Theory - zone name, zone creation, zone lifetime
Theory reads: To represent this data, the world is partitioned into regions whose clocks all agree about time stamps that occur after the somewhat-arbitrary cutoff point of the POSIX Epoch (1970-01-01 00:00:00 UTC). Issue: 1) Where are these regions listed? The only lists of geographical areas I could find are: - the zones in zone.tab - (current) countries in iso3166.tab Theory reads: Use ISO 3166 (see the file iso3166.tab) to help decide whether something is a country. [...] or when locations change countries (e.g. Hong Kong from UK colony to China) Issue: 1) "help to decide" - decision process is undefined. 2) Hong Kong is listed in iso3166.tab as HK.For the purpose of zone creation is it 2a) a country,then it is confusing to say it changed country, or 2b) is it part of China (CN) then it is confusing to mark it with HK 3) In case HK and one of the zones in CN observed the same time since 1970, and that HK is not a country, would there be a merge and would one zone be dropped from zone.tab and any corresponding pre-1970 history be deleted? Theory reads: Within a file name component, use only ASCII letters Issue: 1) It is not defined what the term "file name component" refers to in the context of zone naming. Theory reads: E.g. prefer `Paris' to `France', since France has had multiple time zones. Issue: 1) It does not matter what zones France "had" in the past. 2) It may matter that Europe/Paris might be split in the future. 3) There is no proof in zone.tab that France had multiple time zones at all, there is only one line marked "FR": FR+4852+00220Europe/Paris -- Tobias Conradi Rheinsberger Str. 18 10115 Berlin, Germany http://tobiasconradi.com
I'm forwarding this message from Tobias Conradi, who is not on the time zone mailing list. Those of you who are on the list, please direct replies appropriately. --ado ________________________________________ From: Mr tobias conradi [tobias_conradi@yahoo.de] Sent: Sunday, September 11, 2011 12:19 AM To: tz@elsie.nci.nih.gov Subject: Theory - zone name, zone creation, zone lifetime Theory reads: To represent this data, the world is partitioned into regions whose clocks all agree about time stamps that occur after the somewhat-arbitrary cutoff point of the POSIX Epoch (1970-01-01 00:00:00 UTC). Issue: 1) Where are these regions listed? The only lists of geographical areas I could find are: - the zones in zone.tab - (current) countries in iso3166.tab Theory reads: Use ISO 3166 (see the file iso3166.tab) to help decide whether something is a country. [...] or when locations change countries (e.g. Hong Kong from UK colony to China) Issue: 1) "help to decide" - decision process is undefined. 2) Hong Kong is listed in iso3166.tab as HK.For the purpose of zone creation is it 2a) a country,then it is confusing to say it changed country, or 2b) is it part of China (CN) then it is confusing to mark it with HK 3) In case HK and one of the zones in CN observed the same time since 1970, and that HK is not a country, would there be a merge and would one zone be dropped from zone.tab and any corresponding pre-1970 history be deleted? Theory reads: Within a file name component, use only ASCII letters Issue: 1) It is not defined what the term "file name component" refers to in the context of zone naming. Theory reads: E.g. prefer `Paris' to `France', since France has had multiple time zones. Issue: 1) It does not matter what zones France "had" in the past. 2) It may matter that Europe/Paris might be split in the future. 3) There is no proof in zone.tab that France had multiple time zones at all, there is only one line marked "FR": FR+4852+00220Europe/Paris -- Tobias Conradi Rheinsberger Str. 18 10115 Berlin, Germany http://tobiasconradi.com
On 09/11/11 04:18, Olson, Arthur David (NIH/NCI) [E] wrote:
From: Mr tobias conradi [tobias_conradi@yahoo.de] Sent: Sunday, September 11, 2011 12:19 AM
1) Where are these regions listed? The only lists of geographical areas I could find are: - the zones in zone.tab - (current) countries in iso3166.tab
Yes, that's where the regions are listed in the database. There are efforts to list the boundaries of the regions, but these are outside the scope of the current database.
Theory reads: Use ISO 3166 (see the file iso3166.tab) to help decide whether something is a country. [...] or when locations change countries (e.g. Hong Kong from UK colony to China)
Issue: 1) "help to decide" - decision process is undefined.
It is not a deterministic algorithm; there are border cases that require human judgment.
2) Hong Kong is listed in iso3166.tab as HK. For the purpose of zone creation is it 2a) a country,
Yes, that's right.
Theory reads: Within a file name component, use only ASCII letters
Issue: 1) It is not defined what the term "file name component" refers to in the context of zone naming.
A maximal name component that does not contain '/'.
Theory reads: E.g. prefer `Paris' to `France', since France has had multiple time zones.
Issue: 1) It does not matter what zones France "had" in the past. 2) It may matter that Europe/Paris might be split in the future.
Yes, (2) is the point.
3) There is no proof in zone.tab that France had multiple time zones at all
It did, during both World War I and World War II. Shanks & Pottenger partition France into 28 distinct regions. But this sort of detail isn't needed in the table.
Theory reads: To represent this data, the world is partitioned into regions whose clocks all agree
1) Where are these regions listed? The only lists of geographical areas I could find are: - the zones in zone.tab - (current) countries in iso3166.tab
Yes, that's where the regions are listed in the database. There are efforts to list the boundaries of the regions, but these are outside the scope of the current database. Since iso3166.tab contains geographic areas whose clocks did not all agree, the place seems to be zone.tab.
Zone.tab speaks of "zone" not of "region". Ammendement to Theory: To represent this data, the world is partitioned into regions <NEW>, named time zone,</NEW> whose clocks all agree OR To represent this data, the world is partitioned into regions <NEW>, named time zone and listed in zone.tab,</NEW> whose clocks all agree
Theory reads:
Use ISO 3166 (see the file iso3166.tab) to help decide whether something is a country. [...] or when locations change countries (e.g. Hong Kong from UK colony to China) Issue: 1) "help to decide" - decision process is undefined.
It is not a deterministic algorithm; there are border cases that require human judgment.
Any example?
2) Hong Kong is listed in iso3166.tab as HK. For the purpose of zone creation is it 2a) a country,
Yes, that's right.
Ammendement to Theory: OLD or when locations change countries (e.g. Hong Kong from UK colony to China) NEW or when locations change countries (e.g. Hong Kong from UK to China) Reason: "UK colony" is not a country. Use same level as for "China".
Theory reads:
E.g. prefer `Paris' to `France', since France has had multiple time zones. Issue: 1) It does not matter what zones France "had" in the past. 2) It may matter that Europe/Paris might be split in the future.
Yes, (2) is the point.
3) There is no proof in zone.tab that France had multiple time zones at all
It did, during both World War I and World War II. Shanks & Pottenger partition France into 28 distinct regions. But this sort of detail isn't needed in the table.
Thanks. Since WWI and WWII took place before 1970, that does not matter for tzdata purpose. Ammendement to Theory: OLD E.g. prefer `Paris' to `France', since France has had multiple time zones. NEW E.g. prefer `Paris' to `France', since France may be split into multiple time zones in the future. Reason: Refer to a fact that matters (2) instead of one that does not matter (3). -- Tobias Conradi Rheinsberger Str. 18 10115 Berlin, Germany
participants (3)
-
Mr tobias conradi -
Olson, Arthur David (NIH/NCI) [E] -
Paul Eggert