There's this essay I'd like to write someday, all about how corporate decisions are made and how they are subsequently assessed; and about the role of luck, fortuitous timing, and amazing consumer tolerance in the success of Apple Computer.
The title of the essay would have to be "The Apple Fan," with a double meaning and a twist of irony in each of the meanings.
The essay would tell the story of two products: the 128K Macintosh, cute and cuddly, the triumph of concept and style over performance; and the power Mac, a composite of the Plus, SE, and II.
The absence of a fan in the original Mac would serve as a model for Steve Jobs' insistence on keeping the Mac from looking, sounding, or feeling like a computer, a goal that could only be attained by keeping it from being fully a computer. Thus the fan would also symbolize the serious performance limitations of the early Macintosh.
The essay would portray Mac users symbolically demanding "We want a fan!" and show that the users could then hardly object when the more powerful, fan-equipped Mac they had demanded turned out to be less friendly, less cuddly than the family pet they had initially taken to their hearts. Alter all, power comes at a cost. And the essay would show Mac SE and II users accepting their computers as being in every sense Macintoshes, but with the power of a "real computer."
Could Apple have succeeded as it has if it had delivered a more powerful but less conceptually pure computer in 1984? Or even if it had delivered the Mac II back then? The essay would argue that it could not, that the sequence of "toy" Mac and "real computer" Mac was the key to Apple's current success, because it opened a market and created a specific demand for a more powerful Mac -a machine that only Apple could produce and that Apple could easily produce.
And the essay would argue with malicious delight that the sequence was nobody's decision but merely the accidental result of the successive ascendencies of different factions within the company. A historical accident of the "hardware wars."
One of these days, I'll get around to writing it.
Even Time magazine is warning us about computer viruses. That's all we needed; now the obnoxious pranksters can aspire to make the cover of Time. I deplore the actions of the virus creators, but the real dangers they pose are one that, if encouraged, the practice might become a fad, and two that the general public might become so alarmed as to support crippling of the hardware or operating system or legislation that threatens our civil liberties. Time magazine and others who promote virus hysteria encourage both these possibilities.
One of the most cogent thinkers on civil liberties in the information age is Dean Gengle, who has graced these pages and who reminded me recently about the real danger of computer viruses. I thank him for that, and I thank George Bush for reminding me to renew my ACLU membership.
And while I'm thanking people, let me thank the editor of a short-lived but much-appreciated programming magazine, Turbo Technix. Jeff Duntemann is one of the best writers and editors in the industry. In the October issue of the newsletter he sent to his contributors, he announced that the SeptemberOctober issue of the magazine would be the last, in the interest of Borland's profitability picture. It is characteristic of Jeff that this newsletter, with an audience of a few dozen, was better written than most computer magazines.
Do you get the DAK catalog too? "Forget high stereo prices. Forget changing channels by hand." There's a man who gets his money's worth out of his word processor. Anyway, forget computer viruses. The real menace is exploratory data analysis.
Exploratory data analysis is a legitimate statistical tool, but imagining it in the hands of the average PC user is scary. Here's why:
Using statistical techniques requires recognizing the underlying assumptions. Some statistical techniques are robust to violations of assumptions, and can be applied loosely; others are less forgiving. Few people who work with conventional statistical techniques know the assumptions.
Exploratory data analysis invites a whole different category of assumption violation: the confusion of the context of discovery and the context of justification. Data used to come up with a hypothesis is not valid to use in supporting the hypothesis. Obviously. Just watch that obvious assumption be violated right and left as naive users play around with exploratory data analysis.
Thank You, Thank You, Thank You
Exploratory Data Analysis