[tz] Adding Asia/Beijing timezone into the database

[Daoming Qiu]

"Beijing Time" is the standard time used for China region, so
it is very strange that tzdatabase doesn't have timezone called Asia/Beijing.
Strongly suggest to add "Asia/Beijing" zone into the database,
and its information is very similar with "Asia/Shanghai".

[Philip Newton]

Post by Daoming Qiu
"Beijing Time" is the standard time used for China region, so
?it is very strange that tzdatabase doesn't have timezone called Asia/Beijing.

The database does not follow, in general, country lines but rather
regions that use the same timezone during the same period of time. And
because cities change their names less often than countries do, the
database usually uses city names (rather than country names) to
identify such timezone regions, and in general (in my understanding)
chooses the largest city within a timezone region to represent that
region. This favours Shanghai over Beijing.

You will note that there is also no America/Washington_DC, for
example, or Australia/Canberra, or America/Brasilia.

See also the "Theory" file in the tzcode distribution; this mentions
some of the things taken into account when choosing names of time
zones. (And it even has Shanghai/Beijing as an example: "Use the most
populous among locations in a country's time zone, e.g. prefer
`Shanghai' to `Beijing'.")

Post by Daoming Qiu
Strongly suggest to add "Asia/Beijing" zone into the database,
and its information is very similar with "Asia/Shanghai".

If it is only "similar", can you please provide information on when
Beijing used different time from Shanghai after January 1970? Then the
tz database can be corrected and completed.

Final disclaimer: I'm not a tz database maintainer and do not
represent anyone maintaining it or setting the rules. The above is my
understanding and interpretation.

Cheers,
Philip

[Daoming Qiu]

From local custom and developer's preference, "Asia/Beijing" is natural choice.

Beijing has almost same population as Shanghai, but there are very few
people here to talk about Shanghai timezone.
In windows or linux machine, you can only see "GMT+8 Beijing,
Chongqing,HongKong,Urumuqi" in the timezone panel list,
(there is no shanghai here).

I am not an expert on time issues, and following is suggested zone
info for Asia/Beijing:
# Zone NAME GMTOFF RULES FORMAT [UNTIL]
Zone Asia/Beijing 7:45:32 - LMT 1928
8:00 PRC C%sT

Thanks,
Daoming

Post by Daoming Qiu
"Beijing Time" is the standard time used for China region, so
?it is very strange that tzdatabase doesn't have timezone called Asia/Beijing.

The database does not follow, in general, country lines but rather
regions that use the same timezone during the same period of time. And
because cities change their names less often than countries do, the
database usually uses city names (rather than country names) to
identify such timezone regions, and in general (in my understanding)
chooses the largest city within a timezone region to represent that
region. This favours Shanghai over Beijing.
You will note that there is also no America/Washington_DC, for
example, or Australia/Canberra, or America/Brasilia.
See also the "Theory" file in the tzcode distribution; this mentions
some of the things taken into account when choosing names of time
zones. (And it even has Shanghai/Beijing as an example: "Use the most
populous among locations in a country's time zone, e.g. prefer
`Shanghai' to `Beijing'.")

Post by Daoming Qiu
Strongly suggest to add "Asia/Beijing" zone into the database,
and its information is very similar with "Asia/Shanghai".

If it is only "similar", can you please provide information on when
Beijing used different time from Shanghai after January 1970? Then the
tz database can be corrected and completed.
Final disclaimer: I'm not a tz database maintainer and do not
represent anyone maintaining it or setting the rules. The above is my
understanding and interpretation.
Cheers,
Philip
--
Philip Newton <philip.newton at gmail.com>

[Eliot Lear]

I believe this issue has been discussed repeatedly on this list, the
archives of which are available for your perusal. My suggestion is that
until circumstances change that might impact the decision process, we
not revisit the whole argument. If those circumstances have changed,
perhaps those proposing the change should indicate what they are.

Eliot

Post by Daoming Qiu

From local custom and developer's preference, "Asia/Beijing" is natural choice.

Beijing has almost same population as Shanghai, but there are very few
people here to talk about Shanghai timezone.
In windows or linux machine, you can only see "GMT+8 Beijing,
Chongqing,HongKong,Urumuqi" in the timezone panel list,
(there is no shanghai here).
I am not an expert on time issues, and following is suggested zone
# Zone NAME GMTOFF RULES FORMAT [UNTIL]
Zone Asia/Beijing 7:45:32 - LMT 1928
8:00 PRC C%sT
Thanks,
Daoming

Post by Daoming Qiu
"Beijing Time" is the standard time used for China region, so
it is very strange that tzdatabase doesn't have timezone called Asia/Beijing.

The database does not follow, in general, country lines but rather
regions that use the same timezone during the same period of time. And
because cities change their names less often than countries do, the
database usually uses city names (rather than country names) to
identify such timezone regions, and in general (in my understanding)
chooses the largest city within a timezone region to represent that
region. This favours Shanghai over Beijing.
You will note that there is also no America/Washington_DC, for
example, or Australia/Canberra, or America/Brasilia.
See also the "Theory" file in the tzcode distribution; this mentions
some of the things taken into account when choosing names of time
zones. (And it even has Shanghai/Beijing as an example: "Use the most
populous among locations in a country's time zone, e.g. prefer
`Shanghai' to `Beijing'.")

Post by Daoming Qiu
Strongly suggest to add "Asia/Beijing" zone into the database,
and its information is very similar with "Asia/Shanghai".

If it is only "similar", can you please provide information on when
Beijing used different time from Shanghai after January 1970? Then the
tz database can be corrected and completed.
Final disclaimer: I'm not a tz database maintainer and do not
represent anyone maintaining it or setting the rules. The above is my
understanding and interpretation.
Cheers,
Philip
--
Philip Newton <philip.newton at gmail.com>

[Bill Seymour]

Link Asia/Shanghai Asia/Beijing

Would that suffice? Or do you really mean to say
that Beijing has obeyed PRC rules since 1928?

