September 29, 1998


Jeff Hawkins became something of a legend in the computer industry by designing the PalmPilot, which became a sensation when it hit the market in 1996. At a time when PCs and software were steadily becoming more complex, the PalmPilot turned the model on its head by limiting itself to a few functions, which it did very well. This summer, Hawkins left the Palm Computing division of 3Com Corp. to start a company that will develop new products based on the Palm platform (see BW Enterprise Online, 8/28/98/, "After the PalmPilot, What Do You Do for an Encore?"). Here, Hawkins offers his thoughts on design in response to the Sept. 28 Technology & You column on IBM researcher Clare-Marie Karat's proposed User's Bill of Rights.

It is amazing that the emblem of the Information Age, the Windows PC, is so fault-ridden, confusing, and frustrating to use. I design computers for a living, but I still find my Windows PC crashing several times a week and doing weird things that I don't understand. I shudder to think what it is like for nontechnologists.

The computer industry is mostly to blame for this failure. Consumers also deserve some of the blame. An analogy can be made with airline service. Everyone complains about airline food and tightly packed seats. However, it has been shown time and time again that people buy the lowest priced tickets. They aren't willing to pay slightly more for better service. Windows-based PCs are inexpensive, and that's why people buy them. The Macintosh was a better product but more expensive.

The way out of this mess is to design more reliable, easier-to-use products that cost less, not more. This is what I attempted when I designed the PalmPilot. The PalmPilot has succeeded because it reliably does what it claims to do, is relatively easy to use, and is relatively inexpensive. Here are some of the rules I use when designing.

1. Eliminate options. Often, engineers and designers can't agree on the best way to perform a task. When this happens, their first tendency is to make it an option, allowing both methods. An example of this is the "overstrike" and "insert" modes for entering text from a keyboard. How many times have users accidentally hit the "Insert" key and not have a clue why text starts disappearing as they type! An abundance of preferences and options makes great feature lists on packages, but these "features" are confusing to users and rarely used. Good design is like good editing, eliminate the extraneous, leave only the essential. It is difficult to do.

2. No waiting. I banished the hourglass "wait" cursor. PalmOS doesn't have one. If we found an operation took too much time we either eliminated the feature, redesigned it, or solved the problem a different way so the user doesn't have to wait for common operations. It is better not to do something than to do it poorly.

3. Underpromise, overdeliver. It's tempting to tout all the capabilities of your product, no matter how marginal. Users are happier to find that a product reliably does what it promised -- and then discover that the product does more than expected. If their expectations are set high, they will only be disappointed.

4. Eliminate errors and error messages. It would be welcome if all error messages were comprehensible and explained in detail, but that treats only the symptom. The real challenge is to eliminate errors in the first place. Although not completely possible, it's surprising how much can be done. Better design and more thorough testing go a long way. Whenever possible, I would dictate that no error message will appear and that the system should do the best it can to recover from a problem. After all, if users cannot understand an error message or cannot do anything about the error, why bother showing it to them? A meaningless and cryptic message is worse than none at all.

Twenty years ago, Japanese car manufacturers introduced more reliable and less expensive cars. This forced U.S. auto makers to improve their quality. The same will eventually happen in the personal computer market. Less expensive and more reliable products like the PalmPilot will eventually force Microsoft and PC makers to improve. But as we saw with Detroit, don't expect it to happen overnight.

News Flash Archives

Copyright 1998, Bloomberg L.P.
Terms of Use