Saturday, December 20, 2003

Telirati Newsletter #30

He shoots! He scores! This is one of the best things I ever wrote.

Telirati Newsletter #30: Why tu que?

We are shaped in our outlook by our geography. The north woods of Maine offer a desolation unimaginable to a Western European city dweller. The American west is even more impressive, and the emptiness of some quarters of it gives rise to the romantic notion of escape and refuge. The latest group to pack their pickup trucks with guns and mean dogs and cast their gaze toward the distant hills is year-2000 alarmists.

Positing a frightening cascade of system failures, the alarmists predict global meltdown of economic activity and social order. Most of these alarmists are American. Not coincidentally, I think.

The possibility of running away from a collapsing society and the prediction of such collapse are linked. Whereas a client of mine from the Netherlands pointed out that there they have a) Not very many guns; b) No hills; c) Dogs that are mostly the terrier rug-rat sort; d) No place that is very far away. The Dutch aren’t going anywhere when the Y2K bogeyman comes, except maybe down to the pub for a beer until the computers are rebooted.

The cascade of failures is one problem with Y2K alarmism. Another is that most industrial uses of computers are often no less useless than most desktop deployments. They are just more expensive. What we will find when we reach the millenium and a number of elderly systems crash is that many computers are employed foolishly, on tasks that barely require automation, much less the maintenance of costly legacy systems.

Another thing we will find is that the same methods that work when computers fail now will work when computers fail due to the Y2k bug. A lot of the U.S. air traffic control system runs on alarmingly antique systems. It breaks often. There is a manual backup method or controlling air traffic when the automated system breaks. It is only slightly less alarming than an antique computer, and involves slips of paper on a note-board. But it works when the antique computers break.

It drives my wife nuts when it takes me days to set the date and time on the oven when we have a power failure, but I find clocks that read wrong not very disturbing. Ever hear of the Y2K fax machine problem? Yes, lots of fax machines will start screwing up the dates on faxes. Got to trash the lot of them. The fact is, most of these buggy fax machines will continue to be used, wrong dates and all, until they die a natural death. The “service” light on your dashboard? A square of black electrician’s tape will take care of that more efficiently than replacing the car’s computer.

The enterprises that derive a real competitive advantage from IT have solved their Y2K problems and will be quick to mop up the unforeseen ones when the time comes. Which is why Wal-Mart will be up and running, and their cheesy competitors still won’t have that plastic sandwich container I was looking for, only they will blame it on the Y2K bug and suggest I should have hoarded them when I had the chance.

I was once asked by the proprietor of a bartending school how powerful a PC he should buy to automate his database of alumni. I asked him how many database records he needed to keep track of. He pulled out a stack of index cards. There were about 250 of them. I nodded and decided that I had to ask some more questions to really understand his computing needs, like: “Do you like those 3-D shoot-em-up games?” His academy is in more danger from too much lab work than the Y2K bug.

Even systems that make good use of computers, like just-in-time (JIT) manufacturing supply chains, generally have non-automated precedents. Some Japanese implementations of JIT rely on color-coded parts bins, blackboards, and slips of paper. When the parts bin is brought to the supplier, he fills it. Relatively few industrial processes involve so many operations that must be tracked or calculated in detail as to actually depend on computer automation for their existence.

There is one effect that is underestimated: A number of hardware and software companies are either blaming a current downturn in sales on companies applying their resources to the Y2K bug rather than to deploying newly purchased systems. Other companies, notably voice mail companies, are benefiting from customers deciding to finally scrap antiquated systems in their Y2K update process. This effect will not decrease after January 1, 2000. It will increase as Y2K bugs that are not fatal, but are deeply annoying, drive an intensified Y2K bug hunt. This means continued hard times for big, expensive, hard-to-deploy systems, another nail in the coffin of mainframes that can be replaced by commodity servers, and a continued up-tick for purveyors of moderns, standards-based systems that can be quickly deployed.

To sum it up, we will learn that many computer systems are now misapplied, and could be done entirely without. We will learn that some computer systems are needed, and used, and well taken care of, and that there is a real benefit derived from these systems. And some situations where ancient systems have been kept on in fear of change will find the Y2K bug driving modernization. Finally there will be genuine botches, where some businesses will fail due to the failure of their badly run IT. But it isn’t the end of the world.

Copyright 1998 Zigurd Mednieks. May be reproduced and redistributed with attribution and this notice intact.