--Bill

[Daoming Qiu]

Thanks for all your information.

The fact that this issue has been discussed so many times reflects
that what the user prefers,
and this unreasonable point in the current tz database.
If this issue has not been fixed, it will continue to be raised in the future.

I browsed through the list archive, and I don't agree with the discussion point.
- The timezones listed in the tz database are important and directly
referenced by many kinds of applications.
Requiring applications to do specific mapping (Asia/Shanghai->
Beijing) is not good.
- Using New_York/Washington_DC as selection example, to determine
Shanghai/Beijing selection is not good.
Both city pairs are quite different.
- For a long time(at least recent 100 years), Beijing is used time
reference point, and Shanghai seldom.
I know that tz database uses city's population as the only selection
criteria, but there should be other aspects weighs. tz database should
listen to the voice of the user to be more REASONABLE.

Thanks,
Daoming

Link Asia/Shanghai Asia/Beijing
Would that suffice? Or do you really mean to say
that Beijing has obeyed PRC rules since 1928?
--Bill
I believe this issue has been discussed repeatedly on this list, the
archives of which are available for your perusal. My suggestion is that
until circumstances change that might impact the decision process, we
not revisit the whole argument. If those circumstances have changed,
perhaps those proposing the change should indicate what they are.
Eliot
Have you looked in the list archives? This exact issue has been discussed
many times.
--Ted

[Jaakko Hyvätti]

It's just a code. Same kind of apparently stupid things get added to
database all the time, like this:

SX [phone] America/Lower_Princes

.. Where Lower Princes is more like a suburb of Sint Maarten capital city,
Philipsburg, and should not have been used even as a code. Any reasonable
user interface should use something else.

Maybe someone could maintain a database of reasonable translations in many
languages/locales for each timezone?

Regards,
Jaakko

Post by Daoming Qiu
Thanks for all your information.
The fact that this issue has been discussed so many times reflects
that what the user prefers,
and this unreasonable point in the current tz database.

[Guy Harris]

Post by Daoming Qiu
Thanks for all your information.
The fact that this issue has been discussed so many times reflects
that what the user prefers,

Who are the "users" of the time zone database?

The time zone names are not intended to be directly presented to users other than UN*X command-line users who are directly setting the TZ environment variable.

Post by Daoming Qiu
I browsed through the list archive, and I don't agree with the discussion point.
- The timezones listed in the tz database are important and directly
referenced by many kinds of applications.
Requiring applications to do specific mapping (Asia/Shanghai->
Beijing) is not good.

So would the applications instead have to map Shanghai to Asia/Beijing? In *either* case, they have to map, say, Guangzhou to some other city's name, so there's still mapping.

There should perhaps be a database or databases that can be used to map, say, city names, province/county/etc. names (at least in cases where all of the province/county/etc. is covered by a single tz database entry), latitude/longitude values, etc. to tz database entry names. Perhaps there should also be a database giving names to tz database entries for human use - to use your earlier example, "Beijing, Chongqing, HongKong, Urumuqi" for Asia/{whatever city is used}, although not all systems would choose the same scheme to describe the region covered by a database entry (Mac OS X has a large list of "nearest cities").

Perhaps we should just assign UUIDs to the tz database entries; that should eliminate any complaints about the wrong name being chosen for a database entry. :-)

[yoshito_umaoka]

Post by Guy Harris
Perhaps we should just assign UUIDs to the tz database entries; that
should eliminate any complaints about the wrong name being chosen
for a database entry. :-)

A while ago, I found wikipedia article [
http://en.wikipedia.org/wiki/Tz_database] says -

Post by Guy Harris
Use in other standards
CLDR uses UN/LOCODEs to identify regions.[24] This means all identifiers
are referencing a country, something that the creators of the tz database
wanted to avoid.

I'm actually the person who proposed this. But our (CLDR community) goal
was just assigning unique ID to each "canonical" zone. (BTW, we need such
unique IDs because of the BCP 47 subtag restriction.) If there is an
established coding standard that can reasonably map most of zones
uniquely, I thought I should utilize the standard. It's true that
UN/LOCODES uses ISO country codes, but once it is captured into CLDR name
space, it's nothing more than ID string. Even a city moves to another
country, we'll never change the ID in CLDR.

-Yoshito
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mm.icann.org/pipermail/tz/attachments/20111209/bd60eed2/attachment.html

[lennox]

Post by Guy Harris
Perhaps there should also be a database giving names to tz database entries for human use - to use your earlier example, "Beijing, Chongqing, HongKong, Urumuqi" for Asia/{whatever city is used}, although not all systems would choose the same scheme to describe the region covered by a database entry (Mac OS X has a large list of "nearest cities").

That's what zone.tab is:

CN +3114+12128 Asia/Shanghai east China - Beijing, Guangdong, Shanghai, etc.
CN +4545+12641 Asia/Harbin Heilongjiang (except Mohe), Jilin
CN +2934+10635 Asia/Chongqing central China - Sichuan, Yunnan, Guangxi, Shaanxi, Guizhou, etc.
CN +4348+08735 Asia/Urumqi most of Tibet & Xinjiang
CN +3929+07559 Asia/Kashgar west Tibet & Xinjiang

The descriptions in zone.tab are what is intended to be presented to humans
(and which should be translated by CLDR and the like); the TZID values are
just labels.

