"Mark Davis" <mark.davis@jtcsv.com> writes:
Arctic/Longyearbyen // or Atlantic/Jan_Mayen, don't care which
I'd suggest Longyearbyen over Jan Mayen, as its population is much larger.
3. And if the database had unique TZIDs corresponding to the 'missing' ISO country codes BV, HM
I think someone else has already addressed this issue, but those "countries" are uninhabited. The TZ database attempts to record the clock values that people actually use, so if there are no people then the local time is undefined. More generally: as the number of people (at a location) shrinks, the question "what is the local time?" becomes more and more arbitrary. In some Antarctic locations it seems that the answer really and truly depends on who you're talking to. When the number of people equals zero, the value is undefined, so it doesn't make sense to put it into the database. To give another instance of this problem: the uninhabited island of Clipperton is officially in the PF country code, but it doesn't correspond to any of the TZIDs for PF (Pacific/Tahiti, Pacific/Marquesas, Pacific/Gambier). If Clipperton ever becomes inhabited, it'd undoubtedly have a UTC offset that differed from those three entries (and quite possibly it would no longer belong to PF).
From the TZ database point of view, Clipperton is just like Bouvet: it's a small patch of land without any inhabitants, so it doesn't get an entry, even though it would deserve one if it were inhabited.