Y2k bugs will linger all year, experts say

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

http://www.seattletimes.com/news/technology/html98/ytwo_19991017.html

Link

Nation & World

Posted at 12:33 a.m. PDT; Sunday, October 17, 1999

Y2K bugs will linger all year, experts say

by David Judson Gannett News Service

WASHINGTON - While often billed as a one-day, make-it-or-break-it proposition, the so-called "Y2K" computer bug more likely will linger well into the year 2000 or beyond.

Most likely to fail first: older fax machines, electronic heating and cooling systems, perhaps even processors key to oil refining and the chemical industry - all of which depend on hard-to-find "embedded" computer chips.

But most problems - maybe 90 percent - won't show up until much later in intricately linked computer systems that regulate air-traffic control, banking, the mailing of government checks and countless other business functions.

"Y2K preparation is often perceived as a flat race to the Jan. 1 finish line, when in fact it is more like a steeplechase," said David Henry, the Y2K coordinator for New Zealand, the first densely populated country to enter the new millennium, 18 hours ahead of the eastern United States.

In reality, Henry said, "There are plenty of hurdles to get over both before and after the new year."

Gartner Group technology consultants have gone so far as to estimate that fewer than 10 percent of problems will occur at the new year.

Gartner, among the first to sound the Y2K alarm in the mid-1990s, projects that 25 percent of failures will occur before Jan. 1, half will be spread throughout 2000 and about 15 percent will not occur until 2001.

"This is definitely not going to be a two-week problem," said Lou Marcoccio, Gartner's point man on Y2K.

Consider: In Maine, the state in October accidentally registered 2,000 new cars as antique "horseless carriages" when a computer mistook the model year 2000 for 1900 and assumed the new cars were made before 1916, the cutoff date for antique designation.

One major worry for Y2K experts such as Norman Dean, head of Washington-based think tank Y2K and Society, is that incomplete repairs on secondary computer networks will spill over into primary networks, where most repair work has been focused.

"My fear is less the dramatic event on New Year's Eve than it is a slow accumulation of problems affecting daily life that build up when we are not expecting it," Dean said.

Specific problems are all but impossible to predict, he said, but could range from delayed mailings of government checks to lost hotel reservations. Spillover problems could lead to credit-card rejections, despite the banking industry's highly publicized repair efforts.

While basic phone systems, the focus of intense Y2K repair work, are likely to be fine, for instance, computer systems that monitor customer complaints may not be, Dean said. And it is those systems that, for example, direct repair crews to fix lines downed by trees. If those secondary systems fail, problems in the phone network that have nothing to do with Y2K may go unheeded - days, weeks or months after Jan 1.

Marcoccio predicts problems mounting in day-to-day business tasks such as six-week sales forecasts, reports on future orders or revenue projections as so-called "date-forward" calculations begin tripping over unseen computer glitches. Similar problems could show up in many government systems, he said.

And some experts worry that short-cut Y2K repairs will create more problems later, and cost even more money.

The millennium bug can occur if computers that use two figures to designate a year mistake the year 2000 for 1900 and shut down or produce erroneous information.

The most common repair approach is called "windowing" - which does not convert two-digit program date fields to four digits. Rather, it simply instructs computers to read some dates - say years ending in 01 through 20 - as being in the 21st century and everything else as being in the 20th. That works fine, for example, when an accountant is plotting the retirement benefits of someone retired in 1990 through the year 2020.

But various Y2K repairs have relied on different "date windows," meaning information will be damaged or lost when computers set to different standards attempt to communicate with one another. Those problems will grow worse over time and will necessitate future rounds of costly repairs, said Kappelman, an author on the topic.

"We've attempted to solve the problem by adding to the complexity of computer systems," Kappelman said. "In many ways, we haven't really solved the problem but we've only postponed it and we may have made it worse."

Copyright ) 1999 The Seattle Times Company

-- Homer Beanfang (Bats@inbellfry.com), October 18, 1999

Answers

I very much agree with the end of this article about windowing. Windowing is a horrible "solution" to Y2K, and I believe that what Kappelman says is true: that there will be all kinds of problems with date windowing "standards." Windowing only postpones the problem. Remember, all the Y2K computer work being done is to fix an existing windowing problem - the one with 00-99 being in the 20th century.

That being said, I also feel that expansion will not be maintained throughout the 21st century. Once we get through Y2K, and we get on to the 20's, 30's & 40's, everybody's going to be saying "why do we need two digits anyway?" and we'll be on the way to the "Y2100" problem.

Best solution is for ALL computer internals to store the date year as 4 digits (expansion) with a STANDARDIZED sliding window routine for data entry and input so that humans can be lazy and only enter 2 digits. This routine MUST also allow for century override so that any date can be entered in any century if it is outside the current window.