[JJS]

*Hello,*
*
*
*unless I've missed something in this long thread, there has been (to my
knowledge) no mention made of Taiwan and cities thereof. Although it lies
in the time zone commonly known as CST (China Standard Time) Taiwan, for
all practical purposes, is distinct from other entities in the same time
zone, viz. Australia (Perth), China ??, Malaysia, Philippines, Singapore.
Without revisiting the respective positions of Beijing and Taipei regarding
what China calls its "one China policy" under which Taiwan is its 23rd
province, the practical realities of Taiwan, as well as its political,
monetary and other features, make it separate from China. And if the
demographic criterion determines the inclusion of a country and/or city in
the time zone list, Taiwan is the sixteenth most densely populated country
on earth, and Taipei's population registers above 2.6 million (2006 census).
*
*
*
*In light of the above, and simply as someone who often has to take
different time zones into account, I suggest that, for practical reasons,
Taiwan ?? (or ?? <http://fr.wiktionary.org/wiki/%E5%8F%B0%E6%B9%BE> in
simplified orthography) and at least the city of Taipei ?? (??) be given
the same treatment as, say, Beijing ??, Kuala Lumpur, Manila, Perth, or
Singapore.*
*
*
*Regards,*
*Jean-Jacques Subrenat.*
*
*
2011/12/10 <lennox at cs.columbia.edu>

Post by Guy Harris

Post by Guy Harris
Perhaps there should also be a database giving names to tz database

entries for human use - to use your earlier example, "Beijing, Chongqing,
HongKong, Urumuqi" for Asia/{whatever city is used}, although not all
systems would choose the same scheme to describe the region covered by a
database entry (Mac OS X has a large list of "nearest cities").
CN +3114+12128 Asia/Shanghai east China - Beijing, Guangdong, Shanghai, etc.
CN +4545+12641 Asia/Harbin Heilongjiang (except Mohe), Jilin
CN +2934+10635 Asia/Chongqing central China - Sichuan, Yunnan,
Guangxi, Shaanxi, Guizhou, etc.
CN +4348+08735 Asia/Urumqi most of Tibet & Xinjiang
CN +3929+07559 Asia/Kashgar west Tibet & Xinjiang
The descriptions in zone.tab are what is intended to be presented to humans
(and which should be translated by CLDR and the like); the TZID values are
just labels.
--
Jonathan Lennox
lennox at cs.columbia.edu

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mm.icann.org/pipermail/tz/attachments/20111210/e5af224a/attachment.html

[Russ Allbery]

Post by JJS
In light of the above, and simply as someone who often has to take
different time zones into account, I suggest that, for practical
reasons, Taiwan ?? (or ??
<http://fr.wiktionary.org/wiki/%E5%8F%B0%E6%B9%BE> in simplified
orthography) and at least the city of Taipei ?? (??) be given the
same treatment as, say, Beijing ??, Kuala Lumpur, Manila, Perth, or
Singapore.

Asia/Taipei is already there, so far as I can tell, for the standard
reason that its time has varied from other zones since 1970. Maybe you
didn't check the database before sending this message?

[lennox]

Following up to Russ's comments, the section I exerpted from zone.tab only
includes the zones with the ISO 3166 code "CN". Asia/Taipei has ISO 3166
code "TW" (and is the only zone with that code).

Post by JJS
*Hello,*
*
*
*unless I've missed something in this long thread, there has been (to my
knowledge) no mention made of Taiwan and cities thereof. Although it lies
in the time zone commonly known as CST (China Standard Time) Taiwan, for
all practical purposes, is distinct from other entities in the same time
zone, viz. Australia (Perth), China $BCf9q(B, Malaysia, Philippines, Singapore.
Without revisiting the respective positions of Beijing and Taipei regarding
what China calls its "one China policy" under which Taiwan is its 23rd
province, the practical realities of Taiwan, as well as its political,
monetary and other features, make it separate from China. And if the
demographic criterion determines the inclusion of a country and/or city in
the time zone list, Taiwan is the sixteenth most densely populated country
on earth, and Taipei's population registers above 2.6 million (2006 census).
*
*
*
*In light of the above, and simply as someone who often has to take
different time zones into account, I suggest that, for practical reasons,
Taiwan $BgJ_T(B (or $BBfOQ(B <http://fr.wiktionary.org/wiki/%E5%8F%B0%E6%B9%BE> in
simplified orthography) and at least the city of Taipei $BgJKL(B ($BBfKL(B) be given
the same treatment as, say, Beijing $BKL5~(B, Kuala Lumpur, Manila, Perth, or
Singapore.*
*
*
*Regards,*
*Jean-Jacques Subrenat.*
*
*
2011/12/10 <lennox at cs.columbia.edu>

Post by Guy Harris

Post by Guy Harris
Perhaps there should also be a database giving names to tz database

entries for human use - to use your earlier example, "Beijing, Chongqing,
HongKong, Urumuqi" for Asia/{whatever city is used}, although not all
systems would choose the same scheme to describe the region covered by a
database entry (Mac OS X has a large list of "nearest cities").
CN +3114+12128 Asia/Shanghai east China - Beijing, Guangdong, Shanghai, etc.
CN +4545+12641 Asia/Harbin Heilongjiang (except Mohe), Jilin
CN +2934+10635 Asia/Chongqing central China - Sichuan, Yunnan,
Guangxi, Shaanxi, Guizhou, etc.
CN +4348+08735 Asia/Urumqi most of Tibet & Xinjiang
CN +3929+07559 Asia/Kashgar west Tibet & Xinjiang
The descriptions in zone.tab are what is intended to be presented to humans
(and which should be translated by CLDR and the like); the TZID values are
just labels.
--
Jonathan Lennox

