On 2005-08-15 13:07, "Mark A. Biggar" <[EMAIL PROTECTED]> wrote: > 3) use Astronomical Dates which are kept as the number of days sense > noon Jan-1-4713 BC.
More specifically, that's the astronomical Julian Day, or JD, and JD 0 began at noon Universal Time (a.k.a. GMT) on January 1, 4713 BC in the Julian calendar. Sometimes this is called the Julian Astronomical Day, or JAD, to distinguish it from various other misappropriations of the term "julian day/date". More generally, astronomers use the Gregorian calendar like the rest of us, but often extend it into the past before its adoption, ignoring the Julian calendar switchover. They also tend to use 0 and negative years AD/CE in lieu of BC(E) numbering. On that basis, JD 0 began at noon on Nov 24, -4713. Julian Days make a good neutral format for storing dates (and, with fractional components, times), but they are a bit cumbersome. If you don't actually need the range they provide (because you're only dealing with dates near the present), it might be regarded as overkill to store the current time as e.g. 2453598.225995. In order to get 1-second resolution, the JD requires the same number of significant figures that's required to get millisecond resolution with time_t.
