Year-Round Accuracy RE: Displacement Calculation & RE: Jitter Calculation RE: ....

classic Classic list List threaded Threaded
16 messages Options
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Year-Round Accuracy RE: Displacement Calculation & RE: Jitter Calculation RE: ....

Karl Palmen

Dear Michael and Calendar People

 

Here Michael addresses the problem I found with his definition of year-round accuracy, which involves only the comparison of year lengths.

 

From: East Carolina University Calendar discussion List [mailto:[hidden email]] On Behalf Of Michael Ossipoff
Sent: 31 January 2017 17:34
To: [hidden email]
Subject: Re: Displacement Calculation & RE: Jitter Calculation RE: ....

 

Here's the part that I said I'd reply to later, and I'll reply to it here:


This definition would be OK for one arithmetic calendar with one mean year, but ceases to be OK for a series of arithmetic calendars with different mean years as proposed by Michael. This is because when one changes from one calendar to the next a one-off change of displacement is made. 

Changing the value of Y, as might be done in about a millennium, or a few millennia, doesn't change D. It changes only Y.

But, as, over time, the Y value in use becomes less accurate, due to gradual change in the length of the chosen reference-year (such as the MTY), then that results in a cumulative error in D.

So, when the calendar, after a millennium or a few millennia, accumulates an unacceptable (.4 days?) amount of error (as determined by astronomical observations), then, when changing Y to the correct current value of the reference-year, the value of D could, likewise, be reset to correct the accumulated error resulting from the incorrect Y values.

Maybe you (Karl) have another way you'd like to do the longterm corrections. Fine. You have your way, I have my way, and the people 1000 years, or several thousand years, from now will no doubt have their way. If you think they need your way, and if you think that they won't come up with it on their own, then you need to give them good instructions, and store it in a time-capsule. But you'll have to be a lot more specific and clear about it than you have been.

If you're going to say that there's a problem, then you need to be a lot more specific and clear about what you think the problem is.




We discussed this last year. These changes, if not chosen carefully, can lead to a calendar with accurate mean year lengths to wander with respect the reference year, because drifts are not corrected properly.

That's why I suggested changing Y & D when the astronomical observations indicate that the calendar's center of oscillation is off by .4 day (or whatever other amount is considered unacceptable).

 

KARL REPLIES: The problem is that if one defines year-round accuracy (or any other kind of accuracy) by only comparing year lengths (such as calendar mean year (Y) and tropical year lengths), one is not going see what effect any change in the D is going to have on the accuracy.

 

I don’t understand what Michael means by “the calendar’s centre of oscillation” , but I don’t care what method is used to decide on the change of D, provided it improves the accuracy of the calendar. A measure of accuracy that only compares year lengths cannot show any such improvement of accuracy.

 

If one uses desired dates DDs for measuring accuracy, one can choose the DDs to be exactly the same every year even after Y changes. There must be no substantial change in DDs over a period of time that accuracy is measured, because such a change would allow cheating on accuracy. Note that the DDs are being used only for measuring accuracy and do not form part of the leap year rules (or first day of year rules) of any calendar.

 

Karl

 

16(06(05

 

Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: Year-Round Accuracy RE: Displacement Calculation & RE: Jitter Calculation RE: ....

Michael Ossipoff
(Instead of deleting or moving the intervening text, it's easier to just copy & paste Karl's new text at the top of this post, and leaving the previously-intervening text at the bottom of the post.)


KARL REPLIES: The problem is that if one defines year-round accuracy (or any other kind of accuracy) by only comparing year lengths (such as calendar mean year (Y) and tropical year lengths), one is not going see what effect any change in the D is going to have on the accuracy.


...but it will show what effect it has on the D1 displacement, which is what Karl advocates as the measure of calendar-displacement. And, in a recent post, I told how a mismatch between reference-year and an SEL's tropical year would affect the difference between that SEL's date of occurrence and its DD.


Karl wants to complicate the problem by bringing in a kind of "accuracy" different from his advocated D1 measure of displacement. That's fair, because I invoked accuracy for particular SELs, when I spoke of a reference-year's mean accuracy around the ecliptic.


I told how a mismatch between the lengths of the reference-year and a particular SEL's tropical-year would affect accuracy as measured by the difference between that SEL's DD and its actual date of occurrence. I said it in a recent post, and I shouldn't have to repeat it.



Karl said:


 I don’t understand what Michael means by “the calendar’s centre of oscillation”


[endquote


When something moves back and forth, or in some way changes back and forth, we sometimes say that it "oscillates".  The middle of its range of oscillation can be called its "center of oscillation".


I apologize for using a difficult word.


Karl should feel free to tell me if there are other words that he has trouble with.


To continue:


 Karl says:


A measure of accuracy that only compares year lengths cannot show any such improvement of accuracy.


[endquote]


A departure of a reference-year's length, from the length of the tropical year measured at a particular SEL, will result in a steady unidirectional drift of the mean calendar-date for that SEL, away from that SEL's DD.


The closer the reference-year's length is, to the length of the tropical-year measured at that SEL, then the lower will be that drift.


Karl says:


If one uses desired dates DDs for measuring accuracy


[emdqipte]


I don't advocate that measure of accuracy or displacement. I use the D1 measure of displacement when I speak of the accuracy of the Minimum-Displacement Calendar, and so, for fair comparison, I use it for other calendars too, especially when comparing them to the Minimum-Displaement Calendar.


Yes, I used the D2 measure of displacement for the Gregorian jitter-range, and the Gregorian's maximum displacement in a 400-year cycle.  But I did that because I wanted to state those values by both displacement-measures.


Karl says:


, one can choose the DDs to be exactly the same every year even after Y changes. T


[endquote]


Let's be clear about how much change we're talking about for Y:


I'll find my print-out of Irv's graphs of 4 tropical-year-length  in mean solar days. But, because I don't have it here right now, I ask, how fast does the length of the MTY change? 


Shall I make a rough guess?


Because the MTY is the mean of the lengths of all the tropical years, measured at all of the points of the eclitpic, then presumably there's some degree to which the movements of the graphs of the lengths of those 4 tropical years--with respect to the steady year-length change-rate due to the lengthening of the day--will cancel eachother out.  So, a reasonable first-guess would be to say that the Length of the MTY varies at a rate corresponding to the general common overall downslope of the graphs.


It's said that the mean solar day, today, is about 1/400 of a second longer than it was in 1820.


An increase of 1/400 of second, in 200 years. If that's typical, that amounts to change n day-length of a second every 80,000 years.  In 1000 years, that would be 1/80 of a second, in the length of a mean solar day.


That would amount to about 1/5 of a second change in the length of a year, in 1000 years.


Karl specifically referred to the needed change in Y after 1000 years, so he's referring to a change of about 5 seconds in the length of the year.


Admittedly that's only a guess, but it's probably reasonably close.


Is Karl saying that a need to change Y by 1/5 of a second, from 365.24217 to 365.24222, is going to cause some kind of big accuracy problem for the calendar of 3017 A.D.?


If so, then Karl needs to be a lot more specific about the nature of that problem, and exactly, in detail, how it comes about. A numerical calculation of some error resulting from that problem, whatever its nature, would be much more convincing than a mere alleging of a problem.


Michael Ossipoff





On Wed, Feb 1, 2017 at 11:21 AM, Karl Palmen <[hidden email]> wrote:

Dear Michael and Calendar People

 

Here Michael addresses the problem I found with his definition of year-round accuracy, which involves only the comparison of year lengths.

 

From: East Carolina University Calendar discussion List [mailto:[hidden email]] On Behalf Of Michael Ossipoff
Sent: 31 January 2017 17:34
To: [hidden email]
Subject: Re: Displacement Calculation & RE: Jitter Calculation RE: ....

