Computers Versus Humans

Software is more like a bridge than an edifice. Although it runs on high-technology microprocessors, it must be operated and used by mere mortals. Amid all the attention and excitement about new technology, we overlook the incredible differences between computers and the humans who have to use them.

For example, because computers have memories, we imagine that they must be something like our human memories, but that is simply not true. Computer memories work in a manner alien to human memories. My memory lets me easily recognize the faces of my friends, whereas my own computer never even recognizes me. My computer's memory stores a million phone numbers with perfect accuracy, but I have to stop and think to recall my own.

For software to be robust and strong, it must be written in perfect harmony with the demands of silicon. For programmers to be professional, they must also work in this same harmony.

For users to be happy and effective with software, it must be written in harmony with the demands of human nature. The problem, of course, is that those human demands are so radically different from the demands of silicon.

Clearly, one side of software the inside must be written with technical expertise and sensitivity to the needs of computers. But equally clear, the other side of software the outside must be written with social expertise and sensitivity to the needs of people. It is my contention that programmers can do the former, but it takes interaction designers to do the latter.

Computer industry guru Jerry Weinberg says, "Once you eliminate your number one problem, you promote number two."^[2] For decades, the computer industry's number-one problem has been efficiency. Computers were relatively speaking small, expensive, slow, and weak. We lionized the hacker-gods who could make programs that operated as efficiently as possible so as to maximize the productivity of the expensive mainframe computer. Essentially, it was far cheaper to train people to deal with obscure but efficient software than it was to buy more computers. The driving inevitability of plummeting computer costs has utterly obliterated that problem. Today, it is far more expensive to pay for the human costs of adapting to "efficient" software than it is to make the software conform to the expectations of the humans.

^[2] Gerald Weinberg, The Secrets of Consulting: A Guide to Giving & Getting Advice Successfully, Dorset House, 1985, ISBN 0-932633-01-3.

The solution is obvious: Make the software serve the users. But standing in the way is the culture we've so carefully built over the last 50 years that puts the hacker-gods in the driver's seat. The community of software engineers is generally willing to accept interaction design into the process. They say, "Sure, wait until we're done, and then do all the design you want." Unfortunately, design has to come before construction, so the programmer's openness to design is largely ineffectual. It's like a cement-truck operator telling the carpenters they can build all the forms they want as soon as he is done pouring the concrete.