[tz] Definition for timezone Asia/Urumqi and Asia/Kashgar, CST, and other questions about Chinese timezones.

[Robert Elz]

    Date:        Sun, 1 Jan 2017 12:28:00 +0800
    From:        gfb hjjhjh <c933103 at gmail.com>
    Message-ID:  <CAGHjPPKc5fBCgSeU2sxDx_pbhfsJVJZ6LbGowjKO=DbmW47x8Q at mail.gmail.com>

  | 1. Currently, Asia/Urumqi and Asia/Kashgar are defined as UTC+6, covering
  | most of Xinjiang and Tibet.

I am not going to comment on this one, as I have no info on what people
actually use in the affected areas, but if you can collect any evidence of
what is actually done, that is the kind of information that we seek (direct
evidence from affacted people is usually best).

  | 2. Actually, which file define the boundary of timezones in tz database? I
  | see they are in comments for China but how about generally?

There is none.   The boundaries fluctuate depending on all kinds of things
(there may be some official boundaries, but we don't care much about official
and are more concerened with actual practice.)  There are others who attempt
to draw boundaries, but that's not part of this project.   We are concerned
with which different timezones exist, not with who should use each one.

  | 3. And given Asia/Urumqi is currently defined differently from
  | Asia/Shanghai, and Maoming, Guandong is now using UTC+8, no matter what
  | historical reason might be behind the consideration of putting Maoming,
  | Guangdong in Asia/Urumqi, its current time usage along should have
  | justified putting it in Asia/Shanghai instead

History matters, not just current times - timestamps from the past need to
be interpreted correctly.   Many people are content setting their timezone
to aything that makes the clock show the correct valuee, but that is really
not good.

  | 4. as the name "Central Standard Time" in China was created by ROC,

I always assmed that "CST" in China was "China Standard Time" - but these
names are largely meaningless.

  | 5. Why there are 5 different zones in China when the theory file say "if
  | all the clocks in the region have agreed since 1970, don't bother to
  | inclide more than one location even if subregions' clocks disagreed before
  | 1970" ?

We don't generally create a new timezone if the only difference we have is
pre 1970, but sometimes...  There have been zome zone mergers because of this
(which personally I do not like),   Basically, more accurate timezones are
better, provided the data is correct - the big cost is in maintenance
(a change might affect several zones, rather than just one).

kre