An untested system must be assumed to be noncompliant. This is why most estimates of the cost of a y2k repair put 40% or more into the testing category. Testing is crucial. Problem: What if the test fails? What if the repair shuts down your system? You must then start over. You must also find out what iut was that crashed your system. Then you must fix this and test again.
A few members of the U.S. securities industry see the problem. They also see that Dec. 31, 1999, is too late to begin testing. But when should the deadline be? The industry wants it to be in March, 1999. This is a degree of optimism matched only by the U.S. government, whse deadline in November, 1999.
(This rosy scenario assumes that the banks will still be open in 1999. If they aren't, programmers will not get paid. No testing then!)
Notice especially this comment regarding the exchange of data: "But the panelists said not only must companies work to get their own systems ready for 2000 but they must make sure anybody with whom they exchange data is also prepared." This is the problem of systems. It's not enough to get your computer compliant. You must make certain that every computer that your computer interacts with is also compliant, and compliant with your computer's repair. Whenever you read this, think "banking system."
* * * * * * * * * * * * * * * * * * *
Wednesday April 30 9:13 PM EDT
Year 1999 seen as time to test for year 2000 bugs
NEW YORK, April 30 (Reuter) - The U.S. securities industry is aiming to have computer overhauls ready for testing in the year 1999 as Wall Street gets ready to meet the challenge of the year 2000 date, Michael Tiernan, vice president of information technology at Credit Suisse First Boston CSHZn.S told a panel Wednesday.
Tiernan, who heads the Securities Industry Association's (SIA) Data Management Division Year 2000 Committee said if three months of tests begin in March, 1999, that would leave six months to deal with whatever contingencies come up.
Tiernan and other speakers on the panel, part of a week-long Reuters World '97 event sponsored by Reuters Holdings Plc, warned that the "millennium bug" that business and govrnment are working feverishly to exterminate has the potential to be a very serious problem.
Richard Hunter, research director for applications development technologies at Gartner Group Inc , said the cost of dealing with the bug is likely to be in the $300-$600 billion range worldwide.
He said a recent report by Gartner, a research group, showed that only one of every eight companies surveyed has a plan in place for dealing with the issue.
But the panelists said not only must companies work to get their own systems ready for 2000 but they must make sure anybody with whom they exchange data is also prepared.
"The industry is completely connected," said Maggie Parent, a principal with Morgan Stanley Group Inc . She noted that Wall Street firms have computers which are inter-connected with those of counter-parties in trading, custodian banks, and others.
Henry Guttman, an attorney with Simpson Thacher & Bartlett, said while lawsuits would be inevitable if there are failures and people suffer economic injury, his "business advice" was that companies and their suppliers work in a spirit of "cooperative efforts" and "shared understanding."
David Ure, executive director for global technical development and strategy at Reuters, said his company is thinking about an inititaive for clients that might be called "millennium checks." He said while clients would pay for it, Reuters was not viewing it as a "revenue opportunity."
He also said Reuters has not given an estimate of what it will spend to deal with the millennium bug.
Hunter said that lately a "six-digit" solution has been becoming "increasingly common" in part because it is viewed as less expensive.
He said by remaining with a six-digit date the actual information stored on databases will not be modified, but instead logic will be inserted into the programs using the data to interpret it and treat as if it were an eight-digit date. But he said $300 billion is the still the "rock bottom" estimate of what it will cost to get computers ready for the millennium.