Page #39 (X)
Y2K
Also called the year 2000 problem or the millennium bug (even though it is not actually a bug), a computer problem that was expected to affect older hardware and software on January 1, 2000. Similar problems were expected to arise on other dates, particularly September 9, 1999, and February 28, 2000.
Overview
The Y2K problem represented a family of problems that were anticipated at the turn of the millennium. These problems were expected to arise from three primary sources: internal representation of dates, incorrect programming of leap years, and specialized use of certain key dates.
The issue of internal representation of dates in computer software was expected to be the most critical Y2K- related issue. This is because until recently, most software had been written to store date information in a format that used only the last two digits of a year to represent the year. For example, the year 1960 was internally represented as 60. In the early days of programming, system and disk memory were expensive and had to be carefully optimized, and saving 2 bytes by representing the date 1954 internally as the two-digit number 54 yielded significant savings on a system with limited RAM. This type of "good programming practice" continued even into the mid-1990s. And even though some programmers were aware of potential Y2K problems early on, no one expected programs developed in the 1970s and 1980s to continue being used into the next millennium. This date representation issue was the most commonly identified version of the year 2000 problem because the year 2000 would have been represented internally as 00, which is the same representation for the date 1900. In fact, most older software did interpret the internal date 00 as 1900 instead of 2000, which was expected to lead to problems when calculations were performed using dates. For example, many analysts predicted that financial software might calculate that an employee born in 1954 was actually 00 - 54 = -54 years old in the year 2000, and that such negative ages could lead to such problems as missed paychecks or pension contributions-which on a large scale could have had disastrous social consequences.
This date-representation issue was not just a software application problem but also involved PC hardware, specifically in relation to the real-time clock (RTC) chip, which in most PCs uses only two digits to represent a year, and the basic input/output system (BIOS) routing, which is stored in flash ROM. If the BIOS did not contain code to roll over the century from 19xx to 20xx on January 1, 2000, the operating system would see the date 1900 when the user first turned on his or her PC in the year 2000. In some systems this problem could be fixed with a BIOS upgrade, but other systems had to be replaced with Y2K-compliant hardware.
Another Y2K-related issue was that the year 2000 was a leap year because it was divisible by 4, but some older software applications and hardware BIOS programs were expected to not recognize this fact and would therefore produce dates that were off by one day after February 28, 2000.
Finally, many older software applications used special dates to represent special situations. For example, many older COBOL applications gave the date 9/9/99 a special meaning such as "this item is to be saved forever" or "this item is to be deleted after 30 days." Such applications, therefore, had the potential to produce unpredictable results after September 9, 1999.
Fortunately, the IT (information technology) industry was well aware of these problems years in advance and publicized the potential loss of business that could arise should these issues not be fixed in time. Hardware vendors released new versions of BIOS and other system code to ensure Y2K compliance, and software vendors issued new versions of applications or fixed Y2K issues using service packs and hotfixes. Similarly, IT departments around the world addressed Y2K concerns by allocating resources to upgrading legacy hardware such as PCs built before 1996, upgrading software and applying service packs to eliminate Y2K-related issues, and debugging and rewriting custom software to eliminate other possible date-related bugs. This turned out to be an enormous task-for example, an estimated 180 billion lines of COBOL code had been written for mainframe environments in business, industry, and government, and much, if not most, of this code had the potential to be affected. Y2K problems were not limited to the mainframe arena but also affected server and desktop operating systems and applications on UNIX, Macintosh, and Microsoft Windows platforms, which in turn affected hardware BIOS programming, operating systems, application software, custom code, macros, and data files. The problem was enormous and well publicized, but government and public agencies, enterprise- level businesses, and software and hardware vendors had devoted extensive resources to ensure that systems functioned properly on and after January 1, 2000. The result of these efforts was that instead of the end of the electronic world occurring at the beginning of the new millennium, virtually all government agencies and businesses around the world experienced little more than hiccups in their operation. The Y2K problem, which threatened to be one of humanity's greatest disasters, turned out to be one of IT's greatest triumphs.
