Tuesday, July 11, 2006

James Kunstler and Y2K

James Kunstler, otherwise wonderful, feels that the Y2K problem was averted by busy programmers[July 10th]. He must think that there were no Y2K problems in China, Borneo, Mexico, and France because all the clever programmers in those countries found and fixed all those bugs. Clearly this is a false belief, but, well, no one is perfect.

Another false belief: The Y2K bug had a "deadline" of the night of Dec 31st 1999, presumably as the second hand, the minute hand, and the hour hand all crept or swept by the "12". Not so, as all you cognoscenti are aware. The "mess" caused by the ambiguity in deciding whether one date is before or after another date when the century is not specified was already half over at that point. The first half of the bugs, the forward looking comparisions, had all just begun to start working properly again(!) at that point, and the second set, the backward looking comparisons, had just started to malfunction.

A factor that helped enormously: modern databases have a special "date" data type, that does store the century. Applications using databases (Oracle, DB2, Informix, Access, ...) had limited exposure to the bug.

No comments: