I've received the February 1997 edition of the Standard Schedules Information Manual, published by the International Air Transport Association (IATA). This manual has an appendix of international time zone information, which is intended for use by member airlines in generating and interpreting their service schedules. I compared it line by line with the previous (September 1996) edition. This is a report on the differences. In the following table, "edit" is the date of the edition. "zon" is a code for the time zone, consisting of the ISO 3166 country code followed by a digit, where necessary, to distinguish between different zones in the same country. "stdoff" and "dstoff" are the offsets from GMT to standard time and daylight saving time in the zone, respectively. The rest of each line consists of two pairs of consecutive dates for the beginning and end of daylight saving time in the zone, expressed in UTC (without the century, to save space). For time zones that don't observe DST, the last five columns are blank. For countries in the Southern Hemisphere, the four dates represent the same years. For countries in the Northern Hemisphere, the last two dates reported in the 9/96 manual correspond to the first two dates in the 2/97 manual. To help compare the two rows, I've flagged the items in the second row that differ from the first row, or from what you would expect to find in the continuation of the first row. I hope that line wrapping won't make this table too hard to read. edit zon stdoff dstoff 1st dst start 1st dst end 2nd dst start 2nd dst end ---- --- ------ ------ --------------- --------------- --------------- --------------- 9/96 BR1 -3:00 -2:00 Oct 13 96 03:00 Feb 16 97 02:00 Oct 12 97 03:00 Feb 15 98 02:00 2/97 BR1 -3:00 -2:00 Oct 6 96 03:00 Feb 16 97 02:00 Oct 5 97 03:00 Feb 15 98 02:00 ^^ ^^ 9/96 BR2 -4:00 -3:00 Oct 13 96 04:00 Feb 16 97 03:00 Oct 12 97 04:00 Feb 15 98 03:00 2/97 BR2 -4:00 -3:00 Oct 6 96 04:00 Feb 16 97 03:00 Oct 5 97 04:00 Feb 15 98 03:00 ^^ ^^ 9/96 CL1 -4:00 -3:00 Oct 13 96 04:00 Mar 9 97 03:00 Oct 12 97 04:00 Mar 8 98 03:00 2/97 CL1 -4:00 -3:00 Oct 13 96 04:00 Mar 9 97 03:00 Oct 12 97 04:00 Mar 15 98 03:00 ^^ 9/96 CL2 -6:00 -5:00 Oct 13 96 06:00 Mar 9 97 05:00 Oct 12 97 06:00 Mar 8 98 05:00 2/97 CL2 -6:00 -5:00 Oct 13 96 06:00 Mar 9 97 05:00 Oct 12 97 06:00 Mar 15 98 05:00 ^^ 9/96 CU -5:00 -4:00 Apr 7 96 05:00 Oct 6 96 05:00 Apr 6 97 05:00 Oct 5 97 05:00 2/97 CU -5:00 -4:00 Apr 6 97 05:00 Oct 12 97 05:00 Apr 5 98 05:00 Oct 11 98 05:00 ^^ ^^ 9/96 GE 4:00 5:00 Mar 30 96 20:00 Oct 26 96 19:00 Mar 29 97 20:00 Oct 25 97 19:00 2/97 GE 5:00 9/96 IL 2:00 3:00 Mar 14 96 22:00 Sep 15 96 21:00 Mar 13 97 22:00 Sep 14 97 21:00 2/97 IL 2:00 3:00 Mar 13 97 22:00 Oct 18 97 21:00 Mar 19 98 22:00 Oct 17 98 21:00 ^^^^^^ ^^^^^^ 9/96 KI2 -10:00 2/97 KI2 14:00 ^^^^^^ 9/96 KI3 -11:00 2/97 KI3 13:00 ^^^^^^ 9/96 KG 5:00 6:00 Apr 13 96 19:00 Sep 28 96 18:00 Apr 12 97 19:00 Sep 27 97 18:00 2/97 KG 5:00 6:00 Apr 12 97 19:00 Sep 27 97 18:00 Apr 11 98 19:00 Oct 26 98 18:00 ^^^^^^ 9/96 LY 1:00 2:00 Mar 30 96 01:00 Sep 30 96 01:00 Mar 30 97 01:00 Sep 30 97 01:00 2/97 LY 1:00 2:00 Mar 27 97 01:00 Oct 2 97 01:00 Mar 26 98 01:00 Oct 1 98 01:00 ^^^^^^ ^^^^^^ ^^^^^^ ^^^^^^ 9/96 MN 8:00 9:00 Mar 30 96 16:00 Oct 24 96 15:00 Mar 29 97 16:00 Oct 23 97 15:00 2/97 MN 8:00 9:00 Mar 29 97 16:00 Sep 27 97 15:00 Mar 28 98 16:00 Sep 26 98 15:00 ^^^^^^ ^^^^^^ 9/96 NC 11:00 2/97 NC 11:00 12:00 Nov 30 96 15:00 Mar 1 97 15:00 Nov 29 97 15:00 Mar 7 98 15:00 9/96 PY -4:00 -3:00 Oct 1 96 04:00 Mar 1 97 03:00 Oct 1 97 04:00 Mar 1 98 03:00 2/97 PY -4:00 -3:00 Oct 6 96 04:00 Mar 1 97 03:00 Oct 1 97 04:00 Mar 1 98 03:00 ^^ 9/96 LK 6:30 2/97 LK 6:00 9/96 UA1 2:00 3:00 Mar 31 96 01:00 Oct 27 96 