2000 Will Be Too Late
You come to work on Monday morning and find the building's elevators out of order. After you trudge up the stairs, you realize the computerized phone system is shut off, and your e-mail is nonfunctional. Still you forge bravely ahead, only to discover that your invoicing and purchasing software, and other applications, produce erroneous information.
That could be the scenario for Monday, January 3, 2000, if your company ignores the Year 2000 issue.
by Noriko Takezaki
You've heard the scare scenario: that many of today's computer systems could come to a screeching halt on January 1, 2000, because they store and process calendar years as two-digit (97) rather than four-digit (1997) numbers. If you think your company still has a couple of years before it has to worry about dealing with the problem - think again! Solving the Year 2000 Problem is a thorny issue that will take time, and time is quickly running out.
The Japan Information Service Industry Association (JISA) has been actively warning Japan's corporate computer users about the impending Year 2000 Problem since 1995. But while JISA - a nonprofit organization affiliated with the Ministry of International Trade and Industry (MITI) - has been urging timely action in implementing the required software and hardware changes, so far few corporations have been listening.
| Japan Information Service Industry Association (JISA) : TIME 24 Bldg. 17F, 2-45 Aomi Koto-ku,
Tokyo 135-73 Phone 03-5500-2610 Fax 03-5500-2630 |
JISA was founded in 1984, and has about 600 corporate members. Although JISA's research reports are intended for its members, they are available to non-members for a fee.
JISA's website, http://www.jisa.or.jp, contains an English-language summary of the "Report on User Questionnaire Survey for the Year 2000 Issues." Other English-language materials online at the JISA site include "A Statistical Survey on Software Export/Import 1995," "MITI Annual Survey
of the Information Service Industry 1995," and "The Information Services Industry Japan 1996."
In addition to the Year 2000 Problem, JISA is active in promoting quality control systems and standardization, pursuing research on cutting-edge information technology, and promoting cooperation between industry and academia. JISA also participates in such international organizations as the World Information Technology and Services Alliance (WITSA) and Asian-Oceanian Computing Industry Organization (ASOCIO). The 15th general assembly and symposium of ASOCIO will be held in Tokyo in November 1997. |
A shortage of talent
"There will be shortage of engineers who can handle the Year 2000 issue. This will be a big problem for top management, but they aren't well aware of it," says Akira Ogata, general manager of JISA's research and planning department. "Technically, the Year 2000 issue is not a big problem. But if we don't set about dealing with it in time, it will cause not just problems with computer systems, but Japan will suffer economic losses resulting from the system failures."
JISA carried out its first survey on the Year 2000 Problem among computer users in Japan's public and private sectors in summer 1996. Based on replies from about 750 users, JISA found that although there is significant concern about the issue, most corporations and public organizations have been slow in preparing to deal with it.
According to JISA estimates, there will be an incredible shortage of software engineering expertise (up to 620,000 man-months) in Japan during the 1997 to 2000 period. Many corporations, however, don't understand the magnitude of the problem, or appreciate the resources that are needed to deal with it. [See the "So what's the problem" sidebar-Ed.]
One bottleneck is the number of software engineers proficient in such computer
languages as COBOL and PL/1. Many of the mission-critical business applications in
Japan are written in such languages running on ofucon (office computers) and general-purpose computers. With the diffusion of network computing, and the growing popularity of such languages as C++ and Visual Basic, fewer engineers are being trained in these past-generation languages.
JISA estimates that only 30% to 40% of the software engineers in Japan are technically able to handle the Year 2000 Problem. The association further points out that, in practice, the number of available engineers will be lower than these percentages indicate, since not all such engineers can be spared to tackle the issue; many are busy with other important jobs.
The aforementioned shortage of programming talent is based on the assumption that most corporations will start working on the issue this year. If preparations are delayed, the shortfall of needed software engineers will be even greater in 1998 and 1999, and the situation will be much worse, according to JISA's Ogata.