 

Here's the part that I said I'd reply to later, and I'll reply to it here:


This definition would be OK for one arithmetic calendar with one mean year, but ceases to be OK for a series of arithmetic calendars with different mean years as proposed by Michael. This is because when one changes from one calendar to the next a one-off change of displacement is made. 

Changing the value of Y, as might be done in about a millennium, or a few millennia, doesn't change D. It changes only Y.

But, as, over time, the Y value in use becomes less accurate, due to gradual change in the length of the chosen reference-year (such as the MTY), then that results in a cumulative error in D.

So, when the calendar, after a millennium or a few millennia, accumulates an unacceptable (.4 days?) amount of error (as determined by astronomical observations), then, when changing Y to the correct current value of the reference-year, the value of D could, likewise, be reset to correct the accumulated error resulting from the incorrect Y values.

Maybe you (Karl) have another way you'd like to do the longterm corrections. Fine. You have your way, I have my way, and the people 1000 years, or several thousand years, from now will no doubt have their way. If you think they need your way, and if you think that they won't come up with it on their own, then you need to give them good instructions, and store it in a time-capsule. But you'll have to be a lot more specific and clear about it than you have been.

If you're going to say that there's a problem, then you need to be a lot more specific and clear about what you think the problem is.




We discussed this last year. These changes, if not chosen carefully, can lead to a calendar with accurate mean year lengths to wander with respect the reference year, because drifts are not corrected properly.

That's why I suggested changing Y & D when the astronomical observations indicate that the calendar's center of oscillation is off by .4 day (or whatever other amount is considered unacceptable).

 

KARL REPLIES: The problem is that if one defines year-round accuracy (or any other kind of accuracy) by only comparing year lengths (such as calendar mean year (Y) and tropical year lengths), one is not going see what effect any change in the D is going to have on the accuracy.

 

I don’t understand what Michael means by “the calendar’s centre of oscillation” , but I don’t care what method is used to decide on the change of D, provided it improves the accuracy of the calendar. A measure of accuracy that only compares year lengths cannot show any such improvement of accuracy.

 

If one uses desired dates DDs for measuring accuracy, one can choose the DDs to be exactly the same every year even after Y changes. There must be no substantial change in DDs over a period of time that accuracy is measured, because such a change would allow cheating on accuracy. Note that the DDs are being used only for measuring accuracy and do not form part of the leap year rules (or first day of year rules) of any calendar.

 

Karl

 

16(06(05

 


Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: Year-Round Accuracy RE: Displacement Calculation & RE: Jitter Calculation RE: ....

Michael Ossipoff
I said:

"Is Karl saying that a need to change Y by 1/5 of a second, from 365.24217 to 365.24222, is going to cause some kind of big accuracy problem for the calendar of 3017 A.D.?"

I meant:

Is Karl saying that a need to change Y by 5 seconds, from 365.24217 days to 365.24222 days, is going to cause some kind of big accuracy problem for the calendar of 3017 A.D.?

On Wed, Feb 1, 2017 at 6:23 PM, Michael Ossipoff <[hidden email]> wrote:
(Instead of deleting or moving the intervening text, it's easier to just copy & paste Karl's new text at the top of this post, and leaving the previously-intervening text at the bottom of the post.)


KARL REPLIES: The problem is that if one defines year-round accuracy (or any other kind of accuracy) by only comparing year lengths (such as calendar mean year (Y) and tropical year lengths), one is not going see what effect any change in the D is going to have on the accuracy.


...but it will show what effect it has on the D1 displacement, which is what Karl advocates as the measure of calendar-displacement. And, in a recent post, I told how a mismatch between reference-year and an SEL's tropical year would affect the difference between that SEL's date of occurrence and its DD.


Karl wants to complicate the problem by bringing in a kind of "accuracy" different from his advocated D1 measure of displacement. That's fair, because I invoked accuracy for particular SELs, when I spoke of a reference-year's mean accuracy around the ecliptic.


I told how a mismatch between the lengths of the reference-year and a particular SEL's tropical-year would affect accuracy as measured by the difference between that SEL's DD and its actual date of occurrence. I said it in a recent post, and I shouldn't have to repeat it.



Karl said:


 I don’t understand what Michael means by “the calendar’s centre of oscillation”


[endquote


When something moves back and forth, or in some way changes back and forth, we sometimes say that it "oscillates".  The middle of its range of oscillation can be called its "center of oscillation".


I apologize for using a difficult word.


Karl should feel free to tell me if there are other words that he has trouble with.


To continue:


 Karl says:


A measure of accuracy that only compares year lengths cannot show any such improvement of accuracy.


[endquote]


A departure of a reference-year's length, from the length of the tropical year measured at a particular SEL, will result in a steady unidirectional drift of the mean calendar-date for that SEL, away from that SEL's DD.


The closer the reference-year's length is, to the length of the tropical-year measured at that SEL, then the lower will be that drift.


Karl says:


If one uses desired dates DDs for measuring accuracy


[emdqipte]


I don't advocate that measure of accuracy or displacement. I use the D1 measure of displacement when I speak of the accuracy of the Minimum-Displacement Calendar, and so, for fair comparison, I use it for other calendars too, especially when comparing them to the Minimum-Displaement Calendar.


Yes, I used the D2 measure of displacement for the Gregorian jitter-range, and the Gregorian's maximum displacement in a 400-year cycle.  But I did that because I wanted to state those values by both displacement-measures.


Karl says:


, one can choose the DDs to be exactly the same every year even after Y changes. T


[endquote]


Let's be clear about how much change we're talking about for Y:


I'll find my print-out of Irv's graphs of 4 tropical-year-length  in mean solar days. But, because I don't have it here right now, I ask, how fast does the length of the MTY change? 


Shall I make a rough guess?


Because the MTY is the mean of the lengths of all the tropical years, measured at all of the points of the eclitpic, then presumably there's some degree to which the movements of the graphs of the lengths of those 4 tropical years--with respect to the steady year-length change-rate due to the lengthening of the day--will cancel eachother out.  So, a reasonable first-guess would be to say that the Length of the MTY varies at a rate corresponding to the general common overall downslope of the graphs.


It's said that the mean solar day, today, is about 1/400 of a second longer than it was in 1820.


An increase of 1/400 of second, in 200 years. If that's typical, that amounts to change n day-length of a second every 80,000 years.  In 1000 years, that would be 1/80 of a second, in the length of a mean solar day.


That would amount to about 1/5 of a second change in the length of a year, in 1000 years.


Karl specifically referred to the needed change in Y after 1000 years, so he's referring to a change of about 5 seconds in the length of the year.


Admittedly that's only a guess, but it's probably reasonably close.


Is Karl saying that a need to change Y by 1/5 of a second, from 365.24217 to 365.24222, is going to cause some kind of big accuracy problem for the calendar of 3017 A.D.?


If so, then Karl needs to be a lot more specific about the nature of that problem, and exactly, in detail, how it comes about. A numerical calculation of some error resulting from that problem, whatever its nature, would be much more convincing than a mere alleging of a problem.


Michael Ossipoff





On Wed, Feb 1, 2017 at 11:21 AM, Karl Palmen <[hidden email]> wrote:

Dear Michael and Calendar People

 

Here Michael addresses the problem I found with his definition of year-round accuracy, which involves only the comparison of year lengths.

 

From: East Carolina University Calendar discussion List [mailto:[hidden email]] On Behalf Of Michael Ossipoff
Sent: 31 January 2017 17:34
To: [hidden email]
Subject: Re: Displacement Calculation & RE: Jitter Calculation RE: ....

