Long Road to Simple Systems

The voice on the telephone sounded as if its owner couldnt decide between rage and exhaustion. “My machine re-booted while I was working, and now most of my e-mail in-box is gone. I thought you could only lose work that you hadnt saved!”

I had to break the news that behind the scenes, his e-mail client program—which was open when the system crashed, since he was responding to new messages at the time—had probably corrupted a weakly structured in-box file that had not been properly closed when the system went down. Its just the latest example to cross my desk of the ways that a users simplified, everyday mental model of IT has to struggle against the ugly complexity of the actual hardware and software.

Or, alternatively, its the latest example of IT makers failure to think about how users want to interact with their products and to make their products behave correctly when used in those most likely ways—and to tolerate their most likely failure modes. It makes you realize how immature IT still is and how much work remains to be done—in the same way that other technologies have needed decades of work to make them usable, even after all the effort that went into making them work at all.

To put things into perspective, after this incident, I found these 1915 instructions to new-car owners from an archive compiled by James Higgins at The Detroit News. “Remove the tires and jack up all four wheels. The inner tubes should be removed from the tires and placed in a box. Grease all parts which are liable to rust and do not overlook the motor parts such as the spark plug ends, the flywheel teeth, all chains and open gearing, etc. Remove the battery and let the service station take care of it for you. …” Today, with improved materials and design, we expect our cars to sit outdoors in all sorts of weather and be ready to go at the turn of a key.

Services-based applications, scalable and always up-to-date, will be harder to write than stand-alone programs—just as automatic transmissions are harder to build than stick shifts. Robust database foundations, rather than fragile file systems, will add design complexity in the interest of run-time safety—just like anti-lock brakes. These are some of the kinds of foreseeable, and necessary, improvements that will bring us closer to equal convenience in IT.

Tell me what users need most at peter_coffee@ziffdavis.com.