Think of Y2K as a 'chronic condition'

greenspun.com : LUSENET : TimeBomb 2000 (Y2000) : One Thread

Anchorage Daily News Business News ----------------------------------------------------- ANALYSIS: Think of Y2K as a 'chronic condition'

Copyright ) 2000 Nando Media Copyright ) 2000 Associated Press

NEW YORK (January 4, 2000 8:54 a.m. EST http://www.nandotimes.com) - The biggest risk from the Y2K bug was never to power grids, missile systems or telephone exchanges. Instead, it was to the complicated backroom systems on which the world's corporations and governments run.

That's why the vast majority of Year 2000 computer problems won't turn up for days, weeks or even months, information technology experts say.

So forget the somehow widely disseminated misconception that if planet Earth got past Jan. 1 without any info-disasters we'd be home free.

Think not of Y2K as an information age earthquake avoided but rather as a steady stream of gradually more damaging tremors to come.

For early examples, consider a few of the failures from Monday, the first U.S. business day of the new millennium:

Driver's licenses could not be issued in nearly half of New Mexico's motor vehicle offices.

A vital payroll computer died at an Alabama company.

Doppler weather systems shut down for a few minutes in Chicago.

A small part of a Danish bank's payment system was erased. Millions of small- and medium-sized businesses worldwide have done little or nothing about Y2K and will fix on failure.

"Now is the tough time. The next few months are going to be the toughest Y2K time," said Dale W. Way, bug point man at the Institute of Electrical and Electronic Engineers.

At greatest risk in the private sector are the accounting, inventory, invoicing, billing and other systems integral to survival - a crazy quilt of interconnected programs often cobbled together over decades.

Such "custom applications," also common in government agencies, are nothing like the control systems at power and water plants, which are typically spare, easy to maintain and fortified by built-in redundancy.

Instead, they tend to be a mishmash of different software languages riddled with updates and patches applied over decades like digital duct tape that experts say make them especially susceptible to Y2K errors.

"Every large company says they have some software that they run routinely without even knowing what it does," Way said. "They are afraid not to run it - because they're afraid of screwing up if they don't."

That's why tampering with such code to try to purge it of Y2K bugs can often introduce new unrelated errors. Programming is an art, not a science, and not all programmers are Picassos.

The more complex a system, the more difficult to repair it safely without introducing new errors.

In administrative and accounting software, errors can show up in many ways. Systems can simply crash computers, fail to process certain data and lose it, make an incorrect assumption and corrupt data or destroy data completely.

Errors also may not occur until triggered by a particular event.

"A program whose job is to track the pressure in a chemical plant's boiler ... may not activate until a certain temperature is reached," said Norman Dean, director of the Center for Y2K and Society, a Washington public interest group. "And that may not happen until next week and it may not happen until next year."

Bug-infected systems are apt not to blow up but rather degrade over time - linked in many cases to monthly report generation or billing cycles - and often not even be easily identifiable as Y2K-induced.

Robert X. Cringely, a Silicon Valley commentator, predicts that "Y2K effects will linger far past January as a patina of rust" on information systems.

Or, as the IEEE's Way puts it, the threat has now passed "from systems with low intrinsic risk and high repairability to ones with high risk and low repairability."

For an idea of the complexity of Y2K fixes in business, consider the challenge to Mastercard International, which had to scour 7 million lines of code in at least 10 different programming languages ranging in age from one to 25 years. Or AT&T, which examined 385 million lines in 3,500 systems and applications.

Neither company has so far reported any Y2K-related failures.

Yet while U.S. states and foreign governments dismantled Y2K bunkers Monday and a World Bank-funded international Y2K clearinghouse in Washington canceled all further press briefings, overjoyed that glitch damage has so far been minimal, big companies remained vigilant.

"Our feeling here at AT&T is that we won't close the book on Y2K until Feb. 29," said company spokesman Dave Johnson. "First of all, we need all our billing systems to run a full cycle and then we want to take a close look at the leap year,"

Leap year? This year has 366 days, while 1900 did not.

The Y2K computer problem will not simply go away.

Because of the extra day issue, it will even nag us on Dec. 31.

As programmer Lane Core likes to say in Internet columns that harangue what he considers mainstream media's simplistic coverage of the issue: "Y2K is not a one-time event. It's a chronic condition."

By FRANK BAJAK

Anchorage Daily News - link



-- Teague Harper (tharper@cyberhighway.net), January 04, 2000

Answers

Gotcha. Good face saver. Excellent, in fact.

-- Jim Thompson (jimthompsonmd@attglobal.net), January 04, 2000.

Oh, good, here's Doctor Jim again.

There's a pattern to his posts. His advice to us all is given with the admonition "memorize this", and his evaluation of a problem is always couched in terms of saving (or losing) face. So for this doctor, memorization is a perfectly acceptable alternative to learning, and the appearance of success is at least as importance as success itself.

Aren't you glad he's not YOUR doctor?

-- bw (home@puget.sound), January 04, 2000.


Has you ever noticed that supercilious morons like Jim Thompson are always reactive and never contributive? The true indicator of the mindless, feeling safe in the mass.

Just watch his future comments - they all follow exactly the same knee-jerk pattern.

I'd say that reflects some deeper problems, myself, but we need more evidence. And Jim will surely give us plenty of that...

-- John Whitley (jwhitley@inforamp.net), January 04, 2000.


My prognosis is much more optimistic about the "chronic condition". Unfortunately, I didn't return the reporter's phone call before he had left for the day.

-- Lane Core Jr. (elcore@sgi.net), January 04, 2000.

Moderation questions? read the FAQ