Hi. I have investigated time in Falkland Islands and Moldova, and have found discrepancies with the current tz database. Falkland Islands is represented like this in the tz database (2000h): # Falklands # Rule NAME FROM TO TYPE IN ON AT SAVE LETTER/S Rule Falk 1937 1938 - Sep lastSun 0:00 1:00 S Rule Falk 1938 1942 - Mar Sun>=19 0:00 0 - Rule Falk 1939 only - Oct 1 0:00 1:00 S Rule Falk 1940 1942 - Sep lastSun 0:00 1:00 S Rule Falk 1943 only - Jan 1 0:00 0 - Rule Falk 1983 only - Sep lastSun 0:00 1:00 S Rule Falk 1984 1985 - Apr lastSun 0:00 0 - Rule Falk 1984 only - Sep 16 0:00 1:00 S Rule Falk 1985 1995 - Sep Sun>=9 0:00 1:00 S Rule Falk 1986 max - Apr Sun>=16 0:00 0 - Rule Falk 1996 max - Sep Sun>=8 0:00 1:00 S # Zone NAME GMTOFF RULES FORMAT [UNTIL] Zone Atlantic/Stanley -3:51:24 - LMT 1890 -3:51:24 - SMT 1912 Mar 12 # Stanley Mean Time -4:00 Falk FK%sT 1983 May # Falkland Is Time -3:00 Falk FK%sT 1985 Sep 15 -4:00 Falk FK%sT The rules "Sun>=9" that was later substituted with "Sun>=8", and "Sun>=16", arose my curiosity. The rule "Sun>=9" means the Sunday after the Second Saturday which is pretty awkward to say the least (but possible), and not supported by Windows (which support First, Second, Third, Fourth and Last weekday of a given month). It was replaced in 1996 with the rule "Sun>=8" which means the Second Sunday which *is* supported by Windows. It just so happens that 1996 was the first year in several years back where it would make a difference which rule to use. With the first rule we would get the 15.th. September, while the second rule we would get the 8.th. September. Similarly looking at the rule "Sun>=16" it means the Sunday after the third Saturday, which is an awkward rule and not supported by Windows. I got the suspicion that it should have been "Sun>=15" instead, which means the third Sunday. It just happens that this gives a difference this year, because using the first rule sets the date to Sunday April 22, while the second rule gives Sunday April 15. I decided to investigate the matter. I found a page in Internet that is apparently the official Falkland Island governments web page, http://www.falklands.gov.fk/ In the page http://www.falklands.gov.fk/feedback.htm I entered my question about the planned DST for Falkland Islands this year. The account manager@figo.u-net.com responded me with the following: ===================== Begin Quote ========================== From: General Manager[SMTP:manager@figo.u-net.com] Sent: Lunes 22 de Enero de 2001 7:10 To: 'jnorgard@Prodigy.Net.mx' Subject: Time Zone INformation Thank you for contacting Falkland Islands Government. To help you with your enquiry, we can confirm the clocks revert back to Local Mean Time at 2 am on Sunday 15 April 2001and advance one hour to summer time at 2 am on Sunday 2 September. It is anticipated that the clocks will revert back at 2 am on Sunday 21 April 2002 and advance to summer time at 2 am on Sunday 1 September. I hope this helps you. Do contact us again if we can be of further assistance. Regards Falkland Islands Government Office, London ===================== End Quote ========================== This valuable peace of information gives us at least three details that doesn't correspond with the tz database: (1) Start and end time are 2 am in the morning (I assume local time) (2) My theory about "Sun>=15" instead of "Sun>=16" is apparently correct (3) The rule "Sun>=8" should actually be "Sun>=1" instead. I have sent a question about how far back this rule has been used (which will apparently be used in the years 2001 and 2002) but haven't received any answer yet. I will return to you with that information if I get an answer. Concerning Moldova: A person contacted me to inquire about if the extra time zone GMT+3 in Moldova for Tiraspol really was correct, since I was the only one he had seen using that. Any other sources he had seen was only using Chisinau GMT+2. The current information about Moldova in the tz database (2000h): # Moldova # Zone NAME GMTOFF RULES FORMAT [UNTIL] Zone Europe/Chisinau 1:55:20 - LMT 1880 1:55 - CMT 1918 Feb 15 # Chisinau MT 1:44:24 - BMT 1931 Jul 24 # Bucharest MT 2:00 Romania EE%sT 1940 Aug 15 2:00 1:00 EEST 1941 Jul 17 1:00 C-Eur CE%sT 1944 Aug 24 3:00 Russia MSK/MSD 1990 3:00 - MSK 1990 May 6 2:00 - EET 1991 2:00 Russia EE%sT 1992 2:00 E-Eur EE%sT 1997 # See Romania commentary for the guessed 1997 transition to EU rules. 