Robert Elz <kre@munnari.OZ.AU> writes:
My opinion is that almost all of them should come back. Whatever their history, people had become used to them, and no-one was complaining really (we get far more complaints about the much more irrelevant zone names.)
FWIW (0.2 cents or so?), I agree. Or at least, there should be an *extremely* low standard of proof for external use of a removed abbreviation. Just because people started using what they found in tzdb doesn't make it an invalid citation. (Like it or not, tzdb is the standard so far as most people are concerned; random changes in standards are a bad thing.) In my own downstream project (Postgres), we have so far not removed any of the deleted TZ abbreviations from the list of what we accept for input. Those tzdb changes do affect our output of timestamps, but only in non-default output formats --- the common formats use numeric UTC offsets for zone info. So it's easy to see why we've not seen any pushback from the removal, yet that's not evidence that people aren't using these abbreviations. regards, tom lane