Error? in VMS732_TZ Daylight Savings Patch Rule - US and Canada

[Russ Miller Jr]

I was looking at extracted time rule source file for [SYS$ZONEINFO.SYSTEM.SOURCES]NORTHAMERICA.;1

The new rules for US (and also Canada) appear incorrect, e.g.:
# Rule NAME FROM TO TYPE IN ON AT SAVE LETTER/S

Rule US 2007 max - Mar Sun>=8 2:00 1:00 D

Rule US 2007 max - Nov Sun>=1 2:00 0 S

If counting the second Sunday's in March, shouldn't the rule be as follows (or is it counting Sundays' from first of year?):

Rule US 2007 max - Mar Sun>=2 2:00 1:00 D

=============
Russ Miller

[Dave Gudewicz]

Rule US 2007 max - Mar Sun>=8 2:00 1:00 D

The second Sunday in March has a date greater than or equal to 8.

Rule US 2007 max - Nov Sun>=1 2:00 0 S

The first Sunday in November has a date greater than or equal to 1.

Dave...

[Hoff · Accepted solution]

Effective with the Energy Policy Act of 2006, starting with 2007 the DST switch-over dates are the second Sunday in March and the first Sunday in November.

http://www.doi.gov/iepa/EnergyPolicyActof2005.pdf

By my reading, the rules look correct.

That written, the TZ rules are directly from the US NIST folks, and minimally processed for use on OpenVMS. (If the rules should be wrong, they're wrong for everybody using NIST as a source that caught that update.)

As for other matters...

The rule set is (now) wrong for BC and AB, as that changed after that TZ was kitted. Unconfirmed rumors here in ITRC around an ECO kit for this, but the timing is going to be rather tight for the BC and AB sites regardless. (I'd want to confirm that rumor with HP, and otherwise get ready to make my own DST changes.)

There's a second set of rules for Java, and a Java update is now available.

There is also a report around about problems with an older ZIC, and a brief discussion of tzset on older releases.

The FAQ has a chapter on time and timekeeping on OpenVMS, FWIW.

Stephen Hoffman
HoffmanLabs

[Russ Miller Jr]

Thanks to all for clarifying the time rules.