> The time available in computers (smartphones etc) is normally UTC and a > local time offset, which is sufficient for almost all times, but not > quite all. On average every 18 months leap seconds are inserted. Events > during these leap seconds cannot be unambiguously timestamped by a > representation of time based on UTC. According to the IAU the Julian > day, JD, should be calculated from a date/time presentations using TT > (terrestrial time) rather than UTC. TT is presently 67.184 s ahead of > UTC, this changes with time, especially with every inserted leap second.
The "time" is GMT (UT1). UTC has leap seconds to correct UTC to GMT (UT1). Computers (smartphones etc) all use minutes which have exactly sixty seconds and are syncronized to the Mean Observed Solar Zenith at Grenwich. Just because a cutesy name (UTC) replaces the correct name (GMT), does not mean that the time (or how it is calculated and used) is at all different because of it. > 1) do nothing, or maybe just document more clearly, that the julianday > in Sqlite, for date/time input in UTC, returns JD(UTC) in IAU > terminology, which, for example, does give time differences (or nr of > days elapsed since epoch) ignoring leap seconds and is off by about > 1 minute with respect to astronomical JD. SQLite calculates the Julian Date from the time returned by the Operating System. Except for very special platforms, this is UT1 (GMT). Almost no computer system actually uses UTC -- in fact, UTC is a "display format" only. Some people "call" it UTC, but it is not. It is Mean Observed Solar Time at Grenwich. Each minute contains 60 seconds, each hour contains 60 minutes, and so forth. Leap Seconds only exist to correct the artificial UTC periodically to UT1 (or Mean Observed Solar Time at Grenwich). --- () ascii ribbon campaign against html e-mail /\ www.asciiribbon.org _______________________________________________ sqlite-users mailing list sqlite-users@sqlite.org http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users
