Merrill Lynch released a report, "The Millennium Challenge," on July 21. It is a straightforward assessment of the problem. It is not apocalyptic. Releasing this report was an exercise in the public interest. I hope millions of investors will take it seriously.
It begins with this statement: "Call it what you will: a bug, a quirk, a challenge or a crisis. But don't call it trivial. The Year 2000 computer problem is now one of the most important issues facing businesses, governments and other institutions worldwide."
It continues: "However, most institutions remain unprepared. Industry experts estimate that only 33% of U.S. businesses and federal agencies have started work on the problem. International surveys suggest the rest of the world is even further behind. According to a recent U.K. study, for example, fewer than one in five businesses have acted on the problem" (p. 1).
Its description of THE problem -- shared data -- is excellent.
* * * * * * * * *
Pervasiveness: Dates are present in every type of technology -- they are not restricted to mainframe programs. Dates appear in modern client/server applications, operating systems, micro-code found in computer chips and even in the logic etched into a semiconductor chip itself. In other words, the Millennium Bug exists in almost every place digital technology is used -- including so-called embedded technology that is found in manufacturing systems, medical equipment, elevators, telephone switches, satellites and even automobiles.
Interdependence: Computers do not operate in isolation. Information is exchanged between both internal and external systems. A single uncorrected system can easily spread corrupted data throughout an organization and even affect external institutions. Therefore, addressing the date problem requires an organization to extend the scope of the effort beyond its own systems.
Inconsistency: Designing software code is to some extent an "art." This lack of standardization in programs contributes to the complexity of the Millennium Challenge. Dates are labeled, stored and used differently from program to program -- even varying within a single application. Therefore, identifying and correcting date occurrences within an application requires a close inspection of the code in its entirety.
Size: Most large corporations and government agencies have thousands of programs containing millions of lines of code. The U.S. Social Security Administration, for example, has nearly 30 million lines of code. Locating the date fields in complicated systems among countless lines of computer code obviously makes the process enormously time-consuming (pp. 3-4).