Hi folks,
A distraction from this weekend's ADSL woes.
Specifically: The "calendar"/"clock" application in the Gnome panel is incorrect for the year 1752 (and therefore for all preceding years).
It is described as the "Clock Applet", and its Help only describes its time functions. However, when you left-click on it a calendar comes up, which is what I'm referring to. This is not described anywhere that I can find. Nor can I locate what program is executed for the "Clock Applet".
Any thoughts? [trivial question of course...]
Best wishes to all, Ted.
-------------------------------------------------------------------- E-Mail: (Ted Harding) Ted.Harding@nessie.mcc.ac.uk Fax-to-email: +44 (0)870 094 0861 Date: 08-Oct-06 Time: 15:58:22 ------------------------------ XFMail ------------------------------
On 10/8/06, Ted Harding Ted.Harding@nessie.mcc.ac.uk wrote:
Specifically: The "calendar"/"clock" application in the Gnome panel is incorrect for the year 1752 (and therefore for all preceding years).
How did you get to 1752 without 254 clicks?
How wrong is it? For example, is it correct from Thursday 14th September through to Sunday 31st December?
Tim.
On 08-Oct-06 Tim Green wrote:
On 10/8/06, Ted Harding Ted.Harding@nessie.mcc.ac.uk wrote:
Specifically: The "calendar"/"clock" application in the Gnome panel is incorrect for the year 1752 (and therefore for all preceding years).
How did you get to 1752 without 254 clicks?
I didn't. Just clicked 254 times!
How wrong is it? For example, is it correct from Thursday 14th September through to Sunday 31st December?
Tim.
I suppose so (haven't checked in that much detail). But (as Adam also implies) 3-13 September are there when they shouldn't be, so previous dates have been pushed back 11 days in the weekly cycle. (That's seriously wrong!)
See also forthcomng reply to Adam.
Ted.
-------------------------------------------------------------------- E-Mail: (Ted Harding) Ted.Harding@nessie.mcc.ac.uk Fax-to-email: +44 (0)870 094 0861 Date: 08-Oct-06 Time: 18:10:28 ------------------------------ XFMail ------------------------------
On Sun, Oct 08, 2006 at 06:29:24PM +0100, Ted Harding wrote:
I was prompted to investigate by catching someone referring to the Julian->Gregorian transition on the radio this morning.
I've known about 'cal' since 1984, and about its respect for the transition (documented in 'man cal').
Aaaah, that explains it then, you had me worried that someone could be a Unix user for over 20 years and not know about cal!
The FreeBSD version of 'cal', or rather its new version 'ncal', is more sophisticated that the standard Linux version, in that it has extra options, including "-s":
Hmmmn, appears that the version of cal (and I've got ncal too) on my Ubuntu and Debian boxen are the BSD version and that the GNU version of cal is called gcal. If I was really bored then I'd grab the source code for the Gnome clock and work out why it is wrong, but I think i'll leave that until another day ;)
Adam
On 08-Oct-06 Adam Bower wrote:
On Sun, Oct 08, 2006 at 06:29:24PM +0100, Ted Harding wrote:
I was prompted to investigate by catching someone referring to the Julian->Gregorian transition on the radio this morning.
I've known about 'cal' since 1984, and about its respect for the transition (documented in 'man cal').
Aaaah, that explains it then, you had me worried that someone could be a Unix user for over 20 years and not know about cal!
I have a vivid memory of when I first became aware of this. The man page for 'cal' on that 1984 version of Unix (By SoftQuad as I recall) mentioned it briefly, though I can't now directly recall the exact wording.
I've just shoved aside a pile of "archival material" (don't ask ... ) and delved into dust to retrieve printed manuals for SCO Xenix from 1988 which I've been hoarding. There, 'man cal' says (final para):
The calendar produced is that for England and her colonies. Note that England switched from the Julian to the Gregorian calendar in September of 1752, at which time eleven days were excised from the year. To see the result of this try "cal 9 1752".
