RE: Quetico (and other places in NW Ontario)
Chris, The map you sent of the Quetico park overview is more than sufficient for me to make a reasonable timezone. After all this is not rocket science, I'm intending to give an overview of the timezone for visual use, not for calculating if a specific latitude/longitude is inside the timezone or not. I'm using an ESRI shapefile map, you can check it out from downloading World Time Explorer beta version from http://www.worldtimeexplorer.com/wt20beta.zip and unzip and install it. I can give you a free full license if you want, so you can generate map colors for a specific date between 1850 and today. Otherwise the shareware version will give you 30 days of playing around, and the colors should correspond until 2007, except for a possible (undecided) change in West Australia of course! And PS, I presume you have (access to) a Windows PC, otherwise it will not be of much use! Linux die-hards, sorry! I was wondering if my border between Rainy_River (on GMT-6) area and the rest of Ontario (on GMT-5) is wrong, supposedly it should divide around 90° while in reality it divides around -89.34. This is from another map source where I copied this border from, but perhaps for timezone use I should move the border. It would be nice if you could plot the points from Quetico park (100 points would be fine, perhaps even too much). I don't know if you have a possibility to output to an ESRI shapefile. If not, a collection of the pure latitude/longitudes would be fine (decimal please, not radix 60). I will consider your comments about America/Winnipeg being one hour off each year, but it seems that even tz database is correct in this case? It looks like this: # From Paul Eggert (2006-04-10): # Shanks & Pottenger say Manitoba switched at 02:00 (not 02:00s) # starting 1966. Since 02:00s is clearly correct for 1967 on, assume # it was also 02:00s in 1966. ... Rule Winn 1963 only - Sep 22 2:00 0 S Rule Winn 1966 1986 - Apr lastSun 2:00s 1:00 D Rule Winn 1966 2005 - Oct lastSun 2:00s 0 S Rule Winn 1987 2005 - Apr Sun>=1 2:00s 1:00 D # Zone NAME GMTOFF RULES FORMAT [UNTIL] Zone America/Winnipeg -6:28:36 - LMT 1887 Jul 16 -6:00 Winn C%sT 2006 -6:00 Canada C%sT Since 2:00s standard time means 3:00 local time when leaving DST, we're OK? What would be the difference between the Kenora timezone and the others? Regards, - Jesper -----Original Message----- From: Chris Walton [mailto:Chris.Walton@telus.com] Sent: Domingo, 05 de Noviembre de 2006 1:24 To: jnorgard@prodigy.net.mx; tz@elsie.nci.nih.gov; Paul Eggert Subject: Quetico (and other places in NW Ontario) Jesper, It looks like the entire MNR site is down. It was working on Saturday morning (Toronto time). Hopefully it will come up soon! The maps on this site are actually scanned copies of paper maps; the idea being that you can view an image of the map before you fork out money for a paper copy. The resolution of the scanned images is terrible so the writing on the maps is virtually impossible to read. If you want a rough idea of the shape of the park have a peek at http://www.queticofoundation.org/maps/park_map.gif Let me know if you want me to generate a set of points or if you would prefer to do it yourself; I don't mind either way. I figure that the area could be roughly defined with about 100 points but to properly follow the park boundary would take around 2000 points. And now putting Quetico aside for a minute... you need to decide which zone to use for the "other" areas in Ontario west of 90°. According the TZ database, Fort Frances and Rainy River did not use daylight saving until 1974; this is why there is a dedicated zone called America/Rainy_River. So Rainy River and Fort Frances should use America/Rainy_River without question. Presumably the rest of the region west of 90° has used CST/CDT since WWII. According to the "zonetab" file, the region is supposed to use America/Winnipeg. America/Winnipeg results in incorrect times for one hour each year from 1966 through to 2005. This is because Manitoba "used" to change the clocks at 3:00a.m. every fall. The practice stopped this year. Ontario has always changed the clocks 2:00a.m. America/Rainy_River gives correct times for the entire region from 1974 onwards. Personally I think that America/Rainy_River is the better choice for the entire region because it provides correct rules for the last 32 years. Technically there should be another zone called America/Kenora but I doubt anybody wants to bother creating it. If it were up to me I would merge zones that have been the same for at least 25 years and create entries in the "backward" file for compatibility; but I know this would be against the official rules! -chris
