On Apr 18, 2013, at 11:01 AM, random832@fastmail.us wrote:
On Thu, Apr 18, 2013, at 13:58, Guy Harris wrote:
Presumably by "understand" you mean "have a file with the appropriate pathname"; I don't think software should "understand" the identifiers except by using them to construct pathnames of Olson-database files to load - it shouldn't "know" the pathname syntax or "know" what particular values of particular components of the pathname "mean".
The timezone identifier is used as a key in other databases, such as CLDR (whose maintainers have, IIRC, declined to recognize any changes made to the identifiers used for existing timezones)
Perhaps both databases should be using the same LOCODE-derived identifiers as the "official" identifiers, with all the region/city names used as legacy backwards-compatibility names? Using those as the "official" identifiers has the advantages that: 1) they look like line noise to humans, so UIs for setting the zone will perhaps make an effort to do something better than offer you a choice of zone identifiers or zone identifiers with underscores replaced by spaces; 2) they look like line noise to humans, so perhaps people won't get quite as bent out of shape because The Wrong City was used; 3) they look like line noise to humans, so perhaps people won't get quite as bent out of shape because The Wrong Region was used; 4) advantages 2 and 3 mean they won't have to change over time; 5) advantage 4 might mean that the CLDR folks won't have to worry about identifier changes.