> That has never been true of the tz database.

While that has not been true of the tzdb, it has IMO always been a flaw.

Until recently, every normal* Unicode region code except the following have had at least one unique TZID.
  • AC Ascension Island
  • BV Bouvet Island
  • CP Clipperton Island
  • DG Diego Garcia
  • EA Ceuta and Melilla
  • HM Heard & McDonald Islands
  • IC Canary Islands
  • TA Tristan da Cunha
  • XK Kosovo
It is extremely useful to be able to map all of them to at least one unique TZID, if nothing else than for testing. It does not hurt to have extra links in order to make that true.

But it is more than that. There are definitely people upset to see their timezone reported as being that of a city in a different country, simply because they happened to have the same rules. Most of the above have small populations, where one could argue that it doesn't matter much (and Kosovo is a special case).

* excluding: Unknown Region (ZZ), deprecated codes (like SU), macro regions (like North America)


Mark

— Il meglio è l’inimico del bene —

On Fri, Oct 3, 2014 at 2:27 AM, Paul Eggert <eggert@cs.ucla.edu> wrote:
Thanks for the info, but I'm afraid there's not much new here.

Stephen Colebourne wrote:
The recent change to map Harare to Maputo broke a test in Joda-Time

 Did Joda-Time's regression tests also fail for 2014h's changes to Jamaica time in early 1974?  Those changes are for more-recent time stamps, time stamps that are in scope for the tz database.

there should be at
least one actual Zone entry for each ISO defined territory/country,

 That has never been true of the tz database.  Adding such a requirement now would result in more political churn and maintenance overhead, something we already have too much of.  There is a real benefit to keeping our distance from politics, even if this means some regression tests need to be updated.