 

Here's the part that I said I'd reply to later, and I'll reply to it here:


This definition would be OK for one arithmetic calendar with one mean year, but ceases to be OK for a series of arithmetic calendars with different mean years as proposed by Michael. This is because when one changes from one calendar to the next a one-off change of displacement is made. 

Changing the value of Y, as might be done in about a millennium, or a few millennia, doesn't change D. It changes only Y.

But, as, over time, the Y value in use becomes less accurate, due to gradual change in the length of the chosen reference-year (such as the MTY), then that results in a cumulative error in D.

So, when the calendar, after a millennium or a few millennia, accumulates an unacceptable (.4 days?) amount of error (as determined by astronomical observations), then, when changing Y to the correct current value of the reference-year, the value of D could, likewise, be reset to correct the accumulated error resulting from the incorrect Y values.

Maybe you (Karl) have another way you'd like to do the longterm corrections. Fine. You have your way, I have my way, and the people 1000 years, or several thousand years, from now will no doubt have their way. If you think they need your way, and if you think that they won't come up with it on their own, then you need to give them good instructions, and store it in a time-capsule. But you'll have to be a lot more specific and clear about it than you have been.

If you're going to say that there's a problem, then you need to be a lot more specific and clear about what you think the problem is.




We discussed this last year. These changes, if not chosen carefully, can lead to a calendar with accurate mean year lengths to wander with respect the reference year, because drifts are not corrected properly.

That's why I suggested changing Y & D when the astronomical observations indicate that the calendar's center of oscillation is off by .4 day (or whatever other amount is considered unacceptable).

 

KARL REPLIES: The problem is that if one defines year-round accuracy (or any other kind of accuracy) by only comparing year lengths (such as calendar mean year (Y) and tropical year lengths), one is not going see what effect any change in the D is going to have on the accuracy.

 

I don’t understand what Michael means by “the calendar’s centre of oscillation” , but I don’t care what method is used to decide on the change of D, provided it improves the accuracy of the calendar. A measure of accuracy that only compares year lengths cannot show any such improvement of accuracy.

 

If one uses desired dates DDs for measuring accuracy, one can choose the DDs to be exactly the same every year even after Y changes. There must be no substantial change in DDs over a period of time that accuracy is measured, because such a change would allow cheating on accuracy. Note that the DDs are being used only for measuring accuracy and do not form part of the leap year rules (or first day of year rules) of any calendar.

 

Karl

 

16(06(05

 



Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: Year-Round Accuracy RE: Displacement Calculation & RE: Jitter Calculation RE: ....

Michael Ossipoff
Oops! Because of the mean solar day being 1/80 of a second longer, then a year, of whatever kind, will be very slightly shorter, not longer, as a result.

So, my question should be re-worded:

Is Karl saying that the shortening of the MTY, in terms of mean solar days (as a result of the 1/80 of a second longer mean solar day), and the resulting small reduction in Y,  will cause some significant "inaccuracy" for the calendar of 3017 A.D.?

Michael Ossipoff

Michael Ossipoff

On Wed, Feb 1, 2017 at 6:35 PM, Michael Ossipoff <[hidden email]> wrote:
I said:

"Is Karl saying that a need to change Y by 1/5 of a second, from 365.24217 to 365.24222, is going to cause some kind of big accuracy problem for the calendar of 3017 A.D.?"

I meant:

Is Karl saying that a need to change Y by 5 seconds, from 365.24217 days to 365.24222 days, is going to cause some kind of big accuracy problem for the calendar of 3017 A.D.?

On Wed, Feb 1, 2017 at 6:23 PM, Michael Ossipoff <[hidden email]> wrote:
(Instead of deleting or moving the intervening text, it's easier to just copy & paste Karl's new text at the top of this post, and leaving the previously-intervening text at the bottom of the post.)


KARL REPLIES: The problem is that if one defines year-round accuracy (or any other kind of accuracy) by only comparing year lengths (such as calendar mean year (Y) and tropical year lengths), one is not going see what effect any change in the D is going to have on the accuracy.


...but it will show what effect it has on the D1 displacement, which is what Karl advocates as the measure of calendar-displacement. And, in a recent post, I told how a mismatch between reference-year and an SEL's tropical year would affect the difference between that SEL's date of occurrence and its DD.


Karl wants to complicate the problem by bringing in a kind of "accuracy" different from his advocated D1 measure of displacement. That's fair, because I invoked accuracy for particular SELs, when I spoke of a reference-year's mean accuracy around the ecliptic.


I told how a mismatch between the lengths of the reference-year and a particular SEL's tropical-year would affect accuracy as measured by the difference between that SEL's DD and its actual date of occurrence. I said it in a recent post, and I shouldn't have to repeat it.



Karl said:


 I don’t understand what Michael means by “the calendar’s centre of oscillation”


[endquote


When something moves back and forth, or in some way changes back and forth, we sometimes say that it "oscillates".  The middle of its range of oscillation can be called its "center of oscillation".


I apologize for using a difficult word.


Karl should feel free to tell me if there are other words that he has trouble with.


To continue:


 Karl says:


A measure of accuracy that only compares year lengths cannot show any such improvement of accuracy.


[endquote]


A departure of a reference-year's length, from the length of the tropical year measured at a particular SEL, will result in a steady unidirectional drift of the mean calendar-date for that SEL, away from that SEL's DD.


The closer the reference-year's length is, to the length of the tropical-year measured at that SEL, then the lower will be that drift.


Karl says:


If one uses desired dates DDs for measuring accuracy


[emdqipte]


I don't advocate that measure of accuracy or displacement. I use the D1 measure of displacement when I speak of the accuracy of the Minimum-Displacement Calendar, and so, for fair comparison, I use it for other calendars too, especially when comparing them to the Minimum-Displaement Calendar.


Yes, I used the D2 measure of displacement for the Gregorian jitter-range, and the Gregorian's maximum displacement in a 400-year cycle.  But I did that because I wanted to state those values by both displacement-measures.


Karl says:


, one can choose the DDs to be exactly the same every year even after Y changes. T


[endquote]


Let's be clear about how much change we're talking about for Y:


I'll find my print-out of Irv's graphs of 4 tropical-year-length  in mean solar days. But, because I don't have it here right now, I ask, how fast does the length of the MTY change? 


Shall I make a rough guess?


Because the MTY is the mean of the lengths of all the tropical years, measured at all of the points of the eclitpic, then presumably there's some degree to which the movements of the graphs of the lengths of those 4 tropical years--with respect to the steady year-length change-rate due to the lengthening of the day--will cancel eachother out.  So, a reasonable first-guess would be to say that the Length of the MTY varies at a rate corresponding to the general common overall downslope of the graphs.


It's said that the mean solar day, today, is about 1/400 of a second longer than it was in 1820.


An increase of 1/400 of second, in 200 years. If that's typical, that amounts to change n day-length of a second every 80,000 years.  In 1000 years, that would be 1/80 of a second, in the length of a mean solar day.


That would amount to about 1/5 of a second change in the length of a year, in 1000 years.


Karl specifically referred to the needed change in Y after 1000 years, so he's referring to a change of about 5 seconds in the length of the year.


Admittedly that's only a guess, but it's probably reasonably close.


Is Karl saying that a need to change Y by 1/5 of a second, from 365.24217 to 365.24222, is going to cause some kind of big accuracy problem for the calendar of 3017 A.D.?


If so, then Karl needs to be a lot more specific about the nature of that problem, and exactly, in detail, how it comes about. A numerical calculation of some error resulting from that problem, whatever its nature, would be much more convincing than a mere alleging of a problem.


Michael Ossipoff





On Wed, Feb 1, 2017 at 11:21 AM, Karl Palmen <[hidden email]> wrote:

Dear Michael and Calendar People

 

Here Michael addresses the problem I found with his definition of year-round accuracy, which involves only the comparison of year lengths.