<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex;">On Friday, December 9 2011, &quot;Guy Harris&quot; wrote to &quot;Daoming Qiu, <a href="mailto:tz at iana.org">tz at iana.org</a>&quot; saying:<br>
<div class="im"><br>
&gt; Perhaps there should also be a database giving names to tz database entries for human use - to use your earlier example, &quot;Beijing, Chongqing, HongKong, Urumuqi&quot; for Asia/{whatever city is used}, although not all systems would choose the same scheme to describe the region covered by a database entry (Mac OS X has a large list of &quot;nearest cities&quot;).<br>
<br>
</div>That&#39;s what zone.tab is:<br>
<br>
CN ,A (B ,A (B ,A (B+3114+12128 ,A (B ,A (B Asia/Shanghai ,A (B east China - Beijing, Guangdong, Shanghai, etc.<br>
CN ,A (B ,A (B ,A (B+4545+12641 ,A (B ,A (B Asia/Harbin ,A (B ,A (B Heilongjiang (except Mohe), Jilin<br>
CN ,A (B ,A (B ,A (B+2934+10635 ,A (B ,A (B Asia/Chongqing ,A (Bcentral China - Sichuan, Yunnan, Guangxi, Shaanxi, Guizhou, etc.<br>
CN ,A (B ,A (B ,A (B+4348+08735 ,A (B ,A (B Asia/Urumqi ,A (B ,A (B most of Tibet &amp; Xinjiang<br>
CN ,A (B ,A (B ,A (B+3929+07559 ,A (B ,A (B Asia/Kashgar ,A (B ,A (Bwest Tibet &amp; Xinjiang<br>
<br>
The descriptions in zone.tab are what is intended to be presented to humans<br>
(and which should be translated by CLDR and the like); the TZID values are<br>
just labels.<br>
<span class="HOEnZb"><font color="#888888"><br>
--<br>
Jonathan Lennox<br>
<a href="mailto:lennox at cs.columbia.edu">lennox at cs.columbia.edu</a><br>
</font></span></blockquote></div><br></div>

[Russ Allbery]

Post by JJS
In light of the above, and simply as someone who often has to take
different time zones into account, I suggest that, for practical
reasons, Taiwan ?? (or ??
<http://fr.wiktionary.org/wiki/%E5%8F%B0%E6%B9%BE> in simplified
orthography) and at least the city of Taipei ?? (??) be given the
same treatment as, say, Beijing ??, Kuala Lumpur, Manila, Perth, or
Singapore.

Asia/Taipei is already there, so far as I can tell, for the standard
reason that its time has varied from other zones since 1970. Maybe you
didn't check the database before sending this message?

[lennox]

Following up to Russ's comments, the section I exerpted from zone.tab only
includes the zones with the ISO 3166 code "CN". Asia/Taipei has ISO 3166
code "TW" (and is the only zone with that code).

Post by JJS
*Hello,*
*
*
*unless I've missed something in this long thread, there has been (to my
knowledge) no mention made of Taiwan and cities thereof. Although it lies
in the time zone commonly known as CST (China Standard Time) Taiwan, for
all practical purposes, is distinct from other entities in the same time
zone, viz. Australia (Perth), China $BCf9q(B, Malaysia, Philippines, Singapore.
Without revisiting the respective positions of Beijing and Taipei regarding
what China calls its "one China policy" under which Taiwan is its 23rd
province, the practical realities of Taiwan, as well as its political,
monetary and other features, make it separate from China. And if the
demographic criterion determines the inclusion of a country and/or city in
the time zone list, Taiwan is the sixteenth most densely populated country
on earth, and Taipei's population registers above 2.6 million (2006 census).
*
*
*
*In light of the above, and simply as someone who often has to take
different time zones into account, I suggest that, for practical reasons,
Taiwan $BgJ_T(B (or $BBfOQ(B <http://fr.wiktionary.org/wiki/%E5%8F%B0%E6%B9%BE> in
simplified orthography) and at least the city of Taipei $BgJKL(B ($BBfKL(B) be given
the same treatment as, say, Beijing $BKL5~(B, Kuala Lumpur, Manila, Perth, or
Singapore.*
*
*
*Regards,*
*Jean-Jacques Subrenat.*
*
*
2011/12/10 <lennox at cs.columbia.edu>

Post by Guy Harris

Post by Guy Harris
Perhaps there should also be a database giving names to tz database

entries for human use - to use your earlier example, "Beijing, Chongqing,
HongKong, Urumuqi" for Asia/{whatever city is used}, although not all
systems would choose the same scheme to describe the region covered by a
database entry (Mac OS X has a large list of "nearest cities").
CN +3114+12128 Asia/Shanghai east China - Beijing, Guangdong, Shanghai, etc.
CN +4545+12641 Asia/Harbin Heilongjiang (except Mohe), Jilin
CN +2934+10635 Asia/Chongqing central China - Sichuan, Yunnan,
Guangxi, Shaanxi, Guizhou, etc.
CN +4348+08735 Asia/Urumqi most of Tibet & Xinjiang
CN +3929+07559 Asia/Kashgar west Tibet & Xinjiang
The descriptions in zone.tab are what is intended to be presented to humans
(and which should be translated by CLDR and the like); the TZID values are
just labels.
--
Jonathan Lennox

<blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex;">On Friday, December 9 2011, &quot;Guy Harris&quot; wrote to &quot;Daoming Qiu, <a href="mailto:tz at iana.org">tz at iana.org</a>&quot; saying:<br>
<div class="im"><br>
&gt; Perhaps there should also be a database giving names to tz database entries for human use - to use your earlier example, &quot;Beijing, Chongqing, HongKong, Urumuqi&quot; for Asia/{whatever city is used}, although not all systems would choose the same scheme to describe the region covered by a database entry (Mac OS X has a large list of &quot;nearest cities&quot;).<br>
<br>
</div>That&#39;s what zone.tab is:<br>
<br>
CN ,A (B ,A (B ,A (B+3114+12128 ,A (B ,A (B Asia/Shanghai ,A (B east China - Beijing, Guangdong, Shanghai, etc.<br>
CN ,A (B ,A (B ,A (B+4545+12641 ,A (B ,A (B Asia/Harbin ,A (B ,A (B Heilongjiang (except Mohe), Jilin<br>
CN ,A (B ,A (B ,A (B+2934+10635 ,A (B ,A (B Asia/Chongqing ,A (Bcentral China - Sichuan, Yunnan, Guangxi, Shaanxi, Guizhou, etc.<br>
CN ,A (B ,A (B ,A (B+4348+08735 ,A (B ,A (B Asia/Urumqi ,A (B ,A (B most of Tibet &amp; Xinjiang<br>
CN ,A (B ,A (B ,A (B+3929+07559 ,A (B ,A (B Asia/Kashgar ,A (B ,A (Bwest Tibet &amp; Xinjiang<br>
<br>
The descriptions in zone.tab are what is intended to be presented to humans<br>
(and which should be translated by CLDR and the like); the TZID values are<br>
just labels.<br>
<span class="HOEnZb"><font color="#888888"><br>
--<br>
Jonathan Lennox<br>
<a href="mailto:lennox at cs.columbia.edu">lennox at cs.columbia.edu</a><br>
</font></span></blockquote></div><br></div>

