[Datetime-SIG] Timezone database

[Felipe Ochoa]

>
> It looks like Windows (where a lot of Python users actually are -- just not
>> those who are most active in Open Source)  is left out in the cold here.
>
> sure Windows by now has a timezone info database of its own that is kept
>> up-to-date whenever you install Microsoft updates, and Python should use
>> that in preference over a Python-specific database which is only updated
>> when the user updates Python (which may be never).
>
>
Can't we somehow hook into the Java auto-updater? *wink*

The problem with Windows is that is has its own naming of timezones.
> There are mapping between the two[1], but then you need to make sure
> the mapping is up-to-date (which I think doesn't change very often).
>

Thanks for the pointer to the mapping. I also found the proposal document
<http://cldr.unicode.org/development/development-process/design-proposals/extended-windows-olson-zid-mapping>helpful
in understanding the differences between Windows and Olson time zones.

One design decision that I think has been made but is worth confirming: Do
we want users to specify time zones using the Olson names? I think it's
evident why we have to use consistent names across platforms, but it might
be worth using the 'official' unicode timezone identifiers
<http://www.unicode.org/reports/tr35/tr35.html#Time_Zone_Identifiers>,
which are guaranteed to be stable.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.python.org/pipermail/datetime-sig/attachments/20150729/5f02b42e/attachment.html>