Fourth in an occasional series on Japan’s Y2K preparedness
Akira Ogata does not place full confidence in government survey figures on the nation’s preparedness for the so-called millennium bug.
The general manager of research and planning at the Japan Information Service Industry Association, a nationwide group of major software firms, says most such surveys are conducted without a clear definition of trial tests, which he argues to be the most important element in gauging preparedness for the computer bug.
“What counts most are trial tests to simulate environments (of 2000 and beyond) after conversion work is finished. But the tests are, in fact, very difficult to conduct,” Ogata said in a recent interview.
The millennium bug, or Year 2000 problem, refers to expected errors caused by older computer and software systems that use only two digits to record year entries. Data for 2000 and beyond are subject to error if “00” is mistaken for 1900, rather than 2000.
A number of government surveys have indicated that many industries in Japan are making steady progress in necessary conversion work and trial tests for dealing with the bug.
But company computer systems often consist of hundreds of terminals, complicated networks and other peripheral devices, Ogata pointed out.
Firms rely on such systems for daily operations, and it is often difficult to stop the entire system for trial tests to input mock data, he said.
Engineers therefore are often forced to “simulate” the environment of 2000 and beyond by examining only parts of the total system, he said. “On top of that, there is the question of how far the test should cover outside computer networks connected to the system (of a given firm),” Ogata said.
Many industry experts like Ogata place particular stress on the difficulty of conducting trial tests on the “supply chain management” system, which controls the flow of logistic data and goods between a supplier and distributors via computer networks in real time.
But many of the government surveys do not provide any definition of trial tests, simply citing the number of firms that claim to have “finished” conversion work and trial tests, Ogata said.
“They (the government) merely collected one-sided (voluntary) reports,” he said, adding that JISA is “concerned” about the vagueness of terminology used in a variety of statistics on Y2K problems.
JISA also has other concerns regarding Japan’s Y2K preparedness — a possible shortage of software engineers, Ogata said.
In February 1998, the association announced its estimation of the number of software engineers needed to correct Y2K problems, with the nation being 150,000 short in the worst-case scenario.
JISA does not have updated information on the current situation, but Ogata said he is rather pessimistic, saying still a large number of small firms have yet to place orders with software houses to secure conversion work and staff.
In addition, older computer systems with Y2K bugs often use older computer languages such as COBOL, with which many of today’s young programmers are unfamiliar, he said. “Roughly speaking, only between 20 and 30 percent of all software engineers and operators can deal with those older languages necessary to correct Y2K problems,” Ogata said.
“I am never optimistic. I think the chances are high that (conversion work) will not be made in time, particularly at small firms,” he said.