[JJS]

*Hello,*
*
*
*unless I've missed something in this long thread, there has been (to my
knowledge) no mention made of Taiwan and cities thereof. Although it lies
in the time zone commonly known as CST (China Standard Time) Taiwan, for
all practical purposes, is distinct from other entities in the same time
zone, viz. Australia (Perth), China ??, Malaysia, Philippines, Singapore.
Without revisiting the respective positions of Beijing and Taipei regarding
what China calls its "one China policy" under which Taiwan is its 23rd
province, the practical realities of Taiwan, as well as its political,
monetary and other features, make it separate from China. And if the
demographic criterion determines the inclusion of a country and/or city in
the time zone list, Taiwan is the sixteenth most densely populated country
on earth, and Taipei's population registers above 2.6 million (2006 census).
*
*
*
*In light of the above, and simply as someone who often has to take
different time zones into account, I suggest that, for practical reasons,
Taiwan ?? (or ?? <http://fr.wiktionary.org/wiki/%E5%8F%B0%E6%B9%BE> in
simplified orthography) and at least the city of Taipei ?? (??) be given
the same treatment as, say, Beijing ??, Kuala Lumpur, Manila, Perth, or
Singapore.*
*
*
*Regards,*
*Jean-Jacques Subrenat.*
*
*
2011/12/10 <lennox at cs.columbia.edu>

Post by Guy Harris

Post by Guy Harris
Perhaps there should also be a database giving names to tz database

entries for human use - to use your earlier example, "Beijing, Chongqing,
HongKong, Urumuqi" for Asia/{whatever city is used}, although not all
systems would choose the same scheme to describe the region covered by a
database entry (Mac OS X has a large list of "nearest cities").
CN +3114+12128 Asia/Shanghai east China - Beijing, Guangdong, Shanghai, etc.
CN +4545+12641 Asia/Harbin Heilongjiang (except Mohe), Jilin
CN +2934+10635 Asia/Chongqing central China - Sichuan, Yunnan,
Guangxi, Shaanxi, Guizhou, etc.
CN +4348+08735 Asia/Urumqi most of Tibet & Xinjiang
CN +3929+07559 Asia/Kashgar west Tibet & Xinjiang
The descriptions in zone.tab are what is intended to be presented to humans
(and which should be translated by CLDR and the like); the TZID values are
just labels.
--
Jonathan Lennox
lennox at cs.columbia.edu

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mm.icann.org/pipermail/tz/attachments/20111210/e5af224a/attachment-0002.html

[Random832]

Who are the "users" of the time zone database? The time zone names are
not intended to be directly presented to users other than UN*X
command-line users who are directly setting the TZ environment variable.

Maybe this should be better explained to developers who write time zone
selection tools.

Also, a map with points corresponding to only the cities has the same
problem as presenting the names. Really, the shape data should be
incorporated into the project. (along with some guidance on what to do
when shape data causes problems for disputed borders... which may not be
a problem for the project itself, but can be for distributors; Microsoft
ran into this with a disputed border and Windows 95 was almost banned in
India.)

Perhaps there should also be a database giving names to tz database
entries for human use - to use your earlier example, "Beijing,
Chongqing, HongKong, Urumuqi" for Asia/{whatever city is used}

In fact, such a database does exist, and Asia/Shanghai's entry in it is
"CN / east China - Beijing, Guangdong, Shanghai, etc." (And,
incidentally, Chongqing, Hong_Kong, and Urmuqui are all separate
timezones, so I think you misunderstood something)

[Daoming Qiu]

One example of the user of the tz database is Oracle JDK(Java Dev Kit).
Currently Java applications will get the time zone IDs from the tz database.
So if you call java.util.TimeZone.getTimeZone("Asia/Beijing"), it will
fail in current impl.
calling java.util.TimeZone.getTimeZone("Asia/Shanghai") works.
But for Java developers, using "Asia/Beijing" is more friendly.

Post by Guy Harris

Post by Daoming Qiu
Thanks for all your information.
The fact that this issue has been discussed so many times reflects
that what the user prefers,

Who are the "users" of the time zone database?
The time zone names are not intended to be directly presented to users other than UN*X command-line users who are directly setting the TZ environment variable.

Post by Daoming Qiu
I browsed through the list archive, and I don't agree with the discussion point.
- The timezones listed in the tz database are important and directly
referenced by many kinds of applications.
Requiring applications to do specific mapping (Asia/Shanghai->
Beijing) ?is not good.

