The Year 2000 problem is a direct result of the industry wide data processing dependency on date information and the common programming practice of representing the year in a date field with only two digits. In order to save cost and space, programmers used "shorthand" for handling dates (e.g. 1997 was represented as 97) These two-digit dates exist on millions of files used as input to millions of applications. Perhaps a more intractible problem are microchips, produced as recently as 1996, which also had a time function designed into them that was only structured for this century. When the date goes from '99 to '00 both they and the legacy software that has not been fixed may think it is still the 20th century – not 2000, but 1900.
Every major aspect of modern infrastructure has systems and equipment that rely on such calculations to perform their functions. The world's economic and political infrastructures rely on dense networks of computer systems. Whatever happens in one part of the network has an impact on other parts of the network.
There are somewhat over a billion embedded microprocessors (hardware chips or microchips) located in computer systems worldwide. They sustain the world's manufacturing and engineering base. They exist in traffic lights, elevators, water, gas, and electricity control systems. They are in medical equipment and military and navigation systems. Traffic control systems are dependent upon them. They are in the satellites that circle the earth. Global telecommunications are heavily dependent on them. Modern cars contain about two dozen microprocessors. Many of these chips aren't date sensitive, but a great number are, and engineers looking at long ago installed systems don't know for sure which is which. To complicate things further, not all chips behave the same. Recent tests have shown that two chips of the same model installed in two different computers but performing the same function are not equally sensitive to the year-end problem. One shuts down and the other doesn't.
It is impossible to locate all of these chips in the remaining months, nor can all those that are identified be replaced. Chips more than three years old are obsolete and are probably not available in the marketplace. The solution in those cases is to redesign and remanufacture that part of the system – which often makes starting over with new equipment the best option. That is why some companies are spending millions, even hundreds of millions, to replace their entire computer systems. It at least ensures that their internal systems work.
The larger significance of this Year 2000 timebomb (often called Y2K) extends far beyond single businesses, or even entire industries. Indirect relationships extend like tentacles into many other networks, creating the potential for massive disruptions of service. A few but strategically placed failures would initiate a major economic cascade. In an interdependent system, solving most of the problem is no solution. The integration of these systems requires that we solve virtually all of them.
The computer system at a large London department store recently destroyed tons of food during the process of doing a long term forecast. The computer read 2002 as 1902. Instead of four more years of shelf life, the computer calculated that this food was ninety-six years old. It ordered it thrown out. A similar incident happened in the USA at the warehouse of a freeze dried food manufacturer.
It is not unusual for a company to have more than 100,000,000 lines of computer code. One computer-industry research group, estimates that globally, 180 billion lines of software code will have to be screened. The US Social Security Administration began working on its thirty million lines of code in 1991. After five years of work, in June, 1996, four hundred programmers had fixed only six million lines. The US Internal Revenue Service has 88,000 programs on 80 mainframe computers to debug. By the end of 1997 they had cleaned up 2,000 programs.
A firm that tracks programmer productivity, estimates that finding, fixing and testing all Y2K-affected software would require over 700,000 person-years. It projects that the total cost of dealing with Y2K worldwide will be somewhere between $300 billion to $600 billion – and these are only direct costs associated with trying to remedy the problem. The US Office of Management and Budget (OMB) estimated total government Y2K expense at $3 9 billion. The OMB warned that this estimate might be as much as 90% too low considering the increasing labour shortage and expected growing remediation costs as January 1, 2000 looms nearer. In June 1998, it was announced that US federal agencies had already spent five billion dollars. Of twenty-four agencies, fifteen reported being behind schedule.
The electric power industry is in danger of massive failures because the programmable controllers – which have replaced mechanical relays in virtually all electricity-generating plants and control rooms – may behave badly or even freeze up when 2000 arrives. Many air traffic control centres around the world rely on mid-'80s vintage mainframe computers; flight safety relies additionally on satellites with embedded chips, individual systems in each airplane, and air traffic control systems around the globe. Lufthansa already has announced it will not fly its aircraft during the first days of 2000.
In more and more businesses, expenditures for research and development and modernization are being diverted to Y2K budgets. Business Week in March of 1998 estimated that the Year 2000 economic damage alone would be $119 billion. When potential lawsuits and secondary effects are added to this – people suing over everything from stalled elevators to malfunctioning nuclear power plants – the cost easily could be over $1 trillion. A recent survey of American corporate chief information officers found that 70% of them thought that their companies would not be completely prepared for Y2K. Additionally, 50% of them acknowledged that they would not fly during January 2000. One expert predicts that upward of 70% of the businesses in Asia will fail outright or experience severe hardship because of Y2K. Another has raised the probability of a deep global recession in 2000-2001 as the result of Y2K at about 70%.