We have fixed this by hand-entering the timezone name as "America/Miquelon." The web server we used to obtain the timezone (given lat/lon coordinates) gave "Etc/GMT-3" and that was the primary source of our confusion. Thanks for enduring my new-to-timezones questions, and Cheers! -- David Hancock | dhancock@arinc.com

From: "Hancock, David (DHANCOCK)" <DHANCOCK@arinc.com> Date: Thu, 13 Mar 2008 11:43:10 -0400 To: <tz@elsie.nci.nih.gov>, <tz@lecserver.nci.nih.gov>, "Raymond, Bill (BRAYMOND)" <BRAYMOND@arinc.com> Conversation: Possible error with timezone for Saint Pierre/Miquelon? Subject: RE: Possible error with timezone for Saint Pierre/Miquelon?

Chris: Thanks for the note. We are using a Python wrapper around the TZ database. I will have a look at the chain starting from our side to see where the "no dst" is being introduced. The library we use gets distributed with a copy of the Olson DB, but if we see an update that's newer than the pytz distribution, we obtain the newer distribution and put it where pytz looks for it.

I'll let you know what I discover from our end. Apologies for blaming the database first! We're in the midst of encoding real timezone names for all of our airports, and this was the first error that wasn't simply having the name incorrect.

I should note, though, that worldtimeengine.com (which uses Olson data and a GIS approach for mapping locations to the timezone name) also shows no DST in 2008, but administrative region is Canada. The following URL ends with an underscore:

http://worldtimeengine.com/doTime/miquelon_st_pierre_and_miquelon_saint_pier... _miquelon_

Thanks, and Cheers! -- David Hancock | dhancock@arinc.com

-----Original Message----- From: Chris Walton [mailto:Chris.Walton@telus.com] Sent: Thursday, March 13, 2008 11:26 AM To: tz@lecserver.nci.nih.gov; Hancock, David (DHANCOCK); Raymond, Bill (BRAYMOND) Subject: RE: Possible error with timezone for Saint Pierre/Miquelon?

I am unable to replicate the "Saint Pierre/Miquelon" problem reported by David Hancock and Bill Raymond.

America/Miquelon is configured to follow the "Canada" ruleset. Daylight saving in Canada started on March 9 this year. Zone America/Miquelon -3:44:40 - LMT 1911 May 15 # St Pierre -4:00 - AST 1980 May -3:00 - PMST 1987 # Pierre & Miquelon Time -3:00 Canada PM%sT

# TZ=America/Miquelon date '+%Z %z' PMDT -0200

# zdump -v America/Miquelon | grep 2008 America/Miquelon Thu Mar 13 15:04:41 2008 UTC = Thu Mar 13 13:04:41 2008 PMDT isdst=1 America/Miquelon Sun Mar 9 04:59:59 2008 UTC = Sun Mar 9 01:59:59 2008 PMST isdst=0 America/Miquelon Sun Mar 9 05:00:00 2008 UTC = Sun Mar 9 03:00:00 2008 PMDT isdst=1 America/Miquelon Sun Nov 2 03:59:59 2008 UTC = Sun Nov 2 01:59:59 2008 PMDT isdst=1 America/Miquelon Sun Nov 2 04:00:00 2008 UTC = Sun Nov 2 01:00:00 2008 PMST isdst=0

David and/or Bill, Are you using standard compiled zone files with a standard C library? Or are you doing something more complex that might have introduced an error?

-chris -----Original Message----- From: Olson, Arthur David (NIH/NCI) [E] [mailto:olsona@dc37a.nci.nih.gov] Sent: March 13, 2008 9:19 AM To: tz@lecserver.nci.nih.gov Subject: FW: Possible error with timezone for Saint Pierre/Miquelon?

I'm forwarding this message from David Hancock; since sending it, he has asked to be added to the time zone mailing list and has been added.

--ado

From: Hancock, David (dhancock) [mailto:dhancock@arinc.com] Sent: Thursday, March 13, 2008 8:05 To: tz@lecserver.nci.nih.gov Cc: Raymond, Bill (BRAYMOND) Subject: Possible error with timezone for Saint Pierre/Miquelon?

Please see the note below from one of my colleagues, Bill Raymond. He was following up on a discrepancy in our airport local-time lookup for one of the airports our customers use. He verified that the island follows DST by calling the airport authority there. Neither of us knows the start and end dates for DST, though. One of our destination airports is LFVP - Saint Pierre/Miquelon.  There is a timezone America/Miquelon but the current setting is GMT -3 with no DST.  LFVP is currently is DST and is GMT -2.  It looks like the Olson data has DST info for 2007 but not for 2008.  I did check several sources to confirm, including making a call to Air St. Pierre at the airport to check local time.   For now I set LFVP to Etc/GMT -2.  Please let me know if there will be an update.

We recently integrated the timezone DB into our flight-planning site and it's working out very well. We've also been using worldtimeengine.com, which is great for obtaining timezone information using latitude and longitude.

Cheers! -- David Hancock | dhancock@arinc.com