September 16, 2003

Not everything is easy

Somewhere, somehow, we told people that everything about computers should be easy and intuitive. That you shouldn't have to learn anything, or read manuals. That you should be able to grasp everything in ten or fifteen minutes.

What nonsense.

Some things just aren't easy. Quantum mechanics. Tensor calculus. Navajo verbs forms. Old Norse. Getting rich.

Yesterday, I received mail from a stranger about a fairly esoteric Tinderbox design question. The issue is interesting -- it involves scope rules for UI design of hypertext composites -- and lies at the very edge of current research. It could, for example, easily be the subject of a top-rated research paper. As far as I know, my correspondent isn't a computer scientist -- he might not even have any computer science training at all.

This is one of the great things about computer science right now: you can walk in off the street, roll up your sleeves, and with a little hard work and fortitude you can be right at the research frontier.

But the terms of the correspondence were also all too familiar: since the interface didn't work intuitively, we must all be idiots or scoundrels.

As far as I know, only two research groups have ever thought about the question. As far as I know, nobody has ever discussed it in a journal. This is one of the sad things about the computer world right now: everybody knows better, and hardly anybody seems willing to do the work.