So would the applications instead have to map Shanghai to Asia/Beijing? ?In *either* case, they have to map, say, Guangzhou to some other city's name, so there's still mapping.
There should perhaps be a database or databases that can be used to map, say, city names, province/county/etc. names (at least in cases where all of the province/county/etc. is covered by a single tz database entry), latitude/longitude values, etc. to tz database entry names. ?Perhaps there should also be a database giving names to tz database entries for human use - to use your earlier example, "Beijing, Chongqing, HongKong, Urumuqi" for Asia/{whatever city is used}, although not all systems would choose the same scheme to describe the region covered by a database entry (Mac OS X has a large list of "nearest cities").
Perhaps we should just assign UUIDs to the tz database entries; that should eliminate any complaints about the wrong name being chosen for a database entry. :-)

[yoshito_umaoka]

Post by Guy Harris
Perhaps we should just assign UUIDs to the tz database entries; that
should eliminate any complaints about the wrong name being chosen
for a database entry. :-)

A while ago, I found wikipedia article [
http://en.wikipedia.org/wiki/Tz_database] says -

Post by Guy Harris
Use in other standards
CLDR uses UN/LOCODEs to identify regions.[24] This means all identifiers
are referencing a country, something that the creators of the tz database
wanted to avoid.

I'm actually the person who proposed this. But our (CLDR community) goal
was just assigning unique ID to each "canonical" zone. (BTW, we need such
unique IDs because of the BCP 47 subtag restriction.) If there is an
established coding standard that can reasonably map most of zones
uniquely, I thought I should utilize the standard. It's true that
UN/LOCODES uses ISO country codes, but once it is captured into CLDR name
space, it's nothing more than ID string. Even a city moves to another
country, we'll never change the ID in CLDR.

-Yoshito
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mm.icann.org/pipermail/tz/attachments/20111209/bd60eed2/attachment-0002.html

[lennox]

Post by Guy Harris
Perhaps there should also be a database giving names to tz database entries for human use - to use your earlier example, "Beijing, Chongqing, HongKong, Urumuqi" for Asia/{whatever city is used}, although not all systems would choose the same scheme to describe the region covered by a database entry (Mac OS X has a large list of "nearest cities").

That's what zone.tab is:

CN +3114+12128 Asia/Shanghai east China - Beijing, Guangdong, Shanghai, etc.
CN +4545+12641 Asia/Harbin Heilongjiang (except Mohe), Jilin
CN +2934+10635 Asia/Chongqing central China - Sichuan, Yunnan, Guangxi, Shaanxi, Guizhou, etc.
CN +4348+08735 Asia/Urumqi most of Tibet & Xinjiang
CN +3929+07559 Asia/Kashgar west Tibet & Xinjiang

The descriptions in zone.tab are what is intended to be presented to humans
(and which should be translated by CLDR and the like); the TZID values are
just labels.

[Random832]

Who are the "users" of the time zone database? The time zone names are
not intended to be directly presented to users other than UN*X
command-line users who are directly setting the TZ environment variable.

Maybe this should be better explained to developers who write time zone
selection tools.

Also, a map with points corresponding to only the cities has the same
problem as presenting the names. Really, the shape data should be
incorporated into the project. (along with some guidance on what to do
when shape data causes problems for disputed borders... which may not be
a problem for the project itself, but can be for distributors; Microsoft
ran into this with a disputed border and Windows 95 was almost banned in
India.)

Perhaps there should also be a database giving names to tz database
entries for human use - to use your earlier example, "Beijing,
Chongqing, HongKong, Urumuqi" for Asia/{whatever city is used}

In fact, such a database does exist, and Asia/Shanghai's entry in it is
"CN / east China - Beijing, Guangdong, Shanghai, etc." (And,
incidentally, Chongqing, Hong_Kong, and Urmuqui are all separate
timezones, so I think you misunderstood something)

[Daoming Qiu]

One example of the user of the tz database is Oracle JDK(Java Dev Kit).
Currently Java applications will get the time zone IDs from the tz database.
So if you call java.util.TimeZone.getTimeZone("Asia/Beijing"), it will
fail in current impl.
calling java.util.TimeZone.getTimeZone("Asia/Shanghai") works.
But for Java developers, using "Asia/Beijing" is more friendly.

Post by Guy Harris

Post by Daoming Qiu
Thanks for all your information.
The fact that this issue has been discussed so many times reflects
that what the user prefers,

Who are the "users" of the time zone database?
The time zone names are not intended to be directly presented to users other than UN*X command-line users who are directly setting the TZ environment variable.

Post by Daoming Qiu
I browsed through the list archive, and I don't agree with the discussion point.
- The timezones listed in the tz database are important and directly
referenced by many kinds of applications.
Requiring applications to do specific mapping (Asia/Shanghai->
Beijing) ?is not good.

So would the applications instead have to map Shanghai to Asia/Beijing? ?In *either* case, they have to map, say, Guangzhou to some other city's name, so there's still mapping.
There should perhaps be a database or databases that can be used to map, say, city names, province/county/etc. names (at least in cases where all of the province/county/etc. is covered by a single tz database entry), latitude/longitude values, etc. to tz database entry names. ?Perhaps there should also be a database giving names to tz database entries for human use - to use your earlier example, "Beijing, Chongqing, HongKong, Urumuqi" for Asia/{whatever city is used}, although not all systems would choose the same scheme to describe the region covered by a database entry (Mac OS X has a large list of "nearest cities").
Perhaps we should just assign UUIDs to the tz database entries; that should eliminate any complaints about the wrong name being chosen for a database entry. :-)

[Jaakko Hyvätti]

It's just a code. Same kind of apparently stupid things get added to
database all the time, like this:

SX [phone] America/Lower_Princes

.. Where Lower Princes is more like a suburb of Sint Maarten capital city,
Philipsburg, and should not have been used even as a code. Any reasonable
user interface should use something else.

Maybe someone could maintain a database of reasonable translations in many
languages/locales for each timezone?

Regards,
Jaakko

Post by Daoming Qiu
Thanks for all your information.
The fact that this issue has been discussed so many times reflects
that what the user prefers,
and this unreasonable point in the current tz database.

[Guy Harris]

Post by Daoming Qiu
Thanks for all your information.
The fact that this issue has been discussed so many times reflects
that what the user prefers,

Who are the "users" of the time zone database?

