On Fri 2006-12-29T09:25:33 +0000, Clive D.W. Feather hath writ:
> Why is this challenging? It's whichever of 23 to 29 November (inclusive) is
> a Friday.
Yes, and without bothering to patch a version of the old Tcl/Tk ical
well enough to run on what I have here, I believe that is pretty much
the way its user interface presented one option for repeating that
date annually.
Try the same thing with anything that proclaims conformance with
RFC 2445 and see if it can reproduce that Friday correctly --
a Palm Pilot, Microsoft LookOut, Novell's clone Evolution, etc.
Then produce an ical (= RFC 2445) output from one such and try
feeding it to another such.
Then read RFC 2445 and try to interpret from its rules and examples
what the correct ical snippet for that repeating date should be,
and compare it with any of the output from the above programs.
I don't think it works (and I think that is in part due to one vendor
providing a suboptimal implementation that everyone else has felt
obliged to emulate), and in that is validation of the fears of folks
who worry about the handling of leap seconds. If something as obvious
as describing the annual repetitions of the day after US Thanksgiving
can't be implemented correctly in a non-real-time application, what
sorts of havoc might be wrought by a real-time application that is
supposed to interact with other real-time systems while handling a
leap second.
In the old days when Mr. Astronomer was boinking Miss Quartz clock
she just wasn't able to carry a kid to term. There were repeated
little miscarriages as the terrestrial clocks were reset, but those
left no obvious evidence. Then Miss Quartz was replaced by Ms.
Cesium who was robustly fecund. After an interval of fits and starts
that union started producing almost annual progeny -- a scandal, for
now the relationship was out in the open and every offspring had to be
named. As the word got around some folks found this unacceptable.
What we're pondering is whether Mr. Astronomer is simply going to
stop fooling around with the clocks at the BIPM, or if maybe they can
agree to use some sort protection in order to provide a less
scandalous schedule of planned parenthood.
--
Steve Allen <sla_at_ucolick.org> WGS-84 (GPS)
UCO/Lick Observatory Natural Sciences II, Room 165 Lat +36.99858
University of California Voice: +1 831 459 3046 Lng -122.06014
Santa Cruz, CA 95064 http://www.ucolick.org/~sla/ Hgt +250 m
Received on Fri Dec 29 2006 - 19:58:09 PST