 

From: East Carolina University Calendar discussion List [mailto:[hidden email]] On Behalf Of Michael Ossipoff
Sent: 31 January 2017 17:34
To: [hidden email]
Subject: Re: Displacement Calculation & RE: Jitter Calculation RE: ....

 

Here's the part that I said I'd reply to later, and I'll reply to it here:


This definition would be OK for one arithmetic calendar with one mean year, but ceases to be OK for a series of arithmetic calendars with different mean years as proposed by Michael. This is because when one changes from one calendar to the next a one-off change of displacement is made. 

Changing the value of Y, as might be done in about a millennium, or a few millennia, doesn't change D. It changes only Y.

But, as, over time, the Y value in use becomes less accurate, due to gradual change in the length of the chosen reference-year (such as the MTY), then that results in a cumulative error in D.

So, when the calendar, after a millennium or a few millennia, accumulates an unacceptable (.4 days?) amount of error (as determined by astronomical observations), then, when changing Y to the correct current value of the reference-year, the value of D could, likewise, be reset to correct the accumulated error resulting from the incorrect Y values.

Maybe you (Karl) have another way you'd like to do the longterm corrections. Fine. You have your way, I have my way, and the people 1000 years, or several thousand years, from now will no doubt have their way. If you think they need your way, and if you think that they won't come up with it on their own, then you need to give them good instructions, and store it in a time-capsule. But you'll have to be a lot more specific and clear about it than you have been.

If you're going to say that there's a problem, then you need to be a lot more specific and clear about what you think the problem is.




We discussed this last year. These changes, if not chosen carefully, can lead to a calendar with accurate mean year lengths to wander with respect the reference year, because drifts are not corrected properly.

That's why I suggested changing Y & D when the astronomical observations indicate that the calendar's center of oscillation is off by .4 day (or whatever other amount is considered unacceptable).

 

KARL REPLIES: The problem is that if one defines year-round accuracy (or any other kind of accuracy) by only comparing year lengths (such as calendar mean year (Y) and tropical year lengths), one is not going see what effect any change in the D is going to have on the accuracy.

 

I don’t understand what Michael means by “the calendar’s centre of oscillation” , but I don’t care what method is used to decide on the change of D, provided it improves the accuracy of the calendar. A measure of accuracy that only compares year lengths cannot show any such improvement of accuracy.

 

If one uses desired dates DDs for measuring accuracy, one can choose the DDs to be exactly the same every year even after Y changes. There must be no substantial change in DDs over a period of time that accuracy is measured, because such a change would allow cheating on accuracy. Note that the DDs are being used only for measuring accuracy and do not form part of the leap year rules (or first day of year rules) of any calendar.

 

Karl

 

16(06(05

 




Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: Year-Round Accuracy RE: Displacement Calculation & RE: Jitter Calculation RE: ....

Michael Ossipoff
...meaning that, in the year 3017, the MTY will very slightly shorter in terms of mean solar days, because each mean solar day will be about 1/80 of a second longer than a mean solar day is now.

Michael Ossipoff

On Wed, Feb 1, 2017 at 6:54 PM, Michael Ossipoff <[hidden email]> wrote:
Oops! Because of the mean solar day being 1/80 of a second longer, then a year, of whatever kind, will be very slightly shorter, not longer, as a result.

So, my question should be re-worded:

Is Karl saying that the shortening of the MTY, in terms of mean solar days (as a result of the 1/80 of a second longer mean solar day), and the resulting small reduction in Y,  will cause some significant "inaccuracy" for the calendar of 3017 A.D.?

Michael Ossipoff

Michael Ossipoff

On Wed, Feb 1, 2017 at 6:35 PM, Michael Ossipoff <[hidden email]> wrote:
I said:

"Is Karl saying that a need to change Y by 1/5 of a second, from 365.24217 to 365.24222, is going to cause some kind of big accuracy problem for the calendar of 3017 A.D.?"

I meant:

Is Karl saying that a need to change Y by 5 seconds, from 365.24217 days to 365.24222 days, is going to cause some kind of big accuracy problem for the calendar of 3017 A.D.?

On Wed, Feb 1, 2017 at 6:23 PM, Michael Ossipoff <[hidden email]> wrote:
(Instead of deleting or moving the intervening text, it's easier to just copy & paste Karl's new text at the top of this post, and leaving the previously-intervening text at the bottom of the post.)


KARL REPLIES: The problem is that if one defines year-round accuracy (or any other kind of accuracy) by only comparing year lengths (such as calendar mean year (Y) and tropical year lengths), one is not going see what effect any change in the D is going to have on the accuracy.


...but it will show what effect it has on the D1 displacement, which is what Karl advocates as the measure of calendar-displacement. And, in a recent post, I told how a mismatch between reference-year and an SEL's tropical year would affect the difference between that SEL's date of occurrence and its DD.


Karl wants to complicate the problem by bringing in a kind of "accuracy" different from his advocated D1 measure of displacement. That's fair, because I invoked accuracy for particular SELs, when I spoke of a reference-year's mean accuracy around the ecliptic.


I told how a mismatch between the lengths of the reference-year and a particular SEL's tropical-year would affect accuracy as measured by the difference between that SEL's DD and its actual date of occurrence. I said it in a recent post, and I shouldn't have to repeat it.



Karl said:


 I don’t understand what Michael means by “the calendar’s centre of oscillation”


[endquote


When something moves back and forth, or in some way changes back and forth, we sometimes say that it "oscillates".  The middle of its range of oscillation can be called its "center of oscillation".


I apologize for using a difficult word.


Karl should feel free to tell me if there are other words that he has trouble with.


To continue:


 Karl says:


A measure of accuracy that only compares year lengths cannot show any such improvement of accuracy.


[endquote]


A departure of a reference-year's length, from the length of the tropical year measured at a particular SEL, will result in a steady unidirectional drift of the mean calendar-date for that SEL, away from that SEL's DD.


The closer the reference-year's length is, to the length of the tropical-year measured at that SEL, then the lower will be that drift.


Karl says:


If one uses desired dates DDs for measuring accuracy


[emdqipte]


I don't advocate that measure of accuracy or displacement. I use the D1 measure of displacement when I speak of the accuracy of the Minimum-Displacement Calendar, and so, for fair comparison, I use it for other calendars too, especially when comparing them to the Minimum-Displaement Calendar.


Yes, I used the D2 measure of displacement for the Gregorian jitter-range, and the Gregorian's maximum displacement in a 400-year cycle.  But I did that because I wanted to state those values by both displacement-measures.


Karl says:


, one can choose the DDs to be exactly the same every year even after Y changes. T


[endquote]


Let's be clear about how much change we're talking about for Y:


I'll find my print-out of Irv's graphs of 4 tropical-year-length  in mean solar days. But, because I don't have it here right now, I ask, how fast does the length of the MTY change? 


Shall I make a rough guess?


Because the MTY is the mean of the lengths of all the tropical years, measured at all of the points of the eclitpic, then presumably there's some degree to which the movements of the graphs of the lengths of those 4 tropical years--with respect to the steady year-length change-rate due to the lengthening of the day--will cancel eachother out.  So, a reasonable first-guess would be to say that the Length of the MTY varies at a rate corresponding to the general common overall downslope of the graphs.


It's said that the mean solar day, today, is about 1/400 of a second longer than it was in 1820.


An increase of 1/400 of second, in 200 years. If that's typical, that amounts to change n day-length of a second every 80,000 years.  In 1000 years, that would be 1/80 of a second, in the length of a mean solar day.


That would amount to about 1/5 of a second change in the length of a year, in 1000 years.