The time zone names are not intended to be directly presented to users other than UN*X command-line users who are directly setting the TZ environment variable.

Post by Daoming Qiu
I browsed through the list archive, and I don't agree with the discussion point.
- The timezones listed in the tz database are important and directly
referenced by many kinds of applications.
Requiring applications to do specific mapping (Asia/Shanghai->
Beijing) is not good.

So would the applications instead have to map Shanghai to Asia/Beijing? In *either* case, they have to map, say, Guangzhou to some other city's name, so there's still mapping.

There should perhaps be a database or databases that can be used to map, say, city names, province/county/etc. names (at least in cases where all of the province/county/etc. is covered by a single tz database entry), latitude/longitude values, etc. to tz database entry names. Perhaps there should also be a database giving names to tz database entries for human use - to use your earlier example, "Beijing, Chongqing, HongKong, Urumuqi" for Asia/{whatever city is used}, although not all systems would choose the same scheme to describe the region covered by a database entry (Mac OS X has a large list of "nearest cities").

Perhaps we should just assign UUIDs to the tz database entries; that should eliminate any complaints about the wrong name being chosen for a database entry. :-)

[Thomas]

Post by Daoming Qiu
- For a long time(at least recent 100 years), Beijing is used time

~~~~~~~~~~~~~~~~~~~~~~~~~Daoming，are you really sure?!

This is the wiki page explaining it:
https://en.wikipedia.org/wiki/Historical_time_zones_of_China

Post by Daoming Qiu
reference point, and Shanghai seldom.
Daoming

[Daoming Qiu]

Thanks for all your information.

The fact that this issue has been discussed so many times reflects
that what the user prefers,
and this unreasonable point in the current tz database.
If this issue has not been fixed, it will continue to be raised in the future.

I browsed through the list archive, and I don't agree with the discussion point.
- The timezones listed in the tz database are important and directly
referenced by many kinds of applications.
Requiring applications to do specific mapping (Asia/Shanghai->
Beijing) is not good.
- Using New_York/Washington_DC as selection example, to determine
Shanghai/Beijing selection is not good.
Both city pairs are quite different.
- For a long time(at least recent 100 years), Beijing is used time
reference point, and Shanghai seldom.
I know that tz database uses city's population as the only selection
criteria, but there should be other aspects weighs. tz database should
listen to the voice of the user to be more REASONABLE.

Thanks,
Daoming

Link Asia/Shanghai Asia/Beijing
Would that suffice? Or do you really mean to say
that Beijing has obeyed PRC rules since 1928?
--Bill
I believe this issue has been discussed repeatedly on this list, the
archives of which are available for your perusal. My suggestion is that
until circumstances change that might impact the decision process, we
not revisit the whole argument. If those circumstances have changed,
perhaps those proposing the change should indicate what they are.
Eliot
Have you looked in the list archives? This exact issue has been discussed
many times.
--Ted

[Bill Seymour]

Link Asia/Shanghai Asia/Beijing

Would that suffice? Or do you really mean to say
that Beijing has obeyed PRC rules since 1928?

--Bill

[Eliot Lear]

I believe this issue has been discussed repeatedly on this list, the
archives of which are available for your perusal. My suggestion is that
until circumstances change that might impact the decision process, we
not revisit the whole argument. If those circumstances have changed,
perhaps those proposing the change should indicate what they are.

Eliot

Post by Daoming Qiu

From local custom and developer's preference, "Asia/Beijing" is natural choice.

Beijing has almost same population as Shanghai, but there are very few
people here to talk about Shanghai timezone.
In windows or linux machine, you can only see "GMT+8 Beijing,
Chongqing,HongKong,Urumuqi" in the timezone panel list,
(there is no shanghai here).
I am not an expert on time issues, and following is suggested zone
# Zone NAME GMTOFF RULES FORMAT [UNTIL]
Zone Asia/Beijing 7:45:32 - LMT 1928
8:00 PRC C%sT
Thanks,
Daoming

Post by Daoming Qiu
"Beijing Time" is the standard time used for China region, so
it is very strange that tzdatabase doesn't have timezone called Asia/Beijing.

The database does not follow, in general, country lines but rather
regions that use the same timezone during the same period of time. And
because cities change their names less often than countries do, the
database usually uses city names (rather than country names) to
identify such timezone regions, and in general (in my understanding)
chooses the largest city within a timezone region to represent that
region. This favours Shanghai over Beijing.
You will note that there is also no America/Washington_DC, for
example, or Australia/Canberra, or America/Brasilia.
See also the "Theory" file in the tzcode distribution; this mentions
some of the things taken into account when choosing names of time
zones. (And it even has Shanghai/Beijing as an example: "Use the most
populous among locations in a country's time zone, e.g. prefer
`Shanghai' to `Beijing'.")

Post by Daoming Qiu
Strongly suggest to add "Asia/Beijing" zone into the database,
and its information is very similar with "Asia/Shanghai".

If it is only "similar", can you please provide information on when
Beijing used different time from Shanghai after January 1970? Then the
tz database can be corrected and completed.
Final disclaimer: I'm not a tz database maintainer and do not
represent anyone maintaining it or setting the rules. The above is my
understanding and interpretation.
Cheers,
Philip
--
Philip Newton <philip.newton at gmail.com>

[Daoming Qiu]

From local custom and developer's preference, "Asia/Beijing" is natural choice.

Beijing has almost same population as Shanghai, but there are very few
people here to talk about Shanghai timezone.
In windows or linux machine, you can only see "GMT+8 Beijing,
Chongqing,HongKong,Urumuqi" in the timezone panel list,
(there is no shanghai here).

I am not an expert on time issues, and following is suggested zone
info for Asia/Beijing:
# Zone NAME GMTOFF RULES FORMAT [UNTIL]
Zone Asia/Beijing 7:45:32 - LMT 1928
8:00 PRC C%sT

Thanks,
Daoming

Post by Daoming Qiu
"Beijing Time" is the standard time used for China region, so