What I do recall from 1984 is the words 'try "cal 9 1753"' though I'm not sure about the rest of the context, since I seem to recall that they were sort of dropped in for the reader to experience the consequences, without being prepared for it. But I could be wrong about that. Anyway, I did try it, and was delighted; and that's what I do recall.
In any case, this was one of the details that impressed me about Unix (in addition to the big things). Someone had taken considerable trouble to get details right, and to think of details to get right!
The FreeBSD version of 'cal', or rather its new version 'ncal', is more sophisticated that the standard Linux version, in that it has extra options, including "-s":
Hmmmn, appears that the version of cal (and I've got ncal too) on my Ubuntu and Debian boxen are the BSD version and that the GNU version of cal is called gcal.
I don't seem to have BSD cal or ncal on my most recent Linux (Red Hat 9), nor can I find gcal under that name. However, I have some recent live CDs (Ubuntu, Slackware) and will check them sometime.
If I was really bored then I'd grab the source code for the Gnome clock and work out why it is wrong, but I think i'll leave that until another day ;)
As I recently posted, I suspect that the Gnome clock simply extrapolates the Gregorian algorithm backwards into forbidden territory. Could be wrong, of course!
Best wishes, Ted.
-------------------------------------------------------------------- E-Mail: (Ted Harding) Ted.Harding@nessie.mcc.ac.uk Fax-to-email: +44 (0)870 094 0861 Date: 08-Oct-06 Time: 20:30:31 ------------------------------ XFMail ------------------------------
On 08-Oct-06 Ted Harding wrote:
On 08-Oct-06 Adam Bower wrote:
On Sun, Oct 08, 2006 at 06:29:24PM +0100, Ted Harding wrote:
I was prompted to investigate by catching someone referring to the Julian->Gregorian transition on the radio this morning.
I've known about 'cal' since 1984, and about its respect for the transition (documented in 'man cal').
Aaaah, that explains it then, you had me worried that someone could be a Unix user for over 20 years and not know about cal!
I have a vivid memory of when I first became aware of this. The man page for 'cal' on that 1984 version of Unix (By SoftQuad as I recall) mentioned it briefly, though I can't now directly recall the exact wording.
I've just shoved aside a pile of "archival material" (don't ask ... ) and delved into dust to retrieve printed manuals for SCO Xenix from 1988 which I've been hoarding. There, 'man cal' says (final para):
The calendar produced is that for England and her colonies. Note that England switched from the Julian to the Gregorian calendar in September of 1752, at which time eleven days were excised from the year. To see the result of this try "cal 9 1752".
What I do recall from 1984 is the words 'try "cal 9 1753"' though I'm not sure about the rest of the context, since I seem to recall that they were sort of dropped in for the reader to experience the consequences, without being prepared for it. But I could be wrong about that. Anyway, I did try it, and was delighted; and that's what I do recall.
Thanks to Adam's reference to W. Richard Stevens on the "ADSL" thread, WRS's web page led me to the original () source for the Unix man pages. Under 'cal' I find (stripped of formatiign commands):
prints a calendar for the specified year. If a month is also specified, a calendar just for that month is printed.
Year can be between 1 and 9999. The month is a number between 1 and 12. The calendar produced is that for England and her colonies.
Try September 1752.
and that, I'm now pretty sure, is exactly the laconic (and teasing) wording I encountered back in 1984!
In any case, this was one of the details that impressed me about Unix (in addition to the big things). Someone had taken considerable trouble to get details right, and to think of details to get right!
[...]
Cheers, Ted.
-------------------------------------------------------------------- E-Mail: (Ted Harding) Ted.Harding@nessie.mcc.ac.uk Fax-to-email: +44 (0)870 094 0861 Date: 08-Oct-06 Time: 21:14:50 ------------------------------ XFMail ------------------------------
On 08-Oct-06 Ted Harding wrote:
On 08-Oct-06 Tim Green wrote:
On 10/8/06, Ted Harding Ted.Harding@nessie.mcc.ac.uk wrote:
Specifically: The "calendar"/"clock" application in the Gnome panel is incorrect for the year 1752 (and therefore for all preceding years).