Karl specifically referred to the needed change in Y after 1000 years, so he's referring to a change of about 5 seconds in the length of the year.


Admittedly that's only a guess, but it's probably reasonably close.


Is Karl saying that a need to change Y by 1/5 of a second, from 365.24217 to 365.24222, is going to cause some kind of big accuracy problem for the calendar of 3017 A.D.?


If so, then Karl needs to be a lot more specific about the nature of that problem, and exactly, in detail, how it comes about. A numerical calculation of some error resulting from that problem, whatever its nature, would be much more convincing than a mere alleging of a problem.


Michael Ossipoff





On Wed, Feb 1, 2017 at 11:21 AM, Karl Palmen <[hidden email]> wrote:

Dear Michael and Calendar People

 

Here Michael addresses the problem I found with his definition of year-round accuracy, which involves only the comparison of year lengths.

 

From: East Carolina University Calendar discussion List [mailto:[hidden email]] On Behalf Of Michael Ossipoff
Sent: 31 January 2017 17:34
To: [hidden email]
Subject: Re: Displacement Calculation & RE: Jitter Calculation RE: ....

 

Here's the part that I said I'd reply to later, and I'll reply to it here:


This definition would be OK for one arithmetic calendar with one mean year, but ceases to be OK for a series of arithmetic calendars with different mean years as proposed by Michael. This is because when one changes from one calendar to the next a one-off change of displacement is made. 

Changing the value of Y, as might be done in about a millennium, or a few millennia, doesn't change D. It changes only Y.

But, as, over time, the Y value in use becomes less accurate, due to gradual change in the length of the chosen reference-year (such as the MTY), then that results in a cumulative error in D.

So, when the calendar, after a millennium or a few millennia, accumulates an unacceptable (.4 days?) amount of error (as determined by astronomical observations), then, when changing Y to the correct current value of the reference-year, the value of D could, likewise, be reset to correct the accumulated error resulting from the incorrect Y values.

Maybe you (Karl) have another way you'd like to do the longterm corrections. Fine. You have your way, I have my way, and the people 1000 years, or several thousand years, from now will no doubt have their way. If you think they need your way, and if you think that they won't come up with it on their own, then you need to give them good instructions, and store it in a time-capsule. But you'll have to be a lot more specific and clear about it than you have been.

If you're going to say that there's a problem, then you need to be a lot more specific and clear about what you think the problem is.




We discussed this last year. These changes, if not chosen carefully, can lead to a calendar with accurate mean year lengths to wander with respect the reference year, because drifts are not corrected properly.

That's why I suggested changing Y & D when the astronomical observations indicate that the calendar's center of oscillation is off by .4 day (or whatever other amount is considered unacceptable).

 

KARL REPLIES: The problem is that if one defines year-round accuracy (or any other kind of accuracy) by only comparing year lengths (such as calendar mean year (Y) and tropical year lengths), one is not going see what effect any change in the D is going to have on the accuracy.

 

I don’t understand what Michael means by “the calendar’s centre of oscillation” , but I don’t care what method is used to decide on the change of D, provided it improves the accuracy of the calendar. A measure of accuracy that only compares year lengths cannot show any such improvement of accuracy.

 

If one uses desired dates DDs for measuring accuracy, one can choose the DDs to be exactly the same every year even after Y changes. There must be no substantial change in DDs over a period of time that accuracy is measured, because such a change would allow cheating on accuracy. Note that the DDs are being used only for measuring accuracy and do not form part of the leap year rules (or first day of year rules) of any calendar.

 

Karl

 

16(06(05

 





Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: Year-Round Accuracy RE: Displacement Calculation & RE: Jitter Calculation RE: ....

Michael Ossipoff
And, in general, if we adjust Y & D, when the accumulated error of D reaches some unacceptable value (I've suggested 4/10 of a day, for example), and if that happens only after a thousand or a few thousand years, then, when the error in D just becomes unacceptable, the error in y is thousands of times less than that. (...because D's error is the accumulated time-integration of Y's increasing error over thousands of years).

So, said briefly, when Y (and D) needs to be changed because D has just become off by an unacceptable fraction of a day, then, at that time, Y will only need to be changed by a fraction of a day that's thousands of times less than that barely-unacceptable amount.

In other words: No problem.

...aside from the fact that adjustments resulting from an unacceptable error in D are the business of people a thousand or several thousand years from now. They'll deal with it, as they choose. We needn't worry about it.

I agree with that other poster who suggested that some people here are too future-millennia-oriented.

Michael Ossipoff





On Wed, Feb 1, 2017 at 7:13 PM, Michael Ossipoff <[hidden email]> wrote:
...meaning that, in the year 3017, the MTY will very slightly shorter in terms of mean solar days, because each mean solar day will be about 1/80 of a second longer than a mean solar day is now.

Michael Ossipoff

On Wed, Feb 1, 2017 at 6:54 PM, Michael Ossipoff <[hidden email]> wrote:
Oops! Because of the mean solar day being 1/80 of a second longer, then a year, of whatever kind, will be very slightly shorter, not longer, as a result.

So, my question should be re-worded:

Is Karl saying that the shortening of the MTY, in terms of mean solar days (as a result of the 1/80 of a second longer mean solar day), and the resulting small reduction in Y,  will cause some significant "inaccuracy" for the calendar of 3017 A.D.?

Michael Ossipoff

Michael Ossipoff

On Wed, Feb 1, 2017 at 6:35 PM, Michael Ossipoff <[hidden email]> wrote:
I said:

"Is Karl saying that a need to change Y by 1/5 of a second, from 365.24217 to 365.24222, is going to cause some kind of big accuracy problem for the calendar of 3017 A.D.?"

I meant:

Is Karl saying that a need to change Y by 5 seconds, from 365.24217 days to 365.24222 days, is going to cause some kind of big accuracy problem for the calendar of 3017 A.D.?

On Wed, Feb 1, 2017 at 6:23 PM, Michael Ossipoff <[hidden email]> wrote:
(Instead of deleting or moving the intervening text, it's easier to just copy & paste Karl's new text at the top of this post, and leaving the previously-intervening text at the bottom of the post.)


KARL REPLIES: The problem is that if one defines year-round accuracy (or any other kind of accuracy) by only comparing year lengths (such as calendar mean year (Y) and tropical year lengths), one is not going see what effect any change in the D is going to have on the accuracy.


...but it will show what effect it has on the D1 displacement, which is what Karl advocates as the measure of calendar-displacement. And, in a recent post, I told how a mismatch between reference-year and an SEL's tropical year would affect the difference between that SEL's date of occurrence and its DD.


Karl wants to complicate the problem by bringing in a kind of "accuracy" different from his advocated D1 measure of displacement. That's fair, because I invoked accuracy for particular SELs, when I spoke of a reference-year's mean accuracy around the ecliptic.


I told how a mismatch between the lengths of the reference-year and a particular SEL's tropical-year would affect accuracy as measured by the difference between that SEL's DD and its actual date of occurrence. I said it in a recent post, and I shouldn't have to repeat it.



Karl said:


 I don’t understand what Michael means by “the calendar’s centre of oscillation”


[endquote


When something moves back and forth, or in some way changes back and forth, we sometimes say that it "oscillates".  The middle of its range of oscillation can be called its "center of oscillation".


I apologize for using a difficult word.


Karl should feel free to tell me if there are other words that he has trouble with.


To continue:


 Karl says:


A measure of accuracy that only compares year lengths cannot show any such improvement of accuracy.


[endquote]


A departure of a reference-year's length, from the length of the tropical year measured at a particular SEL, will result in a steady unidirectional drift of the mean calendar-date for that SEL, away from that SEL's DD.


The closer the reference-year's length is, to the length of the tropical-year measured at that SEL, then the lower will be that drift.


