[tz] Proposal for new rules

[Stephen Colebourne]

On 30 August 2013 09:59, Guy Harris <guy at alum.mit.edu> wrote:
> On Aug 30, 2013, at 1:15 AM, Stephen Colebourne <scolebourne at joda.org> wrote:
>> Yep. The difference in LMT is something that isn't important for Rome
>> vs Vatican as they are the same city. But Rome vs San Marino? Also
>> Atikokan vs Panama.
> Rome vs. Florence?  Rome vs. Naples?  Rome vs. Milan?  Rome vs. Turin? etc.

Not time-zone IDs, but I acknowledge with the sentiment.

>> I actually think there is a good case for removing LMT from the main
>> tzdb, and moving it to a separate lookup file. Following this
>> discussion I'm increasingly of the opinion that my current use of
>> naked LMT is a mistake (because the LMT data has been deleted and made
>> inaccurate in the past, so cannot be relied on).
>
> Even if it *hadn't* been deleted, you can't use the TZ database to get LMT for arbitrary locations, so it's *still* a mistake.
>
> Frankly, it's a mistake that it was ever in the TZ database in the first place, given that a "time zone" only has a valid LMT value if its easternmost and westernmost edges are between the same one-second-apart meridians, and even *then* that's true only if you limit the resolution of LMT to one second.
>
> One might be incorrectly led to believe that "Europe/Rome" is a name that refers to the city of Rome rather than to the time zone the most-populous city of which is Rome, but that's not so, and saying it has Rome's notion of LMT is bogus unless all of the locations in that time zone have LMT values that are the same (to a resolution of one second; with sufficiently fine resolution you can't even say *Rome* has *an* LMT value).  (And even if that's the case, it probably falls apart in, for example, North American time zones.)

Yep, LMT is clearly dodgy, but its the best we have. The core problem
is what does the tzdb say that the offset was for times before the
earliest zone. In reality, we know that people didn't really use
clocks and just used local sun time. LMT is one encoding of that.
Another would be a rounded to the nearest hour/half-hour value, which
would be more arbitrary yet also more clearly arbitrary.

Note that saying the tzdb starts at 1970 doesn't solve the problem. In
both cases, users like me still need to have an offset for a zoneID in
the far past.

Stephen