How did you get to 1752 without 254 clicks?
I didn't. Just clicked 254 times!
How wrong is it? For example, is it correct from Thursday 14th September through to Sunday 31st December?
Tim.
I suppose so (haven't checked in that much detail). But (as Adam also implies) 3-13 September are there when they shouldn't be, so previous dates have been pushed back 11 days in the weekly cycle. (That's seriously wrong!)
See also forthcomng reply to Adam.
Ted.
And it gets worse ... (the trouble with starting a hare is that one's tempted to continue chasing it).
In the Julian Calendar, there was a leap year every 4 years regardless (which was the cause of all the trouble, since the calendar year then "runs slow" relative to the annual revolution round the Sun, to the tune of 1 day every 128 years)[1].
The Gregorian Calendar omits the Leap Year in century years not exactly divisible by 400 (so only in 2000, 1600, 1200, ... is the last day of Feb Feb 29).
So I opened up the calendar in Gnome Clock again and clicked back to 1600 -- not a LY; then 1200 -- not a LY, but it should have been; and so on.
Upshot: Gnome Clock seems to use the Gregorian Algorithm for all time. So, even if you use it for pre-1752 dates and adjust by 11 days for the error in Sept 1752, you will get progressively out because it omits LYs where it shouldn't.
(Mind you, "cal 1066" is a damn sight easier than 1,960 mouse clicks).
[1] Not exactly true, since there was some sort of "clerical error" in the early years. The Julian Calendar was decreed by Julius Caesar in 45BC. Initially there were leap years every 3 years, until 9BC (12 in all, not counting 45BC); the next was 8AD, after which it was every 4 years.
The current value of the length of a year is 365.24219 days, hence a Leap Year every 4 years regardless will introduce an error of 4*(0.25 - 0.24219) = 0.03124 days every 4 years, which over 1/0.03124 = 32.0102 4-year cycles builds up to a whole day. Hence 1 day out per 128 years.
The change to the Gregorian Calendar was in fact decreed by Pope Gregory in 1582 (the Church having settled on a Julian system in 325AD [Council of Trent]), by which time the calendar was (1582 - 325)/128 = 9.8 (or 10) days out.
But his writ did not run to England by this time, since Henry VIII had already seen to that. So not until 1752 did Britain decided that there was a need to get into line with other countries (a precursor of current attitudes to the EU?), by which time we were (1752 - 325)/128 = 11.15 (or 11) days out.
Hence
$ 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
Cheers, Ted.
-------------------------------------------------------------------- E-Mail: (Ted Harding) Ted.Harding@nessie.mcc.ac.uk Fax-to-email: +44 (0)870 094 0861 Date: 08-Oct-06 Time: 19:57:50 ------------------------------ XFMail ------------------------------
On Sun, Oct 08, 2006 at 03:58:25PM +0100, Ted Harding wrote:
Specifically: The "calendar"/"clock" application in the Gnome panel is incorrect for the year 1752 (and therefore for all preceding years).
Oh, yes... hmmn, doesn't do Julian > Gregorian correctly... I wonder what Windows does, oh... how disappointing, Windows only does 1980 > 2099 how rubbish ;)
It is described as the "Clock Applet", and its Help only describes its time functions. However, when you left-click on it a calendar comes up, which is what I'm referring to. This is not described anywhere that I can find. Nor can I locate what program is executed for the "Clock Applet".
It is part of the Gnome Panel package on my Ubuntu box.
Any thoughts? [trivial question of course...]
I'm guessing that it was programmed badly not using the standard libraries for doing this kind of thing, but far more importantly how on earth did you discover this? Given that I had to hold down the mouse button for nearly a minute to scroll back to 1752! You do know about the program cal don't you? As I compared the output from gnome calendar to the output from the command cal 1752.
Adam
-
Adam Bower -
Ted.Harding＠nessie.mcc.ac.uk -
Tim Green