[PROPOSED] Namibia was on DST until 2017-10-24
While we're at it, use negative DST offsets in vanguard format, since Namibia (like Ireland) used negative DST in winter. * NEWS, africa (Namibia, Africa/Windhoek), ziguard.awk: Implement this. --- NEWS | 10 ++++++++++ africa | 19 +++++++++++++++++-- ziguard.awk | 28 +++++++++++++++++++++++++--- 3 files changed, 52 insertions(+), 5 deletions(-) diff --git a/NEWS b/NEWS index 0297fd8..3bbb829 100644 --- a/NEWS +++ b/NEWS @@ -1,5 +1,15 @@ News for the tz database +Unreleased, experimental changes + + Changes to past time stamps + + For Namibia's transition from +01/+02 to plain +02, change the + transition date from 2017-09-03 at 03:00 to 2017-10-24 at 00:00. + This does not affect UTC offsets, only whether a DST regime was + in place. + + Release 2018d - 2018-03-22 07:05:46 -0700 Briefly: diff --git a/africa b/africa index 2f7217a..e564906 100644 --- a/africa +++ b/africa @@ -992,9 +992,20 @@ Link Africa/Maputo Africa/Lusaka # Zambia # the same time they would normally start DST, the first Sunday in September: # https://www.timeanddate.com/news/time/namibia-new-time-zone.html +# From Paul Eggert (2017-04-05): +# The official date of the time zone change was 2017-10-24. See: +# http://www.lac.org.na/laws/annoSTAT/Namibian%20Time%20Act%209%20of%202017.pd... + # RULE NAME FROM TO TYPE IN ON AT SAVE LETTER/S +# The following has standard time in summer and negative daylight +# saving time in winter. It is for when negative SAVE values are used. +# For more about this, see the comments in 'europe' for Rule Eire. +#Rule Namibia 1994 only - Mar 21 0:00 -1:00 +01 +#Rule Namibia 1994 2017 - Sep Sun>=1 2:00 0 CAT +#Rule Namibia 1995 2017 - Apr Sun>=1 2:00 -1:00 +01 +# The following is for when SAVE values are always nonnegative. Rule Namibia 1994 only - Mar 21 0:00 0 - -Rule Namibia 1994 2016 - Sep Sun>=1 2:00 1:00 S +Rule Namibia 1994 2017 - Sep Sun>=1 2:00 1:00 S Rule Namibia 1995 2017 - Apr Sun>=1 2:00 0 - # Zone NAME GMTOFF RULES FORMAT [UNTIL] Zone Africa/Windhoek 1:08:24 - LMT 1892 Feb 8 @@ -1003,7 +1014,11 @@ Zone Africa/Windhoek 1:08:24 - LMT 1892 Feb 8 2:00 1:00 SAST 1943 Mar 21 2:00 2:00 - SAST 1990 Mar 21 # independence 2:00 - CAT 1994 Mar 21 0:00 - 1:00 Namibia WA%sT 2017 Sep 3 2:00 +# The next line is for when negative SAVE values are used. +# 2:00 Namibia %s 2017 Oct 24 +# The next line is for when SAVE values are always nonnegative. + 1:00 Namibia WA%sT 2017 Oct 24 +# End of daylight saving. 2:00 - CAT # Niger diff --git a/ziguard.awk b/ziguard.awk index 6da3691..710b630 100644 --- a/ziguard.awk +++ b/ziguard.awk @@ -26,6 +26,7 @@ BEGIN { outfile != "main.zi" { in_comment = /^#/ + uncomment = comment_out = 0 # If this line should differ due to Ireland using negative SAVE values, # uncomment the desired version and comment out the undesired one. @@ -37,11 +38,32 @@ outfile != "main.zi" { if ((Rule_Eire \ || (Zone_Dublin_post_1968 && $(in_comment + 3) == "IST/GMT")) \ == vanguard) { - sub(/^#/, "") - } else if (/^[^#]/) { - sub(/^/, "#") + uncomment = in_comment + } else { + comment_out = !in_comment } } + + # If this line should differ due to Namibia using negative SAVE values, + # uncomment the desired version and comment out the undesired one. + Rule_Namibia = /^#?Rule[\t ]+Namibia[\t ]/ + Zone_using_Namibia_rule \ + = (zone == "Africa/Windhoek" && /^#?[\t ]+[12]:00[\t ]/ \ + && $(in_comment + 2) == "Namibia") + if (Rule_Namibia || Zone_using_Namibia_rule) { + if (in_comment == vanguard) { + uncomment = in_comment + } else { + comment_out = !in_comment + } + } + + if (uncomment) { + sub(/^#/, "") + } + if (comment_out) { + sub(/^/, "#") + } } # If a Link line is followed by a Zone line for the same data, comment -- 2.14.3
On 2018-04-06 00:10, Paul Eggert wrote:
While we're at it, use negative DST offsets in vanguard format, since Namibia (like Ireland) used negative DST in winter.
