I thought this was an interesting discussion of leap seconds. At its heart, the debate is about a tradeoff between hassle (a lot of programming goes into inserting a second into a day every year or so) and how close we want time to match its traditional association with astronomical observations (e.g. noon is exactly noon at Greenwich). This is a debate that has occured at least since the imposition of time zones (mainly at the behest of railroads) which for many cities converted "sun time" to "railroad time." Until then, every town was on a different time with noon set to local astronomic noon. Now, only a few cities actually have noon at noon. Of course, daylight savings time took this even further.