Jesper, Just to clarify... There are no errors in America/Winnipeg. The rules are correct for the entire province of Manitoba including the city of Winnipeg. The problem is that America/Winnipeg is not a perfect fit for the neighbouring communities in northwestern Ontario. You asked what the difference between a "Kenora" time zone and the other time zones would be. Assuming Kenora has been using DST since 1970 (which I don't have proof of) here is a summary of Winnipeg, Kenora, Rainy River, and Atikokan. Winnipeg: 1970-2005: CST/CDT with transitions at 2:00a.m. in spring and "3:00a.m." in fall. 2006-2038: CST/CDT with transitions at 2:00a.m. both spring and fall. Kenora: 1970-2038: CST/CDT with transitions at 2:00a.m. both spring and fall. Rainy River: 1970-1973: CST year round 1974-2038: CST/CDT with transitions at 2:00a.m. both spring and fall. Atikokan: 1970-2038: EST year round ############# And to answer your question about the eastern limit of America/Rainy_River: America/Rainy_River should "not" go as far east as 89.34°. But before defining boundaries for America/Rainy_River, a decision needs on what to do with Kenora, Dryden, and the other communities west of 90°. Do you wish to: a) include them in America/Winnipeg (ignoring the 2:00a.m. vs 3:00a.m. problem which existed up until 2005). b) include them in America/Rainy_River (ignoring the fact that America/Rainy_River has no DST for 1970-1973 and the fact that Kenora is bigger than Rainy_River). c) ask Paul to create a new time zone called America/Kenora Ultimately America/Rainy_River could be defined as any of the following: a) the municipal boundaries of the towns of Rainy River and Fort Frances. b) the district of Rainy River minus Atikokan and Quetico. c) Ontario west of 90° minus Atikokan, Quetico, Upsala, & Shebandowan. This a can of worms; officially another time zone is required but I was rather hoping we would not need to go there. Canada already has 26 zones! Is there any way to flag some of the old historical zones such that they don't appear when running "tzselect"? Paul E., If you are reading, do you have any opinions on whether or not a new zone should be added? One more thing to note: The zone.tab file currently indicates that America/Rainy_River should be used for Rainy River and Fort Frances. The town of Fort Frances has a population of 8315. The town of Rainy River has a population of 981. How come it was not called America/Fort_Frances? ############## By they way... I tried to run worldtimeexplorer. I get a message "comctl32.ocx could not be found in C:\WINDOWS\system32". Is there a quick fix for this? -chris -----Original Message----- From: Jesper Norgaard Welen [mailto:jnorgard@prodigy.net.mx] Sent: November 5, 2006 3:20 PM To: Chris Walton; TZ-list Subject: RE: Quetico (and other places in NW Ontario) Chris, The map you sent of the Quetico park overview is more than sufficient for me to make a reasonable timezone. After all this is not rocket science, I'm intending to give an overview of the timezone for visual use, not for calculating if a specific latitude/longitude is inside the timezone or not. I'm using an ESRI shapefile map, you can check it out from downloading World Time Explorer beta version from http://www.worldtimeexplorer.com/wt20beta.zip and unzip and install it. I can give you a free full license if you want, so you can generate map colors for a specific date between 1850 and today. Otherwise the shareware version will give you 30 days of playing around, and the colors should correspond until 2007, except for a possible (undecided) change in West Australia of course! And PS, I presume you have (access to) a Windows PC, otherwise it will not be of much use! Linux die-hards, sorry! I was wondering if my border between Rainy_River (on GMT-6) area and the rest of Ontario (on GMT-5) is wrong, supposedly it should divide around 90° while in reality it divides around -89.34. This is from another map source where I copied this border from, but perhaps for timezone use I should move the border. It would be nice if you could plot the points from Quetico park (100 points would be fine, perhaps even too much). I don't know if you have a possibility to output to an ESRI shapefile. If not, a collection of the pure latitude/longitudes would be fine (decimal please, not radix 60). I will consider your comments about America/Winnipeg being one hour off each year, but it seems that even tz database is correct in this case? It looks like this: # From Paul Eggert (2006-04-10): # Shanks & Pottenger say Manitoba switched at 02:00 (not 02:00s) # starting 1966. Since 02:00s is clearly correct for 1967 on, assume # it was also 02:00s in 1966. ... Rule Winn 1963 only - Sep 22 2:00 0 S Rule Winn 1966 1986 - Apr lastSun 2:00s 1:00 D Rule Winn 1966 2005 - Oct lastSun 2:00s 0 S Rule Winn 1987 2005 - Apr Sun>=1 2:00s 1:00 D # Zone NAME GMTOFF RULES FORMAT [UNTIL] Zone America/Winnipeg -6:28:36 - LMT 1887 Jul 16 -6:00 Winn C%sT 2006 -6:00 Canada C%sT Since 2:00s standard time means 3:00 local time when leaving DST, we're OK? What would be the difference between the Kenora timezone and the others? Regards, - Jesper -----Original Message----- From: Chris Walton [mailto:Chris.Walton@telus.com] Sent: Domingo, 05 de Noviembre de 2006 1:24 To: jnorgard@prodigy.net.mx; tz@elsie.nci.nih.gov; Paul Eggert Subject: Quetico (and other places in NW Ontario) Jesper, It looks like the entire MNR site is down. It was working on Saturday morning (Toronto time). Hopefully it will come up soon! The maps on this