|
How did Windows end up in the security mess it's in now? There are historical, economic, and sociological factors at play here. This week, I try to upstage James Burke in enumerating those factors, and arranging them in chronological order.
The human autonomic system regulates the many various necessary functions
of the body without the direct intervention of the intellect. It manages your breathing, your heartbeat, your kidney functions, and my writing. Without the autonomic system, you would have to intellectually marshal all the thousands of body management functions with the aid of only your conscious mind. Imagine how difficult the regulation of the distribution of white cells in your bloodstream would be, for instance, if you had to actively consider that function every day your body was invaded by some foreign pathogen. If maintaining every service your body provided for itself were a conscious decision on your part, could you possibly fathom the decision-making process behind each breath, each pulse? You'd want to invent the computer just so it all could be done for you.
For several years, IBM has wondered, what if large systems could be programmed to self-diagnose and self-remedy? Since Ross Perot no longer works for IBM, we can assume the company has not concluded that the result would be the loss of more American jobs. Now, we've talked a lot in these Web pages about component-oriented software. So you should understand how IBM caught itself thinking like this (and a corporation catching itself thinking is scary in itself): Already, one of the chief purposes of modular or component software, such as COM software systems (Windows), is to log errors as they come across such errors, automatically. So such components are already as close as software can come to the source of problems. With a wealth of error data already available to them,
suppose such components could also be capable of making diagnoses, automatically. Then to go one step further, why is it really necessary to bother IT managers with invoking the solution to those problems, especially when more often than these managers themselves can count (surveys have revealed that IT managers do, indeed, count), the solutions generally involve repeating the same remedial methods by rote? In The Need for Autonomic Computing, Stephen Morris studies the concept from a programmer's perspective. Can the protocols for tracing and logging errors be refined in such a way as to facilitate automatic remedial measures as the next logical step?
Speaking of really big errors: The 2000 presidential election is an error I'm not willing to see repeated anytime soon. Regardless of for whom we voted, we all learned to our dismay that the value of any one person's vote is a variable. Believe it or not, there's a science that pollsters and political strategists use to calculate the value of that variable. My colleague Chris Katsaropoulos studies those who have studied how much a parking ticket in Florida diminishes the voting power of the Floridian, in The Battleground: Swing State Campaign Strategy and How It Affects Your Vote. Of course, this hyperlink comes with my dutiful reminder to you to do your duty and vote this November. With the aid of this article, you'll be able to move yourself, your family, and your voting friends to a precinct where all of your opinions truly do matter. One day, when the returns from the pre
cincts are all tallied, and nobody appears to have voted in the "non-counting" states, an IBM autonomic procedure will automatically clean up the mess, wipe several non-counting states off the map, eliminate the needless state governments and trash collection services apportioned to those unoccupied areas, and leave only the 23 most efficient states to co-exist with one another in peaceful, democratic equilibrium. Fox News wouldn't be able to conjure election results from out of the blue anymore. That in itself would be a payoff.
|
