Eric Muller wrote:
There is the situation of changes in administrative boundaries, ... It's unclear to me whether the areas that changed should have resulted in creating new TZ timezones, but they certainly have not.
Since the database associates each zone with a city or similar concentration of population, perhaps these boundary changes should cause creation of a distinct zone only where a city changes hands. Taking the Saudi-Yemen example, the territory changing jurisdiction is a large swath of desert oilfield and a chunk of sea, both only lightly inhabited, not making a pressing case for timezone tracking. Actually the Saudi-Yemen change also has no effect on the TZ database for a stronger reason: both countries set their clocks the same (constant UT+3h since 1950), so the boundary change doesn't actually affect the agreement of wall clocks. When was the last time sovereignty changed over an appreciable conurbation but not over its entire country? There were some in the resolution of World War II; any later? Any since 1970? -zefram