FW: Questions about tzdata implementations

[Olson, Arthur David (NIH/NCI) [E]]

I'm forwarding this message from Chad Eberle, who is not on the time zone mailing list. Those of you who are on the list, please direct replies appropriately.

				--ado

-----Original Message-----
From: Chad Eberle [mailto:ceberle at co.pierce.wa.us] 
Sent: Monday, March 29, 2010 1:15
To: tz at lecserver.nci.nih.gov
Subject: Questions about tzdata implementations

I apologize if this isn't the right place for this question, but would
appreciate any help with tracking down the answer to these questions.

Background: We have Java web services that send back date/time values with GMT
offsets adjusted for DST following historical rules that match what I see in the
northamerica file for America/Los_Angeles. Our client code is largely
JavaScript, and depending on the operating system, will sometimes be incongruous
about when DST starts or ends (dates before 2007 on Windows and before 1950. As
one can imagine, this has caused a few problems.

I can't seem to find any solution for the JavaScript end, so I was planning on
writing one myself (if anyone knows of a solution, I'd appreciate any pointers).

My questions:

1) Is there some reference implementation for the algorithm that selects the
correct rule for a given date?

2) For the general US rules, there is a rule that says DST started in 1945 on
Aug 14 at 23:00; however, my tests in Java indicate that DST was observed from
Feb 9 1942 through Sep 30 1945, continuously. The latter matches the historical
information I've found. Is this rule superfluous?

Again, any tips would be appreciated.

--Chad Eberle