May 1998

Let My Computer Serve Me!

Forget the multimedia distractions. Computer developers should strive for simplicity, eas of use, and machines that are "people-aware."

By Michael Dertouzos

I'm just back from the world economic Forum in Davos, Switzerland, where besides Hillary and a few hundred ministers and prime ministers, there were also some 1,500 CEOs representing $4 trillion of gross revenue-among them the leaders of nearly every computer, software, and communications company, including Michael Dell, Bill Gates, and Larry Ellison. In a session I chaired among these executives, we sought to determine the top issues ahead for the developers of information technology. At the top of the list: Ease of use.

Ease of use-a term I much prefer to "user friendly"-is one of several important issues at the junction of our technology and our humanity that we will explore in this new column.

A machine would be simplest to use if it were intelligent. Never mind if it is swelling with multimedia and colorful three-dimensional animals floating in space. Just give me a keyboard with an impact printer that understands me as well as my human assistant does, and I'll throw away all the fancy hardware and software. Well, that's impossible today and it will be in the near future, at least for the kind of general intelligence we call common sense. We don't know how to do it in practice or in the lab (which precedes practice by a decade).

We can, however, make our systems more intelligent by specializing them into particular, narrow contexts. In fact, it's scandalous that we haven't done so yet. In the industrial era we gave different tools to different workers-doctors, mechanics, plumbers, gardeners-but in the information age we are trudging along with the same word processor, spreadsheet program, database, and graphics editor for musicians, accountants, engineers, and lawyers! This mindless generality must give way to greater specialization among the different human interests and professions. With the right tool, it will be easier to get the job done.

Following specialization, we need to create interfaces tailored to humans. People are not born with keyboard sockets or mice attached to their bodies. Instead, we use our mouths and ears. Why not engage our machines with the same natural human interfaces? Speech-understanding technology has come a long way since researchers began prematurely touting its capabilities. A few useful systems are already commercially available.