Published Sunday, April 18, 1999 Y2K problems will linger for 30 months Steve Alexander / Star Tribune The script for Y2K is all wrong. Instead of Y2K failures occurring at midnight on Dec. 31, they will be spread over a 30-month period that will begin in July. So says the Gartner Group, a leading computer industry consulting firm based in Connecticut. This is revisionist stuff, since practically all Y2K scenarios to date have focused on the potential for serious computer failures at midnight on Dec. 31. But Gartner maintains that only 10 percent of Y2K failures will occur within two weeks of Jan. 1. In an interview, Dale Vecchio, a research director in Gartner Group's Y2K practice in Stamford, Conn., said his firm has been eyeing a multi-month Y2K impact since the middle of last year. But Gartner analysts have been frustrated that the public debate about Y2K has continued to focus on the Jan. 1 millennium date change, he said. "People haven't quite gotten it," Vecchio said. "They have been too focused on the single time boundary and not enough on what they may have to deal with before the time boundary occurs." Why should Y2K start in 1999 and stretch into 2001? In a recent speech in San Diego, Lou Marcoccio, another research director of Gartner's Y2K practice, said the causes will be forecasting software that looks six months into the future, the beginning of new fiscal years for many corporations and some "date-related anomalies in software code." The number of Y2K failures will increase further in October as forecasting software that looks three months ahead runs up against the Jan. 1, 2000, date and still more companies begin new fiscal years, Marcoccio said. In Gartner's view, 25 percent of Y2K computer failures will occur in 1999, 55 percent will occur in 2000 and 15 percent will occur in 2001. The other 5 percent occurred before 1999. While computer system failures will be spread throughout the 30-month period, the failures of devices containing embedded computer chips -- such as factory process control units or building temperature control systems -- will peak at midnight on Dec. 31, Gartner said. The research firm also predicted that 10 percent of all failures will last three days or longer. While Gartner's extended Y2K impact scenario is at odds with most discussions of Y2K impact, the company's leadership role in raising awareness of Y2K practically guarantees that its findings will not be ignored. Gartner has consistently been the leader in estimating the technical cost of Y2K and projects that the software costs alone will be $300 billion to $600 billion worldwide. Other analysts have estimated that related costs, including lawsuits, could push total Y2K expenses to $1 trillion. Vecchio said it's hard to say what specific kinds of computer failures will occur before Dec. 31 and which ones will occur afterward. But the root causes of the failures will be different, he said. During 1999, programs that look forward in time will be the primary sources of failure, Vecchio said. But after January, the problems probably will come from programs that deal with year-end 1999 corporate results and from programs that are used so seldom that their Y2K flaws won't become apparent until well into the year 2000. In addition, efforts to fix Y2K problems inevitably will introduce other software errors that will extend the impact of Y2K for many months, he said. Vecchio said that other post-Jan. 1 problems will be caused by "corrupted data" -- information that accumulates errors during data processing by Y2K-flawed software -- and new versions of packaged commercial software that have acquired Y2K flaws as they were upgraded to include new functions. Worldwide preparedness As part of the San Diego conference, Gartner also offered a worldwide snapshot of Y2K preparedness. It listed the countries that it believes are the most and least prepared for Y2K and predicted how bad things could be for those nations. The countries that Gartner said are most prepared for Y2K are the United States, Australia, Belgium, Bermuda, Canada, Denmark, the Netherlands, Ireland, Israel, Switzerland, Sweden and the United Kingdom. The nations in the second category are the Bahamas, Brazil, Chile, Finland, France, Germany, Greece, Hungary, Iceland, Italy, Japan, South Korea, Mexico, New Zealand, Norway, Peru, Portugal, Singapore, Spain, Taiwan and Thailand, Gartner said. Less well off will be Argentina, Armenia, Austria, Bulgaria, Colombia, the Czech Republic, the Dominican Republic, Egypt, Guatemala, India, Jamaica, Jordan, Kuwait, Malaysia, Panama, Poland, Puerto Rico, Saudi Arabia, South Africa, Sri Lanka, Turkey, and United Arab Emirates, Venezuela and Yugoslavia, Gartner said. Worst off are Afghanistan, Bahrain, Bangladesh, Cambodia, Chad, China, Costa Rica, Ecuador, El Salvador, Ethiopia, Fiji, Haiti, Indonesia, Kenya, Laos, Lithuania, Morocco, Mozambique, Nepal, Nigeria, Pakistan, the Philippines, Romania, Russia, Somalia, Sudan, Uruguay, Vietnam, Zaire and Zimbabwe, Gartner said. The worst thing likely to happen to countries in the best-prepared category is a moderate interruption in government services, Gartner said. The worst scenario for countries in the least-prepared category includes "widespread and severe" interruption of government services, plus widespread and moderate interruptions in air transportation, and moderate to severe interruptions of electrical power, telephone services and the flow of imports and exports, Gartner said. © Copyright 1999 Star Tribune. All rights reserved.