2:00 EU EE%sT Zone Europe/Tiraspol 1:58:32 - LMT 1880 1:55 - CMT 1918 Feb 15 # Chisinau MT 1:44:24 - BMT 1931 Jul 24 # Bucharest MT 2:00 Romania EE%sT 1940 Aug 15 2:00 1:00 EEST 1941 Jul 17 1:00 C-Eur CE%sT 1944 Aug 24 3:00 Russia MSK/MSD 1991 Mar 31 2:00 2:00 Russia EE%sT 1992 Jan 19 2:00 3:00 Russia MSK/MSD I decided to check it out with some searches in Internet, but didn't find anything remotely authoritative, but did find some english-speaking persons with e-mail that were apparently living in Tiraspol. I sent mails to 4 persons and got 2 replies: ===================== Begin Quote ========================== From: Moldavizolit[SMTP:moldavizolit@tirastel.md] Reply To: Moldavizolit Sent: Miércoles 24 de Enero de 2001 1:03 To: Jesper Nørgaard Subject: Re: Time in Tiraspol
Hello Jesper Nørgaard !
Tiraspol is on the same time as Chisinau (Kishinev, GMT+2) Regards,
Lubov ===================== End Quote ==========================
... and this one ===================== Begin Quote ========================== From: ??? ???????????[SMTP:mk@tirastel.md] Sent: Miércoles 24 de Enero de 2001 4:28 To: Jesper Nørgaard Subject: Re: Time in Tiraspol Hello, Yes, time in Tiraspol is on the same time as Chisinau (Kishinev, GMT+2) best ragards Dmitry Goncharenco ----- Original Message ----- From: Jesper Nørgaard To: 'mk@mk.tiraspol.net' Sent: Wednesday, January 24, 2001 5:13 AM Subject: Time in Tiraspol ===================== End
Date: Thu, 25 Jan 2001 00:10:15 -0600 From: <jnorgard@Prodigy.Net.mx>
I decided to check it out with some searches in Internet, but didn't find anything remotely authoritative, but did find some english-speaking persons with e-mail that were apparently living in Tiraspol. I sent mails to 4 persons and got 2 replies:
Thanks for checking into this. On further investigation, the existing Europe/Tiraspol entry (which is taken from Shanks) is most likely incorrect. I'll correct this in my next proposed patch.
# Falklands... Rule Falk 1985 1995 - Sep Sun>=9 0:00 1:00 S Rule Falk 1986 max - Apr Sun>=16 0:00 0 - Rule Falk 1996 max - Sep Sun>=8 0:00 1:00 S
The rules "Sun>=9" that was later substituted with "Sun>=8", and "Sun>=16", arose my curiosity. The rule "Sun>=9" means the Sunday after the Second Saturday which is pretty awkward to say the least (but possible)
Yes. That rule was inferred from the Shanks data before 1990 and the IATA data through February 2000. For example, both IATA and Shanks give 1991-09-15, and Shanks gives 1985-09-15. Unfortunately this data lists only dates and times, so I don't know the actual rules. The Sun>=9 rule resembles the rule that we know is used in Chile, so it's not entirely implausible, but I admit that there's a lot of guesswork here.
It just so happens that 1996 was the first year in several years back where it would make a difference which rule to use. With the first rule we would get the 15.th. September, while the second rule we would get the 8.th. September.
Just to make things interestings, Shanks says that the 1996 transition occurred on 15 September, but I was going with the IATA data which gives 8 September. This is the only date between 1990 and 2000 (inclusive) for which the two sources disagree.
Similarly looking at the rule "Sun>=16" it means the Sunday after the third Saturday, which is an awkward rule and not supported by Windows.
Yes. However, Shanks gives 1990-04-22, which is a counterexample to Sun>=15. (Perhaps they didn't use Windows much in the Falklands back then. :-) Our IATA data do not go back that far. Thanks for finding out the official Falklands rule for this year. I'll update the tables and (for lack of better info) assume the switch occurred at the start of this year.
participants (2)
-
Jesper Nørgaard -
Paul Eggert