On 12/17/2012 10:22 AM, Jonathan Gregory wrote:
Dear Steve
This isn't the central issue, but I wonder if I have missed your point. I
agree that there is a gap in the legal dates in the real-world mixed Julian-
Gregorian calendar, but this is just an inconvenient problem of translation,
isn't it? There is no discontinuity in real-world time!
Hi Jonathan,
We are arguably descending into a Glass Bead Game of standards, and I
think we should strive to reach closure quickly. (N.B. backwards
compatibility notwithstanding, because that *is* a vital issue.) I'll
try to head towards closure along two paths:
1. path 1: Why do we create standards?
It is to solve practical problems -- tools for more effective use of
computing power and improved interoperability. Every design choice
in standards ought to be seen as a compromise, weighing benefits to
users against degradation of the standard through accreted
complexity and implementation difficulty. The question before us --
whether to actively support a mixed Gregorian-Julian calendar --
seems exceptionally clear cut: we have _no known users_ and we are
debating a nasty, ugly encoding. On a cost-benefit basis the mixed
Julian-Gregorian calendar should be scheduled for immediate death!
(If we were to identify one or two obscure users that would not
change the conclusion.)
2. path 2: Is mixed Julian-Gregorian a real "calendar"?
No. It is not. It is a narrative -- a description of the history
of evolving methods of recording dates. Envision a method of
measuring distance that consisted of using furlongs for lengths up
to 20 furlongs but for lengths greater than this convert to meters
and subtract 11. You wouldn't want to adopt that as a scientific
measure, even it were in common usage among obscure members of (say)
the horse racing community.
1 days since 1582-10-14 = 0 days since 1582-10-25
because in the real-world calendar 14 Oct 1582 is followed by 25 Oct 1582.
If you code a time coordinate in "days since 1582-10-1" as 0, 1, 2, 3, ... 22
it should be translated into dates of 1582-10-1, 2, 3, ..., 14, 25, ..., 31.
_This is what really happened._ How could we_stamp it out_?
This is not an adequate description of "what really happened". What
really happened is that he transition from Julian to Gregorian occurred
in other countries on different dates (as you point out, yourself, in
your next paragraph.) The deeper you look into this so-called calendar
the less computable it becomes.
I am not proposing that we "stamp it out". We also need not "stamp out"
furlongs as a measure of distance. What we need do is only to liberate
CF from it. (Please.)
I note that cal(1) has the date change in Sep 1752, when it happened in England
I rest my case. Great grand daddy Unix disagrees with CF and udunits.
We should not risk enraging our own ancestors.
- Steve
$ cal 9 1752
September 1752
Su Mo Tu We Th Fr Sa
1 2 14 15 16
17 18 19 20 21 22 23
24 25 26 27 28 29 30
but anyone who works with the dates of real historical events should be aware
of the need to put them into the same calendar.
Cheers
Jonathan
----- Forwarded message from Steve Hankin <steven.c.han...@noaa.gov> -----
Date: Fri, 14 Dec 2012 15:45:27 -0800
From: Steve Hankin <steven.c.han...@noaa.gov>
User-Agent: Mozilla/5.0 (Windows NT 6.1; WOW64; rv:16.0) Gecko/20121026
Thunderbird/16.0.2
To: Jonathan Gregory <j.m.greg...@reading.ac.uk>
CC: cf-metadata@cgd.ucar.edu
Subject: Re: [CF-metadata] CF calendars (was: problem with times in PSD
dataset)
On 12/14/2012 9:35 AM, Jonathan Gregory wrote:
Dear Cecilia, Steve et al.
Steve is right that mostly we use the Gregorian calendar. That is what I meant
mostly when I said that the default is the calendar we use. The real world
is mixed Julian-Gregorian, and I don't think dealing with this calendar is an
issue only for Renaissance historians. I can't give you examples, but I
think it is conceivable or likely that at some point people would want to
record real-world data in CF earlier than the Renaissance, or have already
done so. For instance, what about astronomical data, such as the dates of
eclipses. These are real-world events, on precise dates which are translated
into the mixed Julian-Gregorian calendar.
Hi Jonathan,
If scientists somewhere have encoded the dates of these historical
events as data(*) using a mixed Gregorian-Julian calendar Lord help
'em. Those poor folks have to face an 11 day discontinuity in their
own data, as well as in ours. I'm not meaning to be snarky. I just
want to stamp out this pesky calendar issue. It has been tripping
us up for too many years.
Your points below are definitely the real guts of the discussion,
but in this email I am addressing just the one single point. I'm
afraid we will never heal ourselves from this virus if we do not
eradicate it from our thinking.
- Steve
(*) attaching a date to an historical narrative is different from
using a date as a time coordinate. It's metadata versus data.
----- End forwarded message -----
_______________________________________________
CF-metadata mailing list
CF-metadata@cgd.ucar.edu
http://mailman.cgd.ucar.edu/mailman/listinfo/cf-metadata
_______________________________________________
CF-metadata mailing list
CF-metadata@cgd.ucar.edu
http://mailman.cgd.ucar.edu/mailman/listinfo/cf-metadata