[tz] Correcting daylight-savings time for central Missouri prior to 1967
zoidsoft at gmail.com
Sat Nov 24 03:38:35 UTC 2018
> I only became aware of how uncertain time change history is when I began running across astrology software that uses the tz database. They consistently miscalculate my Columbia, MO August 1950 birth chart due to assuming daylight savings time was in effect. No other astrology software I’ve used has this problem.
> I would like to be part of the solution, but don’t know how to beyond sharing the Doane information.
We’ve known about the Doris Doane collection for a long time. The issue is that the farther back in time change history you go, the more iffy the data. The fog of history encroaches. The TZ database doesn’t try to overstate accuracy and IMO has a reasonable cutoff date of Jan 1, 1970. We can try and be more accurate for time stamps in the past and perhaps the ACS Atlas is better for use in astrological software, but the TZ database wasn’t built for that purpose. There are attempts to expand the format of the TZ data to include time change history pre-unix epoch date of 1970, but the growth of the database will grow exponentially the farther back in time that you go. My guess is that bringing full history all the way back to the onset of standard time circa 1900 (earlier in some areas), would cause the tz database to expand to 10,000+ zones (it’s roughly 470 now). There are approximately 500 researchers subscribed to this group (last I heard from ADO). If we can’t accurately account for that level of depth, then how is a small company of a handful of researchers going to do this?
I think it would be better if we set aside confirmation bias and really see the issues, which can’t really be fully understood unless you’ve read through the entire database portion of the TZ data (Pauls notes and notes from other researchers). Take all such information with a large grain of salt.
More information about the tz