"Markus Scherer" <markus.icu@gmail.com> writes:
All of France, and Paris, was occupied. It makes sense to apply the occupier's rules that were in effect there during the occupation.
Our sources report that Vichy France used different DST rules than German-occupied France did. The "France" rules are for Vichy France. Europe/Paris uses the occupied-Europe rules; Europe/Monaco (not part of Vichy, but reportedly used Vichy rules) uses the French rules.
In Berlin there were multiple occupiers. The eastern sector [2] was smaller and less populous than the three western ones together [1].
The tz database uses a single latitude/longitude coordinate to identify a city by its center; the current format does not have the capability to represent the entire city outline. So the question, from the tz point of view, is: what is the center of Berlin? If it's in east Berlin the tz database uses the rules appropriate for that location. Currently the database simply accepts Shank & Pottenger's definition for the center of Berlin (as being in east Germany) but if that's incorrect then we can fix it.
Should the TZ rules not reflect the transitions in the larger part of the territory, for historical data where the TZ database won't split the time zone to describe each of the parts?
Not the way it's currently constructed. It currently specifies one location for each zone, and gives accurate data for that location.