Canada's some little problems

Paul Eggert eggert at twinsun.com
Tue May 27 05:46:49 UTC 2003


> From: "Ciro Discepolo" <discepol at tin.it>
> Date: Fri, 13 Jul 2001 14:50:36 +0200
> 
> I have noted some little problems respect DST in Canada analyzing
> between Thomas Shanks data and the Time zone data

I finally got around to decoding your attachment (sorry, I don't use
Microsoft Word, so Word documents are a bit of a pain for me).  Thanks
very much for taking the time to correct the errors in the timezone
tables; you must have gone to a lot of work!

I have the following reaction to your comments:

> Halifax does not use DST in the year 1932, while in the 1950 and
> 1955 use a DST.

Thanks for catching this error; I'll fix this in my next proposed
patch.


> From 1942 to 1944 Shanks says that ST JOHNS saved 1 hour (war time),
> while in the timezone I've a normal DST from 1942 to 1944 without
> stop.

Here the timezone database was following Whitman rather than Shanks.
However, Shanks seems more reliable about Canadian data than Whitman,
so I'll change to Shanks here.


> In the following table there are other date with difference from
> Shanks, TIMEZONE and SHANKS contains the offset from GMT in DATE
> TESTED.
> 
> ZONE       DATE TESTED    TIMEZONE   SHANKS
> ---------------------------------------------------------------------------
> GOOSE BAY     15-1-1943       3,3     2,3
> GOOSE BAY     15-1-1944       3,3     2,3

These discrepancies are for the same reason as St Johns, so changing
St Johns will also remove these discrepancies.


> GOOSE BAY     15-8-1917       2,3     3,3
>
> NIPIGON       15-1-1941       5       4
> NIPIGON       15-1-1942       5       4
> 
> THUNDER BAY   15-1-1941       5       4
> THUNDER BAY   15-1-1942       5       4
>
> RAINY RIV.    15-1-1941       6       5
> RAINY RIV.    15-1-1942       6       5
>
> REGINA        15-8-1958       6       7

Thanks for catching these transcription errors; I'll fix them too.


> MONTREAL      15-8-1917       5       4

But Montreal uses Canadian table #119, which says:

25 Mar 1917  2:00  4:00
24 Apr 1917  0:00  5:00

This means daylight saving time was not observed after April 24 of
that year, so it seems to me that the SHANKS column should be "5".
Perhaps the discrepancy is because you are using a different edition
of Shanks?  I am using the Revised 5th edition of Shanks's
International Atlas <http://astrocom.com/books/full/b112x.php>.


> WINNIPEG      15-1-1942       6       5

But Winnipeg uses Canadian table #171, which says:

26 Sep 1937  2:00  6:00
 9 Feb 1942  2:00  5:00

so Shanks is saying that daylight saving time was not observed until
February 9 of that year, so the SHANKS column should say "6".


> EDMONTON      15/08/1918      6       7

But Edmonton uses Canadian table #179, which defers to table #5 for
1918, and table #5 says:

14 Apr 1918  2:00  6:00
31 Oct 1918  2:00  7:00

so the SHANKS column should say "6".


> DAWSON FROM 1947 TO 1972      8       9

The timezone table says that Dawson was at 9 from 1947 to 1965; this
agrees with Shanks, who uses Canadian time table #253.  But that "1947
TO 1972" suggests to me that you've confused Dawson with Dawson Creek.
Dawson Creek uses Canadian time table #251 and was at 8 (7 during
summer) from 1947 to 1972.


Thanks again for checking the database; your fixes should appear
in the next version.



More information about the tz mailing list