01:00 Mar 30 97 01:00 Oct 26 97 01:00 2/97 UA1 2:00 3:00 Mar 30 97 03:00 Oct 26 97 04:00 Mar 29 98 03:00 Oct 25 98 04:00 ^^^^^ ^^^^^ ^^^^^ ^^^^^ Descriptions of the changes shown in these manuals are as follows. I've had to make assumptions in order to state the rules. All I can say for sure is that they're consistent with the data. The rules are expressed in local time, with midnight counting as the following day. Brazil has changed the end of DST from the second to the first Sunday in October, starting in 1996. (Actually, the rule implied by previous manuals was Oct Sun>10 or Sun>11, from 1993 to 1997.) Chile has changed the end of DST from the second to the third Sunday in March, starting in 1998. Cuba has changed the end of DST from Oct Sun>4 to Sun>5 (or more), starting in 1997. Georgia has gone on year-round standard time using its former DST offset, probably by failing to change the clocks back in October, 1996. Israel has changed the end of DST from the third Monday in September to the third Sunday in October, starting in 1997. The parts of Kiribati that were east of the International Date Line have set their calendars ahead one day, so that at most times Kiribati will have a uniform date. There's no evidence in the manuals as to when this took effect, but it's also been reported in the news. Kyrgyzstan has changed the end of DST from the last Sunday in September to the last Tuesday in October, starting in 1998. This may be a typo in the manual. If it said Sep 26 instead of Oct 26, that would be consistent with the old rule. Besides, Tuesday is a rather unusual day for DST to end on. Libya has changed the start and end of DST from fixed dates to the last Thursday in March and the first Thursday in October, respectively, starting in 1997. Mongolia has changed the end of DST from the fourth Friday in October to the fourth Sunday in September, starting in 1997. (The October rule only lasted for one year, 1996.) New Caledonia has adopted DST, beginning December 1, 1996 (local time). Paraguay started DST on October 6 instead of October 1 in 1996 only. Sri Lanka has set its clocks back a half hour. There's no evidence in the manuals as to when this took effect. Ukraine, except for Symferopol, has changed the time at which DST begins to 05:00, and the end time to 07:00 (local time). This is rather unusual, and I suspect an error in the manual. There were three other entries in the manual where I am convinced that IATA made an error. I fixed the errors and didn't list them as changes above. Just for the record, here they are. The manual says that: (1) Gibraltar has moved its winter time an hour earlier, but not its summer time, so that in the winter it matches the United Kingdom and in the Summer it matches mainland Spain, with a two-hour change each time. (2) DST ends one day earlier in Kaliningrad Oblast than in the rest of Russia, although it starts on the same day and hour. This is an error that has occurred in previous editions. When the times are expressed in UTC, Kaliningrad is the only one that changes at or after midnight, so the error is easily explained. Someone copied the date for all the Russian entries without noticing that it should have been the following date in Kaliningrad. (3) In Symferopol, the offset from UTC to standard time is three hours. The manual reports that the offset for DST is four hours in 1997, but three hours in 1998. If that were really true, there would be no point in their reporting the DST start and end dates for 1998, because DST would be the same as standard. Note to Paul Eggert: I'll send you new files with these changes incorporated as soon as I have time (probably not in the next week). -- Gwillim Law Gwillim Law gwil@mindspring.com