On Tue, Aug 13, 2013, at 17:32, Paul Eggert wrote:
> enh wrote:
> > Android currently uses zone.tab to get a list of time
> > zones in use in a given country.
>
> Yes, as does tzselect (which is in the tz code).
> This continues to work with the proposed changes:
> zone.tab still lists all time zones in use in a
> given country, and every entry in that list continues
> to work.

The problem is, most tools that use it list city names, and that is not
likely to change.

> Marc Lehmann wrote:
> > reducing the quality of the tz distribution(s)
> > (which were beyond outstanding) for political
> > reasons
>
> No real reduction in quality is being proposed.
> There's no significant difference in behavior between
> the current and proposed data.

----

What happened here is _you_ made a blunder in 1995, by rigidly applying
the "most populated city" rule [ignoring the principle of leaving small
violations like a hypothetical Rome vs Milan alone] to a city you
believed was the most populated one in the region while ignorant of the
dispute as to whether, in fact, it is in the region.

This went unnoticed for a while, and when it was pointed out, you
covered for it by trying to turn "politics" into a massive bogeyman
rather than going for the simple non-disruptive solution of causing the
database to once again not comment on the status of Jerusalem.

The idea that no reduction in quality is being proposed is ridiculous.
The reality of the fact that country codes and city names are listed
alongside each other by the vast majority of tools that use zone.tab
should not be ignored. And politics is not a significant enough issue
often enough to justify completely destroying the "Each region is only
in one country" rule rather than simply avoiding naming them after
cities in disputed border regions.