HAVE you ever wondered why the astonishing growth in computing power has not made PCs a whole lot cheaper? Or easier to use? Or both these things?
Since 1965, the power of computers as measured by the number of transistors per microchip has doubled every 18 months, just as Gordon Moore said it would. By the same exponential formula, a laptop that cost $3000 in 1990 could cost $6 today. But you don鈥檛 see any $6 laptops, and despite the hike in processing power, the experience of using a computer has sunk to new lows. Even when installed on the fastest laptop on the market, my software runs slower, is less reliable and is more frustrating to use than ever. Any new software release is distinctly worse than its predecessor. Why?
The answer is 鈥渇eaturitis鈥�: a drive to pack new releases with new features and options. What you actually get is 10 different ways to do the same thing, with fewer and fewer of them intuitively obvious. The result is obese software that consumes all the benefits (and more) of the speed and memory improvements of the hardware. For example, the latest releases of Windows, Acrobat and Lotus Notes all run much slower than earlier ones.
Advertisement
It is featuritis that has kept the price of a laptop more or less constant for 15 years. How do I know? I have used a laptop since 1979, when I bought an early portable called a Sony Typecorder. Today, I spend all my working hours 鈥� 6 hours a day, every day of the year 鈥� in front of a top-of-the-range laptop. People find it odd that I carry two such laptops around with me, but one is basically a back-up and is better thought of as a carrying case for two spare batteries. Although this weighs a lot, it can provide as much as 20 hours of use without mains power. But this is more or less what I had in 1979 鈥� from four AA batteries.
Granted, we have come a long way since then, with spectacular flat-screen colour displays, vast disc drives and features like the magnificent USB port 鈥� which automatically senses the need for drivers for scanners, printers or webcams, say, and installs them. But by and large, today鈥檚 approach to general-purpose machines is like adding blades to a Swiss army knife: most people wouldn鈥檛 be surprised if their next handset made coffee. This featuritis has made computers exasperating to use.
The same is happening with cellphones. We are packing so much into one small device that you need to be 15 years old, with tiny fingers and great eyesight to use it. Packing supplementary features into an itsy-bitsy device just cannot continue, not least because batteries will not be able to cope.
Happily, early evidence of new thinking can be found in the great popularity of the no-frills Blackberry cellphone and emailer. Its appeal is its simplicity. And it is simplicity in personal technology that the MIT Media Lab is working on. We see two paths to simplicity which need to be pursued in parallel, as one addresses the short term and the other the long term.
The short-term solution to bloated systems is good design combined with a diet: cutting down on options and features. Mainstream devices should stay simple. For those who really want 鈥渁dvanced鈥� features, the idea would be to create dedicated, special-purpose devices by adapting software to make the hardware behave in different ways.
Of course, the longer-term solution is to give computers common sense. How many times have you accidentally hit the Control key instead of Shift, only to watch your work disappear? Most six-year-olds could figure out that you had made a mistake, and know you meant upper-case. Likewise, an easy-to-use computer should do what I mean, not what I say 鈥� and by no means should it send me a dancing paper clip to ask. For sure, common-sense computing is a tough task and won鈥檛 happen in my lifetime. But it is time to start trying.
In the near term, more than any one plan of action, we users need a course correction in attitude. Too many people I know believe that machines simply have to be built the way they are, and that they (the users) are the stupid ones. They actually expect costs to rise or at best to stay the same.
Simpler machines can be much less expensive, and while consumers, I believe, want this badly, the manufacturers have little interest in making it happen because the high end of any market is more profitable. It is in the industry鈥檚 interest to force us to use and believe in bloated, feature-extravagant systems that use more and more memory and processing power.
This has got to stop. We need to use Moore鈥檚 law the other way around, for the user鈥檚 benefit.