-- Jim (x@x.x), October 18, 1999.


More backpeddling in store I think...from the doomers that is!

Lot of difference between bugs that show up all the time anyway...even if they are "new" Y2K bugs...that are an irritation...and TEOTWAWKI..or the "collapse" of power grids or whatever. Fact is lots of the predictions were for LOTS of trouble before now. Anybody who wants to argue just go back and read Yourdon's book. Remember April 1st? October 1st? Came and went..with nary a peep.

Basically, the predicted disasters just haven't happened. I do think there will be Y2K problems next year, just no TEOTWAWKI or "collapse". I make guesses just like everybody else -- and don't care what any of you do to prep...I'm not against "prepping"...I just tire of the stupid drivel from some of the doomers here. Only know about the systems I work on...and they test out fine. But....problem is....doomers wont accept any good news because it's not what they WANT....that's the problem!

So, what do they do, now that no more "trigger dates" have been announced? Mainly watch the stock market, hoping for a fall. How pathetic. I don't care what the stock market does. The issue is Y2K...and the doomer predictions haven't happened...yet....and thus they now say "wait till next year!"...and ....fearing the world just might survive new year's day...."wait ALL OF NEXT YEAR!"...now we're into decades...year 2020 will blow!

I'm a programmer/analyst. Know all about year end processing...and how some problems might "compound" and show up on annual reports and job runs that execute in early 2001. But...again...lots of difference between some reports that are screwed up or other minor irritants....and problems that would bring about collapse of society or even shortages and/or cause portions of our infrastructure to crash. Think about that example of those people whose cars were designated "horseless carriages"....I'll bet this glitch didn't stop a one of them from getting their vehicle licensed and on the road. Just an aggravation...not a show stopper. Bet the people at DMV were able to work around the problem...and even laugh at it...as they FIXED IT...like organizations do! Bet nobody ran for the hills to await DOOOOOOOOMMMMMM!!!!!

Think people think! Get a grip! How about some sense of proportion....instead of posting any computer fart as an "evil wind that portends TEOTWAWKI"....or is that too grown up for some of you.....or maybe just beyond your level of technical understanding. And don't try the 1000 cuts argument.....bet there are many more than 1000 computer glitches every day...probably over a million...and they get fixed, life goes on, and...here's the key...all "1000" cuts don't happed to any one person...and...capitalism IS redundant, as Decker points out. One burger joint goes belly up, and you just go to the next one's drive thru! Isn't society wonderful! YES!!

Come on people. Accept reality...whatever it brings....whether you are right or wrong. Just don't be mealy mouth......and backpeddle...or try to insist you were "right all along" no matter what next year brings. Of course, for some of you, this is too much to ask, isn't it? Well, start reading more doomer stuff so you can have another reason to expect doom after 2001 comes and goes. Emerging plauges? Terrorism? You'll think of something, I'm sure!

-- Genius (codeslinger@work.now), October 18, 1999.


more like linger for 15 years or more

-- code slammer (last@brickby.midnight), October 18, 1999.

Duuuuuuuhhhhhhh!!

Thanks for that hot tip, Mr. "Expert"!

Sheeesh, you really need to be an expert to figure that out!

-- @ (@@@.@), October 18, 1999.


Whoa, CodeSlinger (aren't you DominantSniper, too?), take a breath.

Doomers don't accept good news, true enough, but being GI is not the same as being Doomer. I'm prepped not because I want the world to end, but because I recognize the risks. I've posted good news here when I find it, but it's pretty thin on the ground. Gary North aside (now, THERE'S a Doomer) I haven't seen anyone who sounds like they really LIKE bad news. But I see many who are hypersensitive to the news that will trigger a panic, and the panic is a given. Scoffers like you, and self-reported compliance statements, have pretty much guaranteed that. Perhaps you're mistaking hypersensitivity for enthusiasm.

For a programmer to say "no more trigger dates have been announced" makes me wonder how good a programmer you are. The dates don't get "announced", we thought them up a couple years ago, sitting around talking about the Y2k problem, trying to figure out what might break. Weren't you in on those meetings? We aren't out here waiting for the next trigger date to be "announced" like this week's Lotto winners.

The problems with any one date have been low, so far. That's what we in the computer business call a "good thing", though a few failures might have awakened a few more people. Most would be delighted if nothing at all broke, but we just won't know for a few more months.

It's not backpedalling to keep rethinking your forecasts - weathermen rethink their 6-day forecast 5 times before it arrives! NASA didn't do enough rethinking, and shot their last orbiter right into the planet. I'll take a little rethinking, any day. Try rethinking your own view about what motivates GI's.

-- bw (home@puget.sound), October 18, 1999.



Moderation questions? read the FAQ