I must overlook something here -- but I do not see why you use negative SAVE values before 2017-10-24. From 1994-03 until 2017-10-24, the Namibian Time Act 1994 was in effect, which specifies: " 1(1) The standard time of Namibia shall - (a) during the summer period, be two hours in advance of Greenwich mean time; and (b) during the winter period, be one hour in advance of Greenwich mean time. " See [www.lac.org.na/laws/1994/811.pdf]. So both summer and winter time are called "standard" (which differs from the use in Ireland), and the proposed tzdb dst bit just indicates the "winter period". Michael Deckers.
[Changing the Subject: line from [PROPOSED] Namibia was on DST until 2017-10-24"".] Michael H Deckers via tz wrote:
See [www.lac.org.na/laws/1994/811.pdf].
So both summer and winter time are called "standard" (which differs from the use in Ireland),
Thanks, good catch. So: not only was the proposed patch incorrect for Namibia, tzdb has been wrong since tzdata1997i, since it said that Namibian summer time was daylight saving time, and this wasn't correct either. zic does not have a convenient notation for Namibia's timekeeping practices in 1994-2017, when standard time was changed twice per year in a regular pattern. Although tzdb can explicitly list each transition, that's awkward and error-prone; it's better to use Rule lines. Attached is a proposed patch to do that, by slightly extending the syntax of Rule lines so that an entry in their SAVE columns can have an 's' suffix if you want the corresponding rule to specify standard time. For completeness, the new 'd' suffix specifies daylight saving time. Since this is an extension to zic input format, it's a vanguard feature, which means that the patch causes the new feature to be used in vanguard.zi but not in the traditional 'africa' file. We will wait a while before migrating it into 'africa'. As usual, people can test their tzdata parsers on bleeding-edge-format data by running the parsers on 'vanguard.zi'.
I think there's a good chance Namibia is using the word "standard" in a different way than is meant by the distinction between "standard" and "daylight saving" times in this instance. They may mean "standard" in the sense of "the time used in Namibia is standardized to X." While I don't disagree with the proposed expansion of the zi format to *allow* for this, for the specific case of Namibia, it would probably be a good idea to try and find evidence of whether the people of Namibia consider their "standard" time as shifting back and forth, or whether they consider one to be the "standard" offset. On 04/07/2018 09:21 PM, Paul Eggert wrote:
[Changing the Subject: line from [PROPOSED] Namibia was on DST until 2017-10-24"".]
Michael H Deckers via tz wrote:
See [www.lac.org.na/laws/1994/811.pdf].
So both summer and winter time are called "standard" (which differs from the use in Ireland),
Thanks, good catch. So: not only was the proposed patch incorrect for Namibia, tzdb has been wrong since tzdata1997i, since it said that Namibian summer time was daylight saving time, and this wasn't correct either.
zic does not have a convenient notation for Namibia's timekeeping practices in 1994-2017, when standard time was changed twice per year in a regular pattern. Although tzdb can explicitly list each transition, that's awkward and error-prone; it's better to use Rule lines. Attached is a proposed patch to do that, by slightly extending the syntax of Rule lines so that an entry in their SAVE columns can have an 's' suffix if you want the corresponding rule to specify standard time. For completeness, the new 'd' suffix specifies daylight saving time.
Since this is an extension to zic input format, it's a vanguard feature, which means that the patch causes the new feature to be used in vanguard.zi but not in the traditional 'africa' file. We will wait a while before migrating it into 'africa'. As usual, people can test their tzdata parsers on bleeding-edge-format data by running the parsers on 'vanguard.zi'.
The Wikipedia article (https://en.wikipedia.org/wiki/Time_in_Namibia) leads me to this story suggesting that Namibians consider themselves as switching to "Winter Time" (as is the case in Ireland): https://www.newera.com.na/2016/03/24/grn-evaluates-winter-time-change/ This one similarly suggests that they think of themselves as switching to "Winter time": https://www.namibian.com.na/161621/archive-read/Time-change-divides-lawmaker... Regardless of how the law is worded, I think people considered summer to be the standard time rather than always being on 'standard time' and the offset switching twice per year. On 04/07/2018 10:19 PM, Paul G wrote:
I think there's a good chance Namibia is using the word "standard" in a different way than is meant by the distinction between "standard" and "daylight saving" times in this instance. They may mean "standard" in the sense of "the time used in Namibia is standardized to X."
