Re: Date Trivia

In article <1993Apr9.030226.7113@informix.com> aland@informix.com (Colonel Panic) writes: > >Are any UNIX versions set up to automatically handle leap seconds? After >all, we get our time from the O/S. (e.g. we don't provide device drivers to >use an atomic clock as an OnLine chunk). > I don't believe there can be an "algorithm" that tells us when we are going to need to insert a leap second, because it is basically an empirical thing. There are minute variations in the length of the year, probably due to massive fluxes of material in the earth's core which result in changes in the planet's moment of inertia. Since angular momentum is conserved, the spin rate has to vary to compensate. But the need for a leap second is essentially unpredictable, and is only revealed by measurements as time goes by. In fact, it is the same with leap years, given a long enough perspective. We know the "formula" (every 4, except every hundred, except every four hundred [and possibly a few more terms after that]) to a certain degree of precision, but we cannot know the "true and complete" formula, because to do so would be to know the ratio of the length of the day to the length of the year to infinite precision, and we can't know empirical constants (even if this one were constant) to infinite precision. Well, I'll shut up now.