[tz] Question about historical TZData for North Carolina

Paul Eggert eggert at cs.ucla.edu
Mon Dec 7 17:15:14 UTC 2015

On 12/07/2015 04:25 AM, Clifford Ribaudo wrote:
> 	http://ncpedia.org/time-zones
> The relevant section of the article states that:
> 	On 28 Sept. 1947 North Carolina adopted Eastern Standard Time statewide.
> 	During World War I, North Carolina adopted Daylight Time for the periods 31 Mar.-27 Oct. 1918 and 30 Mar.-26 Oct. 1919. Between 1 Aug. 1941 and 9 Feb. 1942, state offices operated on Daylight Time; from 9 Feb. 1942 to 30 Sept. 1945, North Carolina implemented what President Franklin D. Roosevelt called "War Time." From 1946 to 1965 North Carolina did not observe Daylight Time, but since 1966 the state has followed the national schedule.
> So in summary that means we have:
> 	1) Sep 28 1947 statewide adoption of ET
> 	2) Mar 31 - Oct 27 1918 Adopted Daylight Time
> 	3) Mar 30 - Oct 26 1919 Adopted Daylight Time
> 	4) Feb 9 1942 - Sep 30 1945 Adopted War Time
> 	5) 1946 - 1965 Daylight time NOT observed
> 	6) 1966 - Entire state adopts national schedule for Daylight time.
> So I would propose that this data be added to an America/North_Carolina entry in backzone

That's useful info. Let me try to summarize what work would need to be 
done to get this in (and this may help to explain why it hasn't been 
done yet....).

First, that source is better than nothing, but it has real problems. 
First, it cites secondary sources (Doane is an old astrology book if I'm 
not mistaken, and the DOT source is also secondary), and these secondary 
sources are often in error. Second, it summarizes the secondary sources 
without giving full details, e.g., it doesn't say which part of North 
Carolina formerly observed Central Time, and it doesn't give the times 
that switchovers occurred. Someone should read the secondary sources 
directly, or better yet, read the primary sources that they are based 
on, to get as many details as we can.

Second, whatever data we come up with needs to be put in tzdata form, 
with commentary citing the sources. You can look at the existing 
'northamerica' and 'backzone' files for examples, and read the tzfile 
man page (and the file 'Theory', which explains naming and in particular 
why we shouldn't use the name 'America/North_Carolina') for details.  
The result should be a patch to the file 'backone' in Git form. If 
that's all done and the patch goes in, you should be able to run 'make 
posix_packrat', and use the result in your programs.

More information about the tz mailing list