I don't do windows. Let me rephrase that: I will cheerfully wash windows -- I simply won't deal with them on my computer.

You know windows -- that software split-screen technique that lets you handle a spreadsheet on one part of the display, a word processing document on another side and that project scheduler on the third side. (Kind of reminds me of that old joke where the huckster says he can divide that wonderful-looking chocolate cake into three halves.)

Assorted PC operating systems offer windows as an integral part of the package. The designers, in their seemingly relentless quest to squeeze more and more "functionality" (read that as "goldplating") into the software, offer things like windows in the misbegotten belief that they actually will be used.

A casual, thoroughly unscientific but revealing survey of some dozen people I know who use computers indicates that only one uses more than two windows. (The guy is an engineer, OK?) Virtually everyone else simply uses a single split screen to hop back and forth between spreadsheet and text or text and text or spreadsheet and spreadsheet. In sum, only two windows at a time, please.

Why don't people like multiple windows? The answer should be obvious. The darn screen's too small to be divvied up into four or five pieces. Moreover, screen resolution is sufficiently poor that windows ain't the way to view your data world without catching a nasty case of eyestrain.

Admittedly, a high-resolution screen such as the one available on the Macintosh makes multiple windows more appealing -- but guess what? The Mac users I talked with rarely use more than two windows either!

What can one conclude from this?

The obvious conclusion is that most people still don't think in multiple-display windowesque terms. The other is that there may not be the computer equivalent of Say's Law. Say's Law is that "supply creates its own demand." Say has become real popular in this age of supply-side economics. Alas, computer programmers hoping to piggyback on Say are discovering that most of those nifty functions they design into their programs aren't even being touched -- let alone actively used.

Why pick on windows? Because they are, in my mind, the classic example of something that sounds terrific in concept but works rather poorly in practice. Windows illuminate the technology-driven rather than the utility-driven nature of most PC software.

Let me resurrect the name of another long-dead economist -- the Italian Pareto -- who is known for the Pareto optimum. The Pareto optimum, you may recall, also is known as the 20-80 rule. That is, 20 percent of the participants are responsible for 80 percent of the activity. To put it more crudely, in any organization, 20 percent of the people are responsible for getting 80 percent of the work done. Similarly, in computer programs, 20 percent of the functions do 80 percent of the work.

Why do I keep quoting economists? Do you know how much all this software costs? The reason is that designers are ignoring Pareto. People are building in technical functionality rather than practical functionality.

I will harp on this theme until my word processor short-circuits. Virtually all the software that passes over my desk for "review" fails fundamental tests of practicality. It is designed by people who are good at designing computer systems, not human systems.

Here's an experiment: Go to your favorite piece of software. Use it. Count the number of different functions in the program you have to use to get something you want done. Then compare that number to the number of functions your documentation lists the program as having. Note the discrepancy. Then ask yourself, how could the software be redesigned to let you deal with your most-used functions most easily?

Some day, more software designers will take that last question more seriously. For now, there's a lot of software on the market (and on my desk) that isn't worth the floppy discs it's printed on.