that darken corridors as you're walking halfway down them -- French regimentation taking precedence over considerations of fluorescent ballast and tube lifetime, power, or convenience.

My second thoughts: my, this is awfully convenient for the talk, and a very illuminating point about the assumptions of product designers to boot. Nicely done.

All in all, Norman went over a lot of familiar ground, and he's affable and a pleasure to listen to. I was surprised by the reaction of much of the audience to what Norman did say; it's almost as if they'd never heard of any of this stuff before, as they laughed and nodded at seemingly novel points. Now that I find scary. But I'm not sure why. Maybe it's because they're simply more human. Because their lifelong learning never saw fit to cover this stuff. Or because they're older.

L.

Copyright © 1999 by Lloyd Wood <L.Wood@surrey.ac.uk>. All rights reserved. For permission to reprint, please contact Lloyd.