Karl says:


If one uses desired dates DDs for measuring accuracy


[emdqipte]


I don't advocate that measure of accuracy or displacement. I use the D1 measure of displacement when I speak of the accuracy of the Minimum-Displacement Calendar, and so, for fair comparison, I use it for other calendars too, especially when comparing them to the Minimum-Displaement Calendar.


Yes, I used the D2 measure of displacement for the Gregorian jitter-range, and the Gregorian's maximum displacement in a 400-year cycle.  But I did that because I wanted to state those values by both displacement-measures.


Karl says:


, one can choose the DDs to be exactly the same every year even after Y changes. T


[endquote]


Let's be clear about how much change we're talking about for Y:


I'll find my print-out of Irv's graphs of 4 tropical-year-length  in mean solar days. But, because I don't have it here right now, I ask, how fast does the length of the MTY change? 


Shall I make a rough guess?


Because the MTY is the mean of the lengths of all the tropical years, measured at all of the points of the eclitpic, then presumably there's some degree to which the movements of the graphs of the lengths of those 4 tropical years--with respect to the steady year-length change-rate due to the lengthening of the day--will cancel eachother out.  So, a reasonable first-guess would be to say that the Length of the MTY varies at a rate corresponding to the general common overall downslope of the graphs.


It's said that the mean solar day, today, is about 1/400 of a second longer than it was in 1820.


An increase of 1/400 of second, in 200 years. If that's typical, that amounts to change n day-length of a second every 80,000 years.  In 1000 years, that would be 1/80 of a second, in the length of a mean solar day.


That would amount to about 1/5 of a second change in the length of a year, in 1000 years.


Karl specifically referred to the needed change in Y after 1000 years, so he's referring to a change of about 5 seconds in the length of the year.


Admittedly that's only a guess, but it's probably reasonably close.


Is Karl saying that a need to change Y by 1/5 of a second, from 365.24217 to 365.24222, is going to cause some kind of big accuracy problem for the calendar of 3017 A.D.?


If so, then Karl needs to be a lot more specific about the nature of that problem, and exactly, in detail, how it comes about. A numerical calculation of some error resulting from that problem, whatever its nature, would be much more convincing than a mere alleging of a problem.


Michael Ossipoff





On Wed, Feb 1, 2017 at 11:21 AM, Karl Palmen <[hidden email]> wrote:

Dear Michael and Calendar People

 

Here Michael addresses the problem I found with his definition of year-round accuracy, which involves only the comparison of year lengths.

 

From: East Carolina University Calendar discussion List [mailto:[hidden email]] On Behalf Of Michael Ossipoff
Sent: 31 January 2017 17:34
To: [hidden email]
Subject: Re: Displacement Calculation & RE: Jitter Calculation RE: ....

 

Here's the part that I said I'd reply to later, and I'll reply to it here:


This definition would be OK for one arithmetic calendar with one mean year, but ceases to be OK for a series of arithmetic calendars with different mean years as proposed by Michael. This is because when one changes from one calendar to the next a one-off change of displacement is made. 

Changing the value of Y, as might be done in about a millennium, or a few millennia, doesn't change D. It changes only Y.

But, as, over time, the Y value in use becomes less accurate, due to gradual change in the length of the chosen reference-year (such as the MTY), then that results in a cumulative error in D.

So, when the calendar, after a millennium or a few millennia, accumulates an unacceptable (.4 days?) amount of error (as determined by astronomical observations), then, when changing Y to the correct current value of the reference-year, the value of D could, likewise, be reset to correct the accumulated error resulting from the incorrect Y values.

Maybe you (Karl) have another way you'd like to do the longterm corrections. Fine. You have your way, I have my way, and the people 1000 years, or several thousand years, from now will no doubt have their way. If you think they need your way, and if you think that they won't come up with it on their own, then you need to give them good instructions, and store it in a time-capsule. But you'll have to be a lot more specific and clear about it than you have been.

If you're going to say that there's a problem, then you need to be a lot more specific and clear about what you think the problem is.




We discussed this last year. These changes, if not chosen carefully, can lead to a calendar with accurate mean year lengths to wander with respect the reference year, because drifts are not corrected properly.

That's why I suggested changing Y & D when the astronomical observations indicate that the calendar's center of oscillation is off by .4 day (or whatever other amount is considered unacceptable).

 

KARL REPLIES: The problem is that if one defines year-round accuracy (or any other kind of accuracy) by only comparing year lengths (such as calendar mean year (Y) and tropical year lengths), one is not going see what effect any change in the D is going to have on the accuracy.

 

I don’t understand what Michael means by “the calendar’s centre of oscillation” , but I don’t care what method is used to decide on the change of D, provided it improves the accuracy of the calendar. A measure of accuracy that only compares year lengths cannot show any such improvement of accuracy.

 

If one uses desired dates DDs for measuring accuracy, one can choose the DDs to be exactly the same every year even after Y changes. There must be no substantial change in DDs over a period of time that accuracy is measured, because such a change would allow cheating on accuracy. Note that the DDs are being used only for measuring accuracy and do not form part of the leap year rules (or first day of year rules) of any calendar.

 

Karl

 

16(06(05

 






Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: Year-Round Accuracy RE: Displacement Calculation & RE: Jitter Calculation RE: ....

Karl Palmen
In reply to this post by Michael Ossipoff

Dear Michael and Calendar People

 

It is obvious to me that Michael has not understood what I’ve been attempting to express at all.  So I’ll make one more attempt with a simple example possibly exaggerated for clarity.

 

Consider two calendars whose calendar mean years (Y) are exactly equal every year. The Y values are changed at exactly the same year to exactly the same new value. The only difference is that the first calendar never changes the displacement (D) whenever Y is changed and the second calendar always reduces the displacement by 0.2 day, whenever Y is changed. Then after five Y changes the second calendar will be running one day ahead of the first calendar (on average if they are leap week calendars).

 

Now we’d like to know which one of these two calendars is the more accurate. Any method based only on comparing year lengths will give exactly the same accuracy to both calendars and so will not indicate which one of the two calendars is more accurate.

 

I made another criticism of Michael’s definition of year-round accuracy and it is that it will not indicate the year-round accuracy of a calendar of the type Walter Ziobro  is suggesting and not show that such a calendar has better year-round accuracy.

 

Karl

 

16(06(06

 

From: East Carolina University Calendar discussion List [mailto:[hidden email]] On Behalf Of Michael Ossipoff
Sent: 02 February 2017 00:13
To: [hidden email]
Subject: Re: Year-Round Accuracy RE: Displacement Calculation & RE: Jitter Calculation RE: ....

 

...meaning that, in the year 3017, the MTY will very slightly shorter in terms of mean solar days, because each mean solar day will be about 1/80 of a second longer than a mean solar day is now.

Michael Ossipoff

 

On Wed, Feb 1, 2017 at 6:54 PM, Michael Ossipoff <[hidden email]> wrote:

Oops! Because of the mean solar day being 1/80 of a second longer, then a year, of whatever kind, will be very slightly shorter, not longer, as a result.

So, my question should be re-worded:

Is Karl saying that the shortening of the MTY, in terms of mean solar days (as a result of the 1/80 of a second longer mean solar day), and the resulting small reduction in Y,  will cause some significant "inaccuracy" for the calendar of 3017 A.D.?

Michael Ossipoff

 

Michael Ossipoff

 

On Wed, Feb 1, 2017 at 6:35 PM, Michael Ossipoff <[hidden email]> wrote:

I said:


"Is Karl saying that a need to change Y by 1/5 of a second, from 365.24217 to 365.24222, is going to cause some kind of big accuracy problem for the calendar of 3017 A.D.?"