site are actually scanned copies of paper maps; the idea being that you can view an image of the map before you fork out money for a paper copy. The resolution of the scanned images is terrible so the writing on the maps is virtually impossible to read. If you want a rough idea of the shape of the park have a peek at http://www.queticofoundation.org/maps/park_map.gif Let me know if you want me to generate a set of points or if you would prefer to do it yourself; I don't mind either way. I figure that the area could be roughly defined with about 100 points but to properly follow the park boundary would take around 2000 points. And now putting Quetico aside for a minute... you need to decide which zone to use for the "other" areas in Ontario west of 90°. According the TZ database, Fort Frances and Rainy River did not use daylight saving until 1974; this is why there is a dedicated zone called America/Rainy_River. So Rainy River and Fort Frances should use America/Rainy_River without question. Presumably the rest of the region west of 90° has used CST/CDT since WWII. According to the "zonetab" file, the region is supposed to use America/Winnipeg. America/Winnipeg results in incorrect times for one hour each year from 1966 through to 2005. This is because Manitoba "used" to change the clocks at 3:00a.m. every fall. The practice stopped this year. Ontario has always changed the clocks 2:00a.m. America/Rainy_River gives correct times for the entire region from 1974 onwards. Personally I think that America/Rainy_River is the better choice for the entire region because it provides correct rules for the last 32 years. Technically there should be another zone called America/Kenora but I doubt anybody wants to bother creating it. If it were up to me I would merge zones that have been the same for at least 25 years and create entries in the "backward" file for compatibility; but I know this would be against the official rules! -chris
"Chris Walton" <Chris.Walton@telus.com> writes:
You asked what the difference between a "Kenora" time zone and the other time zones would be. Assuming Kenora has been using DST since 1970 (which I don't have proof of) ...
Is there any way to flag some of the old historical zones such that they don't appear when running "tzselect"?
We could have a table for people who don't care about historical times, which maps zone names to "modern" zone names, and tzselect could have an option to use that table.
Paul E., If you are reading, do you have any opinions on whether or not a new zone should be added?
Since we don't have reliable data about the difference between Kenora and its neighbors, I'm inclined to leave it alone for now. But if we get reliable data, we should probably create a new zone. It is possible that our data for Rainy River are wrong as well, so perhaps no new zone would be needed.
The zone.tab file currently indicates that America/Rainy_River should be used for Rainy River and Fort Frances. The town of Fort Frances has a population of 8315. The town of Rainy River has a population of 981. How come it was not called America/Fort_Frances?
Because I didn't know that when I created the table. :-) I suppose it should be renamed....
Paul, I agree there is not much in the way of reliable data for Northern Ontario and a I agree that creating a zone without sufficient data would not be wise. Jesper, My recommendation is to use America/Rainy_River for all areas in Ontario that observe CST/CDT. But the choice is yours. I will send you some points to define a boundary between Central and Eastern zones. -chris -----Original Message----- From: Paul Eggert [mailto:eggert@CS.UCLA.EDU] Sent: November 6, 2006 12:13 PM To: Chris Walton Cc: tz@elsie.nci.nih.gov; jnorgard@prodigy.net.mx Subject: Re: Quetico (and other places in NW Ontario) "Chris Walton" <Chris.Walton@telus.com> writes:
You asked what the difference between a "Kenora" time zone and the other time zones would be. Assuming Kenora has been using DST since 1970 (which I don't have proof of) ...
Is there any way to flag some of the old historical zones such that they don't appear when running "tzselect"?
We could have a table for people who don't care about historical times, which maps zone names to "modern" zone names, and tzselect could have an option to use that table.
Paul E., If you are reading, do you have any opinions on whether or not a new zone should be added?
Since we don't have reliable data about the difference between Kenora and its neighbors, I'm inclined to leave it alone for now. But if we get reliable data, we should probably create a new zone. It is possible that our data for Rainy River are wrong as well, so perhaps no new zone would be needed.
The zone.tab file currently indicates that America/Rainy_River should be used for Rainy River and Fort Frances. The town of Fort Frances has a population of 8315. The town of Rainy River has a population of 981. How come it was not called America/Fort_Frances?
Because I didn't know that when I created the table. :-) I suppose it should be renamed....
participants (3)
-
Chris Walton -
Jesper Norgaard Welen -
Paul Eggert