[tz] Reason for removal of several TZ abbreviations

Guy Harris guy at alum.mit.edu
Tue Dec 5 03:21:56 UTC 2017


On Dec 4, 2017, at 6:26 PM, Michael Douglass <mikeadouglass at gmail.com> wrote:

> On 12/4/17 19:38, David Patte ₯ wrote:
>> Using such a scheme, a database such as geonames would then map a location to the tz id (as it does now), defering political arguments to geonames, and removing them from this list. It seems appropriate.
>> 
>> But what about the tz designations (such as EDT), do we have a solution to that conundrum?
> I think under such as scheme the differentiation between names and aliases disappears. They are all names mapped onto a unique tzid, so America/NewYork and US/Eastern would refer to the same tzid.

As I pointed out in an earlier email in this thread, the "tz designations (such as EDT)" are separate items from tzids (such as America/New_York and US/Eastern).

tzids are names for tzdb regions, with every tzdb region having a tzid (and some aliases) that refer to it, and every tzid referring to one and only one tzdb region.

Tz designations are just things that are shown in times displayed to humans, and do *not* refer to a particular tzdb region and cannot always be used to uniquely identify a tzdb region (for example, does "CST" refer to America/Chicago, America/Indiana/Indianapolis, America/Winnipeg, etc.?)

So nothing done about tzids will have any effect whatsoever on tz designations or on the conundrum in question, namely "what to do about tz designations that the tzdb maintainers invented in order that, at certain times in certain tzdb regions, UN*X/POSIX API implementations using the tzdb could provide a tz designation to use at those particular times".


More information about the tz mailing list