I meant:

Is Karl saying that a need to change Y by 5 seconds, from 365.24217 days to 365.24222 days, is going to cause some kind of big accuracy problem for the calendar of 3017 A.D.?

 

On Wed, Feb 1, 2017 at 6:23 PM, Michael Ossipoff <[hidden email]> wrote:

(Instead of deleting or moving the intervening text, it's easier to just copy & paste Karl's new text at the top of this post, and leaving the previously-intervening text at the bottom of the post.)

 

KARL REPLIES: The problem is that if one defines year-round accuracy (or any other kind of accuracy) by only comparing year lengths (such as calendar mean year (Y) and tropical year lengths), one is not going see what effect any change in the D is going to have on the accuracy.

 

...but it will show what effect it has on the D1 displacement, which is what Karl advocates as the measure of calendar-displacement. And, in a recent post, I told how a mismatch between reference-year and an SEL's tropical year would affect the difference between that SEL's date of occurrence and its DD.

 

Karl wants to complicate the problem by bringing in a kind of "accuracy" different from his advocated D1 measure of displacement. That's fair, because I invoked accuracy for particular SELs, when I spoke of a reference-year's mean accuracy around the ecliptic.

 

I told how a mismatch between the lengths of the reference-year and a particular SEL's tropical-year would affect accuracy as measured by the difference between that SEL's DD and its actual date of occurrence. I said it in a recent post, and I shouldn't have to repeat it.

 

 

Karl said:

 

 I don’t understand what Michael means by “the calendar’s centre of oscillation”

 

[endquote

 

When something moves back and forth, or in some way changes back and forth, we sometimes say that it "oscillates".  The middle of its range of oscillation can be called its "center of oscillation".

 

I apologize for using a difficult word.

 

Karl should feel free to tell me if there are other words that he has trouble with.

 

To continue:

 

 Karl says:

 

A measure of accuracy that only compares year lengths cannot show any such improvement of accuracy.

 

[endquote]

 

A departure of a reference-year's length, from the length of the tropical year measured at a particular SEL, will result in a steady unidirectional drift of the mean calendar-date for that SEL, away from that SEL's DD.

 

The closer the reference-year's length is, to the length of the tropical-year measured at that SEL, then the lower will be that drift.

 

Karl says:

 

If one uses desired dates DDs for measuring accuracy

 

[emdqipte]

 

I don't advocate that measure of accuracy or displacement. I use the D1 measure of displacement when I speak of the accuracy of the Minimum-Displacement Calendar, and so, for fair comparison, I use it for other calendars too, especially when comparing them to the Minimum-Displaement Calendar.

 

Yes, I used the D2 measure of displacement for the Gregorian jitter-range, and the Gregorian's maximum displacement in a 400-year cycle.  But I did that because I wanted to state those values by both displacement-measures.

 

Karl says:

 

, one can choose the DDs to be exactly the same every year even after Y changes. T

 

[endquote]

 

Let's be clear about how much change we're talking about for Y:

 

I'll find my print-out of Irv's graphs of 4 tropical-year-length  in mean solar days. But, because I don't have it here right now, I ask, how fast does the length of the MTY change? 

 

Shall I make a rough guess?

 

Because the MTY is the mean of the lengths of all the tropical years, measured at all of the points of the eclitpic, then presumably there's some degree to which the movements of the graphs of the lengths of those 4 tropical years--with respect to the steady year-length change-rate due to the lengthening of the day--will cancel eachother out.  So, a reasonable first-guess would be to say that the Length of the MTY varies at a rate corresponding to the general common overall downslope of the graphs.

 

It's said that the mean solar day, today, is about 1/400 of a second longer than it was in 1820.

 

An increase of 1/400 of second, in 200 years. If that's typical, that amounts to change n day-length of a second every 80,000 years.  In 1000 years, that would be 1/80 of a second, in the length of a mean solar day.

 

That would amount to about 1/5 of a second change in the length of a year, in 1000 years.

 

Karl specifically referred to the needed change in Y after 1000 years, so he's referring to a change of about 5 seconds in the length of the year.

 

Admittedly that's only a guess, but it's probably reasonably close.

 

Is Karl saying that a need to change Y by 1/5 of a second, from 365.24217 to 365.24222, is going to cause some kind of big accuracy problem for the calendar of 3017 A.D.?

 

If so, then Karl needs to be a lot more specific about the nature of that problem, and exactly, in detail, how it comes about. A numerical calculation of some error resulting from that problem, whatever its nature, would be much more convincing than a mere alleging of a problem.

 

Michael Ossipoff

 

 

 

 

On Wed, Feb 1, 2017 at 11:21 AM, Karl Palmen <[hidden email]> wrote:

Dear Michael and Calendar People

 

Here Michael addresses the problem I found with his definition of year-round accuracy, which involves only the comparison of year lengths.

 

From: East Carolina University Calendar discussion List [mailto:[hidden email]] On Behalf Of Michael Ossipoff
Sent: 31 January 2017 17:34
To: [hidden email]
Subject: Re: Displacement Calculation & RE: Jitter Calculation RE: ....

 

Here's the part that I said I'd reply to later, and I'll reply to it here:


This definition would be OK for one arithmetic calendar with one mean year, but ceases to be OK for a series of arithmetic calendars with different mean years as proposed by Michael. This is because when one changes from one calendar to the next a one-off change of displacement is made. 

Changing the value of Y, as might be done in about a millennium, or a few millennia, doesn't change D. It changes only Y.

But, as, over time, the Y value in use becomes less accurate, due to gradual change in the length of the chosen reference-year (such as the MTY), then that results in a cumulative error in D.

So, when the calendar, after a millennium or a few millennia, accumulates an unacceptable (.4 days?) amount of error (as determined by astronomical observations), then, when changing Y to the correct current value of the reference-year, the value of D could, likewise, be reset to correct the accumulated error resulting from the incorrect Y values.

Maybe you (Karl) have another way you'd like to do the longterm corrections. Fine. You have your way, I have my way, and the people 1000 years, or several thousand years, from now will no doubt have their way. If you think they need your way, and if you think that they won't come up with it on their own, then you need to give them good instructions, and store it in a time-capsule. But you'll have to be a lot more specific and clear about it than you have been.

If you're going to say that there's a problem, then you need to be a lot more specific and clear about what you think the problem is.




We discussed this last year. These changes, if not chosen carefully, can lead to a calendar with accurate mean year lengths to wander with respect the reference year, because drifts are not corrected properly.

That's why I suggested changing Y & D when the astronomical observations indicate that the calendar's center of oscillation is off by .4 day (or whatever other amount is considered unacceptable).

 

KARL REPLIES: The problem is that if one defines year-round accuracy (or any other kind of accuracy) by only comparing year lengths (such as calendar mean year (Y) and tropical year lengths), one is not going see what effect any change in the D is going to have on the accuracy.

 

I don’t understand what Michael means by “the calendar’s centre of oscillation” , but I don’t care what method is used to decide on the change of D, provided it improves the accuracy of the calendar. A measure of accuracy that only compares year lengths cannot show any such improvement of accuracy.

 

If one uses desired dates DDs for measuring accuracy, one can choose the DDs to be exactly the same every year even after Y changes. There must be no substantial change in DDs over a period of time that accuracy is measured, because such a change would allow cheating on accuracy. Note that the DDs are being used only for measuring accuracy and do not form part of the leap year rules (or first day of year rules) of any calendar.

 

Karl

 

16(06(05

 

 

 

 

 

Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: Year-Round Accuracy RE: Displacement Calculation & RE: Jitter Calculation RE: ....

Michael Ossipoff


On Thu, Feb 2, 2017 at 11:24 AM, Karl Palmen <[hidden email]> wrote:

Dear Michael and Calendar People

 

It is obvious to me that Michael has not understood what I’ve been attempting to express at all.


