On 10 July 2014 21:20, Marc Lehmann <schmorp@schmorp.de> wrote:
field. In practice it seems that end users don't much care about things like the time zone of Guadeloupe in 1899 -- which is probably a good thing, since the pre-2013e database was wrong anyway.
I think almost everybody agrees to that.
One major difference I noticed when comparing the recent Pacific <https://github.com/eggert/tz/commit/662016b64ae64bc8d2680f349d1a7813dc1ee536> and Africa <https://github.com/eggert/tz/commit/4b4e789d5c5ee79366b4606d139cbb9eb1d5a28d> changes to the September 2013 changes in the Caribbean <https://github.com/eggert/tz/commit/c7bc2f8e1a65146a920977f297524e25df642674> is that, in the previous changes, the affected and coalesced zones were geographically close, and in all but one case, the most recent information lost was from 1912 and was only LMT. Needless to say, it's hard to sustain serious objection to such relatively minor changes, since as we know, most end-users don't really care about 100-year-old timestamps. In contrast, the removals from the Pacific and Africa are sweeping and diverse. The coalesced zones are geographically disparate and many have differences as recent as the late 1960s that have completely been removed. Additionally, I find it ironic that after a series of patches clarifying abbreviations, including the proper use of JCST versus JST, it has been decided that many are to be thrown out, and that we should link Palau into JST. I have little doubt that many of these, even those we invented, are in use... and at the very least, greater care should be taken to ensure that we aren't giving an "unfriendly" abbreviation to a region. Whatever happened to having one zone per ISO-coded region? Yes, technically, we have all those zones, often as links, but I think there's something to be said for decoupling these and letting each region remain independent, as they truly are. (For those genuinely concerned about the extra space requirements, I'd recommend that links conditional upon a winnowing threshold could be a feature that we should consider building into zic to achieve that end.) Although we have never claimed to be an authoritative compendium of all timekeeping facts, tz and its associated commentary have been highly regarded in the past for this purpose. It is a shame to see some of this historical documentation discarded for the sake of expediency that is marginal at best. Let's not forget that our code is art, and that art is often messy. -- Tim Parenti