The alternatives
JISA considers that preparation work done by in-house engineers is the easiest and least expensive method of retrofitting computer systems to handle 21st century dates. For companies that cannot manage this, however, there are two alternative solutions: replacing their entire legacy systems with new hardware and applications that are Year 2000 compliant, or outsourcing the entire project to a third-party solution provider.
The first alternative is expensive, but JISA concludes that some companies may have to fall back on this method. If a company does not have the right staff for adequate system administration, or lacks sufficient documentation on its existing systems that would allow engineers to alter the date-setting format, there may be no choice but to replace the entire system.
Regarding the second alternative, outsourcing, JISA suggests that the shortage of engineers in Japan will make it necessary to seek help from abroad. JISA learned that many American companies have started asking for help from software engineers in India to solve their Year 2000 Problem. As part of its study on the potential use of foreign engineers, therefore, JISA visited India in February.
JISA found no problem with the technological expertise of Indian engineers, but does have a concern about the language barrier. Unlike the case of the US, here foreign engineers will have to deal with Japanese-language computing environments. Therefore, JISA warns that Japanese companies can only count on the assistance of foreign engineers who understand the Japanese language.
Techno-apocalypse
With every passing month, the magnitude of the Year 2000 Problem will become more and more evident. Companies need to tackle the issue now, to be assured of having the time and resources they need to finish the job.
The millennium may not bring the apocalypse, but for businesses that are not prepared, it could bring the
next-worst thing.
So what's the problem?
The "Year 2000 Problem" is one that affects all time-based calculation: invoicing systems,
interest calculations, credit card expiration dates, computerized equipment maintenance schedules.... Why does it exist? Blame the high cost of data storage and processing capacity prior to the 1980s, and the shortsightedness of yesterday's managers and programmers. How much will it cost businesses worldwide to solve? An estimated $600 billion or more.
Since the first generation of business computers, introduced in the late 1950s,
calendar years have been handled as two-digit (decimal) numbers, 00 to 99. When calculating based on such a system, centuries don't exist; the computer reckons the time between 1997 (97) and 2000 (00) not as 3 years, but as -97 years.
The obvious question is, "Why did early programmers decide to use only two digits?" Well, the computers of those early days were expensive, with memory and processing capacity being limiting factors. Even an incremental boost in either meant a substantial increase in system cost, so handling years as a two-digit rather than a four-digit number seemed like a valid compromise. Besides, the new century was still decades away - plenty of time to deal with any problem.
But even if a vendor had warned a client in 1970 that a more powerful system would enable years to be stored as four digits instead of two digits, and thus forestall future problems, the likely response would have been, "Why spend money now for something that won't even be an issue for another 30 years?" In most businesses, the focus is nearly always on immediate cost savings, profits, and dividends. Why would managers (or programmers) lose sleep over a problem that their successors would have to deal with?
"Let's just put those two digits back in," is the solution, but that isn't as easy as it sounds. It isn't unusual for legacy systems in a moderately large company to have 10 million (or even 100 million) lines of code. It takes time to even look at that much code and pinpoint all the references to dates (well over one programmer-year for 10 million lines of code). Then comes trying to reprogram around the problem. Estimates for the cost of correcting the Year 2000 Problem range between JPY 90 to JPY 250 per line of code.
And the deadline for reprogramming isn't really 2000. Unless a company is foolhardy enough to try and run new, un-debugged business-critical software, several months of testing will be required for a full suite of applications.
You say you're hoping that some ingenious third party will devise a simple solution you can implement before the big day arrives? You're obviously the type who counts on the US Cavalry to come riding to the rescue of the settlers so the movie will have a happy ending. It might happen, but are you willing to stake your business on it? And when will you stop hoping for the cavalry and start tackling the issue? When your system fails on January 1, 2000, it will be too late.
How much will it cost your company to correct the Year 2000 Problem? A better question might be: How much will it cost your company if you don't?- Wm.Auckerman |
Back to table of contents
 |