?it is very strange that tzdatabase doesn't have timezone called Asia/Beijing.

The database does not follow, in general, country lines but rather
regions that use the same timezone during the same period of time. And
because cities change their names less often than countries do, the
database usually uses city names (rather than country names) to
identify such timezone regions, and in general (in my understanding)
chooses the largest city within a timezone region to represent that
region. This favours Shanghai over Beijing.
You will note that there is also no America/Washington_DC, for
example, or Australia/Canberra, or America/Brasilia.
See also the "Theory" file in the tzcode distribution; this mentions
some of the things taken into account when choosing names of time
zones. (And it even has Shanghai/Beijing as an example: "Use the most
populous among locations in a country's time zone, e.g. prefer
`Shanghai' to `Beijing'.")

Post by Daoming Qiu
Strongly suggest to add "Asia/Beijing" zone into the database,
and its information is very similar with "Asia/Shanghai".

If it is only "similar", can you please provide information on when
Beijing used different time from Shanghai after January 1970? Then the
tz database can be corrected and completed.
Final disclaimer: I'm not a tz database maintainer and do not
represent anyone maintaining it or setting the rules. The above is my
understanding and interpretation.
Cheers,
Philip
--
Philip Newton <philip.newton at gmail.com>

[Steven Abner]

Hi
Not taking a position, but one possibility as the user of the database is to
install a file or bundle with your application the locale you want. You can easily
and a new zone anytime you feel like deviating from the norm. Just make sure to bundle
with an installer.
Myself, I have patches to the database to create a database. It then gets installed on a
system. I also use the patched files and zic so that, if using a different application, it can
access the Olson database along with any changes or new information.
This doesn't just apply to a database, including compatibly source code has been done
for ages.
Hope this helpful, just add a patch to backwards file with your new link, or bundle with your app,
depending on how you need to use.

Steve

[Daoming Qiu]

"Beijing Time" is the standard time used for China region, so
it is very strange that tzdatabase doesn't have timezone called Asia/Beijing.
Strongly suggest to add "Asia/Beijing" zone into the database,
and its information is very similar with "Asia/Shanghai".

[Philip Newton]

Post by Daoming Qiu
"Beijing Time" is the standard time used for China region, so
?it is very strange that tzdatabase doesn't have timezone called Asia/Beijing.

The database does not follow, in general, country lines but rather
regions that use the same timezone during the same period of time. And
because cities change their names less often than countries do, the
database usually uses city names (rather than country names) to
identify such timezone regions, and in general (in my understanding)
chooses the largest city within a timezone region to represent that
region. This favours Shanghai over Beijing.

You will note that there is also no America/Washington_DC, for
example, or Australia/Canberra, or America/Brasilia.

See also the "Theory" file in the tzcode distribution; this mentions
some of the things taken into account when choosing names of time
zones. (And it even has Shanghai/Beijing as an example: "Use the most
populous among locations in a country's time zone, e.g. prefer
`Shanghai' to `Beijing'.")

Post by Daoming Qiu
Strongly suggest to add "Asia/Beijing" zone into the database,
and its information is very similar with "Asia/Shanghai".

If it is only "similar", can you please provide information on when
Beijing used different time from Shanghai after January 1970? Then the
tz database can be corrected and completed.

Final disclaimer: I'm not a tz database maintainer and do not
represent anyone maintaining it or setting the rules. The above is my
understanding and interpretation.

Cheers,
Philip

[Steven Abner]

Hi
Not taking a position, but one possibility as the user of the database is to
install a file or bundle with your application the locale you want. You can easily
and a new zone anytime you feel like deviating from the norm. Just make sure to bundle
with an installer.
Myself, I have patches to the database to create a database. It then gets installed on a
system. I also use the patched files and zic so that, if using a different application, it can
access the Olson database along with any changes or new information.
This doesn't just apply to a database, including compatibly source code has been done
for ages.
Hope this helpful, just add a patch to backwards file with your new link, or bundle with your app,
depending on how you need to use.

Steve

[Alois Treindl]

It's a while since this had been discussed and the proposed
Link Asia/Shanghai Asia/Beijing
has not been added.

This makes sense.

What makes much less sense is using Shanghai TZ history for most of
China. The city area of Shanghai had (according to Shank, as used in tz)
Daylight saving time in the summers of 1940 and 1941.

The rest of China had definitely no Daylight saving time in those two
years, except for four small city pockets around Wuhan (Hubei), Nanjing
(Jiangsu), Hangzhou (Zhejiang), Suzhou (Jiangsu), which, again according
to Shanks, had DST in 1941, but not in 1940.

In consequence, I propose to create zone Asia/Beijing by copying (not
linking) Asia/Shangai, but leaving oit the DST in 1940 and 1941.

Post by Daoming Qiu
"Beijing Time" is the standard time used for China region, so
it is very strange that tzdatabase doesn't have timezone called Asia/Beijing.
Strongly suggest to add "Asia/Beijing" zone into the database,
and its information is very similar with "Asia/Shanghai".

[Lester Caine]

In consequence, I propose to create zone Asia/Beijing by copying (not linking)
Asia/Shangai, but leaving oit the DST in 1940 and 1941.

Alois ... the rule is that only changes after 1970 are managed, so like a number
of other recent correction/improvements to existing pre 1970 data, this will not
be allowed through. We do need a better way of managing pre-1970 history as it
is important, but currently iana tz is not going to provide that.

I have set up http://timedb.co.uk/ as a base to allow recording of all of the
pre-history stuff, but other activity to keep on top of other upgrades has been
eating up all my spare time. If people have pre-1970 material then copy it over
to me and I'll create pages for it as a starting point.