On 2013-09-09 23:03, Paul Eggert wrote:
Clive D.W. Feather wrote:
don't put LMT values in the files; remove the ones that are there
Unfortunately the underlying interface requires that *some* offset be there, and LMT is more accurate than UTC would be, so there is an argument for keeping it in the database so long as it's understood to be a somewhat-flaky approximation.
That being said....
In hindsight perhaps we should have not put LMT in there, and instead designed a data format that allows one to say "undefined" (so presumably localtime fails).
The same goes for the transition from LMT to standard time, which is often not reliably known even to the nearest decade (even though the data format requires that it be specified to one-second precision!). But how would that be modeled so that the caller of localtime could find this out? Maybe localtime should return randomish values each time you call it, if the result is not reliably known, sort of like randomized rounding? (I'm mostly kidding here....)
Would time standards have been set from local or national observatories, universities with astronomy departments, or admiralty equivalents, prior to and after standard time? Those institutions are most likely to have been involved in production and use of ephemerides and keeping relevant records.