While I don't disagree with the proposed expansion of the zi format to *allow* for this, for the specific case of Namibia, it would probably be a good idea to try and find evidence of whether the people of Namibia consider their "standard" time as shifting back and forth, or whether they consider one to be the "standard" offset.
On 04/07/2018 09:21 PM, Paul Eggert wrote:
[Changing the Subject: line from [PROPOSED] Namibia was on DST until 2017-10-24"".]
Michael H Deckers via tz wrote:
See [www.lac.org.na/laws/1994/811.pdf].
So both summer and winter time are called "standard" (which differs from the use in Ireland),
Thanks, good catch. So: not only was the proposed patch incorrect for Namibia, tzdb has been wrong since tzdata1997i, since it said that Namibian summer time was daylight saving time, and this wasn't correct either.
zic does not have a convenient notation for Namibia's timekeeping practices in 1994-2017, when standard time was changed twice per year in a regular pattern. Although tzdb can explicitly list each transition, that's awkward and error-prone; it's better to use Rule lines. Attached is a proposed patch to do that, by slightly extending the syntax of Rule lines so that an entry in their SAVE columns can have an 's' suffix if you want the corresponding rule to specify standard time. For completeness, the new 'd' suffix specifies daylight saving time.
Since this is an extension to zic input format, it's a vanguard feature, which means that the patch causes the new feature to be used in vanguard.zi but not in the traditional 'africa' file. We will wait a while before migrating it into 'africa'. As usual, people can test their tzdata parsers on bleeding-edge-format data by running the parsers on 'vanguard.zi'.
I agree with Paul G. The meaning of "standard" here is "standardized" or "common across the country". https://www.namibian.com.na/161621/archive-read/Time-change-divides-lawmaker... "Iivula-Ithana told the NA that Cabinet decided that Namibia should have summer time as standard time." implying that use of "summer time" trumps reference to "standard time. "The country changes to winter time in April, but the Zambezi region maintains summer time all year." https://www.namibian.com.na/index.php?page=archive-read&id=2875 "NAMIBIA enters its five-month-long winter time period at 02h00 on Sunday when the official time will be changed an hour back, to 01h00. Official summer time returns at 02h00 on the first Sunday in September [September 5 this year], when the country's standard time shifts an hour forward again." https://www.namibian.com.na/164449/archive-read/Poor-attendance-at-public-he... " the new law seeks to repeal the Namibian Time Act of 1994 to set the standard time to two hours in advance of Greenwich Mean Time" Note it is "the standard time" not "standard time". https://www.youtube.com/watch?v=FV9jSBqEepg Video, no mention of "standard time" https://www.youtube.com/watch?v=HF-dx3Km9GQ Video, refers to winter, summer and "national time", not "standard time" Together with other links, it clearly suggests that "winter time" and "summer time" are in common parlance. That the word "standard" is also used simply doesn't have the meaning TZDB uses for that word. The TZDB is_dst flag should express that there is a relationship between the times within the year, ie. one is for summer (and therefor the other is implicitly winter). Given this, there is no need to introduce a change to the source format. I do agree that the evidence is that Namibia viewed switching to winter time as DST (a negative SAVE), although I continue to think that is irrelevant to the purpose and goals of TZDB. The old law is here: http://www.lac.org.na/laws/1994/811.pdf The new law is here: http://www.lac.org.na/laws/annoSTAT/Namibian%20Time%20Act%209%20of%202017.pd... And just for fun, here are some references about the area around Zambezi in eastern Namibia. Maybe this is enough historical evidence for that region to have its own ID? https://www.namibian.com.na/index.php?id=52856&page=archive-read https://www.newera.com.na/2017/06/01/zambezi-residents-support-standard-time... Stephen On 8 April 2018 at 03:25, Paul G <paul@ganssle.io> wrote:
The Wikipedia article (https://en.wikipedia.org/wiki/Time_in_Namibia) leads me to this story suggesting that Namibians consider themselves as switching to "Winter Time" (as is the case in Ireland): https://www.newera.com.na/2016/03/24/grn-evaluates-winter-time-change/
This one similarly suggests that they think of themselves as switching to "Winter time": https://www.namibian.com.na/161621/archive-read/Time-change-divides-lawmaker...
Regardless of how the law is worded, I think people considered summer to be the standard time rather than always being on 'standard time' and the offset switching twice per year.
On 04/07/2018 10:19 PM, Paul G wrote:
I think there's a good chance Namibia is using the word "standard" in a different way than is meant by the distinction between "standard" and "daylight saving" times in this instance. They may mean "standard" in the sense of "the time used in Namibia is standardized to X."
While I don't disagree with the proposed expansion of the zi format to *allow* for this, for the specific case of Namibia, it would probably be a good idea to try and find evidence of whether the people of Namibia consider their "standard" time as shifting back and forth, or whether they consider one to be the "standard" offset.
