On 18.08.11 04:46, Paul Eggert wrote:
Yes, the situation in China was quite different back then, but luckily for us our cutoff date is the POSIX epoch of 1970-01-01 00:00:00 UTC. In the tz database we don't create a new zone merely because some locality had different clocks before 1970. If we did, we'd need hundreds of zones for Indiana alone.
We (a group of timezone researchers for astrology purposes) are planning to do exactly that, create the required 300+ Indiana zones (Shanks/Pottenger uses 350+ but is not complete), to cover pre-1970 Indiana timezone history. It will take a lot of work and we understand that these extra zone will not get entry into standard tzdata file distribution. But I think if we maintain a separate 'history' zone fileset outside, tzcode will work with these extra zones of we include them in the zic compilation step. For China as well. Writing the zone files is only one part of the work, the other is describing precisely for which area the extra zones apply, and which area is covered by the existing standard tzdata zones.