On 04/07/2018 09:21 PM, Paul Eggert wrote:
[Changing the Subject: line from [PROPOSED] Namibia was on DST until 2017-10-24"".]
Michael H Deckers via tz wrote:
See [www.lac.org.na/laws/1994/811.pdf].
So both summer and winter time are called "standard" (which differs from the use in Ireland),
Thanks, good catch. So: not only was the proposed patch incorrect for Namibia, tzdb has been wrong since tzdata1997i, since it said that Namibian summer time was daylight saving time, and this wasn't correct either.
zic does not have a convenient notation for Namibia's timekeeping practices in 1994-2017, when standard time was changed twice per year in a regular pattern. Although tzdb can explicitly list each transition, that's awkward and error-prone; it's better to use Rule lines. Attached is a proposed patch to do that, by slightly extending the syntax of Rule lines so that an entry in their SAVE columns can have an 's' suffix if you want the corresponding rule to specify standard time. For completeness, the new 'd' suffix specifies daylight saving time.
Since this is an extension to zic input format, it's a vanguard feature, which means that the patch causes the new feature to be used in vanguard.zi but not in the traditional 'africa' file. We will wait a while before migrating it into 'africa'. As usual, people can test their tzdata parsers on bleeding-edge-format data by running the parsers on 'vanguard.zi'.
On 04/08/2018 02:55 AM, Stephen Colebourne wrote:
Together with other links, it clearly suggests that "winter time" and "summer time" are in common parlance. That the word "standard" is also used simply doesn't have the meaning TZDB uses for that word.
OK, but that's pretty much the same meaning that "standard time" originally had in North America. The terminology diverged only when DST is considered. In North America we now call only winter time "standard time", whereas in Namibia they call both summer and winter time "standard time".
I do agree that the evidence is that Namibia viewed switching to winter time as DST (a negative SAVE),
Thanks for looking into this. This inspired me to look for non-Wikipedia English-language sources that talk about the popular terminology for time in Namibia (as opposed to the legal terminology). They were pretty uniformly in agreement with you. Proposed patch attached.
On 2018-04-09 18:07, Paul Eggert proposed the text:
# I invented the following abbreviations; corrections are welcome! -# +02 WAST West Africa Summer Time +# +02 WAST West Africa Summer Time (no longer used)
I am confused -- does this mean that we have to expect a change in Africa/Ndjamena? Michael Deckers.
On 04/09/2018 01:01 PM, Michael H Deckers via tz wrote:
does this mean that we have to expect a change in Africa/Ndjamena?
No, WAST is like CAST and SAST, which are also marked as "(no longer used)". For example, the most recent use of CAST in a timestamp was in 1985; see Africa/Khartoum and Africa/Juba. WAST is like that, except that its only use was in Africa/Ndjamena in 1979/80 (where the UTC and DST offset data are from Shanks; I invented the abbreviation "WAST").
On 2018-04-08 01:21, Paul Eggert wrote:
Attached is a proposed patch to do that, by slightly extending the syntax of Rule lines so that an entry in their SAVE columns can have an 's' suffix if you want the corresponding rule to specify standard time. For completeness, the new 'd' suffix specifies daylight saving time.
Considering that this proposal makes the setting of the dst bit an issue that is independent from the circumstance of whether the value of a SAVE entry is zero or not, is it not desirable to completely avoid the use of negative SAVE values to express the dst bit setting, so as to make it easier for parsers of zic source files to follow the changes in the "bleeding edge" format? Also, for systematicness, shouldn't the same suffixes and defaults also be applicable to a numeric entry in the RULE column of ZONE lines? Just my 0.04 $. Michael Deckers.
Michael Deckers via tz wrote:
Considering that this proposal makes the setting of the dst bit an issue that is independent from the circumstance of whether the value of a SAVE entry is zero or not, is it not desirable to completely avoid the use of negative SAVE values to express the dst bit setting, so as to make it easier for parsers of zic source files to follow the changes in the "bleeding edge" format?
It's more the other way around, I would think. zic has supported negative SAVE values for decades and most other tzdata parsers have followed suit. So if a timekeeping practice can be specified either way, it'll be easier on the installed base if we use this long-established feature than if we use the proposed new feature. I get the point that we'd rather keep the syntax simpler and more regular, and that lessening the number of syntax features helps do that. But here I think the compatibility win outweighs the regularity loss.
for systematicness, shouldn't the same suffixes and defaults also be applicable to a numeric entry in the RULE column of ZONE lines?
Good point, thanks. Proposed further patch attached.
participants (5)
-
Michael Deckers -
Michael H Deckers
-
Paul Eggert -
Paul G -
Stephen Colebourne