We definitely agree on that.

 

 So I’ll make one more attempt with a simple example possibly exaggerated for clarity.


Ok.
 

 

Consider two calendars whose calendar mean years (Y) are exactly equal every year. The Y values are changed at exactly the same year to exactly the same new value. The only difference is that the first calendar never changes the displacement (D) whenever Y is changed and the second calendar always reduces the displacement by 0.2 day, whenever Y is changed. Then after five Y changes the second calendar will be running one day ahead of the first calendar (on average if they are leap week calendars).

 

Now we’d like to know which one of these two calendars is the more accurate.


Between the years when Y is changed, Y has been kept constant, while the reference-year that Y represents has been continually changing. The annual changes in D have been calculated by using that un-updated value of Y. So, at the end of that period, when Y is updated, D should be corrected too.

One of your 2 calendars doesn't do that correction of D..That calendar is leaving an error that gets bigger and bigger.

You other calendar has been reducing D by 0.2 day whenever Y is changed. If +0.2 is the amount by which astronomical observations say that D is off, at that time, then that calendar is the one that is the more accurate one.

 

Any method based only on comparing year lengths will give exactly the same accuracy to both calendars and so will not indicate which one of the two calendars is more accurate.


Yes, it's necessary to correct D, when it gets too far off.

Of course. I never denied that it's necessary to correct D (in addition to correcting Y) when D differs too much from what it would be if the actual reference year, instead of the constant Y, were used for each annual change in D,

At that time, D should be changed to the value that it would have if the actual value of the reference year, instead of the constant Y, were used when calculating each annual change in Y.

But, nevertheless, for any SEL, the calendar will be more accurate, in the way that I've been repeating, if the the length of the reference year is closer to the length of the tropical year measured at that SEL.

Could it be that I still don't know what you're talking about?

Yes, that's entirely possible!
 

 

I made another criticism of Michael’s definition of year-round accuracy and it is that it will not indicate the year-round accuracy of a calendar of the type Walter Ziobro  is suggesting and not show that such a calendar has better year-round accuracy.


When I spoke of "year-round accuracy", I merely meant the average closeness of the reference-year's length to the lengths of the tropical years measured at the various SELs.

As for the accuracy-benefit of that, I'll repeat something that I said above:

"But, nevertheless, for any SEL, the calendar will be more accurate, in the way that I've been repeating, if the the length of the reference year is closer to the length of the tropical year measured at that SEL."

I've repeated several times, the way in which that calendar is more accurate for that SEL, if its reference-year is close to the tropical year measured at that SEL. I'm not going to repeat that again.

I never claimed that that definition indicates the accuracy of a type of calendar that I don't know about. There are many kinds of calendars. I spoke about the choice of Y only for the Minimum-Displacement Calendar.

If we have different kinds of cars, their parts might not be interchangeable. The constant "Y" was intended solely for the Minimum-Displacement Calendar. Other calendars will have different rules, and that's as it should be.

Michael Ossipoff



 

 

Karl

 

16(06(06

 

From: East Carolina University Calendar discussion List [mailto:[hidden email]] On Behalf Of Michael Ossipoff
Sent: 02 February 2017 00:13
To: [hidden email]
Subject: Re: Year-Round Accuracy RE: Displacement Calculation & RE: Jitter Calculation RE: ....

 

...meaning that, in the year 3017, the MTY will very slightly shorter in terms of mean solar days, because each mean solar day will be about 1/80 of a second longer than a mean solar day is now.

Michael Ossipoff

 

On Wed, Feb 1, 2017 at 6:54 PM, Michael Ossipoff <[hidden email]> wrote:

Oops! Because of the mean solar day being 1/80 of a second longer, then a year, of whatever kind, will be very slightly shorter, not longer, as a result.

So, my question should be re-worded:

Is Karl saying that the shortening of the MTY, in terms of mean solar days (as a result of the 1/80 of a second longer mean solar day), and the resulting small reduction in Y,  will cause some significant "inaccuracy" for the calendar of 3017 A.D.?

Michael Ossipoff

 

Michael Ossipoff

 

On Wed, Feb 1, 2017 at 6:35 PM, Michael Ossipoff <[hidden email]> wrote:

I said:


"Is Karl saying that a need to change Y by 1/5 of a second, from 365.24217 to 365.24222, is going to cause some kind of big accuracy problem for the calendar of 3017 A.D.?"

I meant:

Is Karl saying that a need to change Y by 5 seconds, from 365.24217 days to 365.24222 days, is going to cause some kind of big accuracy problem for the calendar of 3017 A.D.?

 

On Wed, Feb 1, 2017 at 6:23 PM, Michael Ossipoff <[hidden email]> wrote:

(Instead of deleting or moving the intervening text, it's easier to just copy & paste Karl's new text at the top of this post, and leaving the previously-intervening text at the bottom of the post.)

 

KARL REPLIES: The problem is that if one defines year-round accuracy (or any other kind of accuracy) by only comparing year lengths (such as calendar mean year (Y) and tropical year lengths), one is not going see what effect any change in the D is going to have on the accuracy.

 

...but it will show what effect it has on the D1 displacement, which is what Karl advocates as the measure of calendar-displacement. And, in a recent post, I told how a mismatch between reference-year and an SEL's tropical year would affect the difference between that SEL's date of occurrence and its DD.

 

Karl wants to complicate the problem by bringing in a kind of "accuracy" different from his advocated D1 measure of displacement. That's fair, because I invoked accuracy for particular SELs, when I spoke of a reference-year's mean accuracy around the ecliptic.

 

I told how a mismatch between the lengths of the reference-year and a particular SEL's tropical-year would affect accuracy as measured by the difference between that SEL's DD and its actual date of occurrence. I said it in a recent post, and I shouldn't have to repeat it.

 

 

Karl said:

 

 I don’t understand what Michael means by “the calendar’s centre of oscillation”

 

[endquote

 

When something moves back and forth, or in some way changes back and forth, we sometimes say that it "oscillates".  The middle of its range of oscillation can be called its "center of oscillation".

 

I apologize for using a difficult word.

 

Karl should feel free to tell me if there are other words that he has trouble with.

 

To continue:

 

 Karl says:

 

A measure of accuracy that only compares year lengths cannot show any such improvement of accuracy.

 

[endquote]

 

A departure of a reference-year's length, from the length of the tropical year measured at a particular SEL, will result in a steady unidirectional drift of the mean calendar-date for that SEL, away from that SEL's DD.

 

The closer the reference-year's length is, to the length of the tropical-year measured at that SEL, then the lower will be that drift.

 

Karl says:

 

If one uses desired dates DDs for measuring accuracy

 

[emdqipte]

 

I don't advocate that measure of accuracy or displacement. I use the D1 measure of displacement when I speak of the accuracy of the Minimum-Displacement Calendar, and so, for fair comparison, I use it for other calendars too, especially when comparing them to the Minimum-Displaement Calendar.

 

Yes, I used the D2 measure of displacement for the Gregorian jitter-range, and the Gregorian's maximum displacement in a 400-year cycle.  But I did that because I wanted to state those values by both displacement-measures.

 

Karl says:

 

, one can choose the DDs to be exactly the same every year even after Y changes. T

 

[endquote]

 

Let's be clear about how much change we're talking about for Y:

 

I'll find my print-out of Irv's graphs of 4 tropical-year-length  in mean solar days. But, because I don't have it here right now, I ask, how fast does the length of the MTY change? 

 

Shall I make a rough guess?

 

Because the MTY is the mean of the lengths of all the tropical years, measured at all of the points of the eclitpic, then presumably there's some degree to which the movements of the graphs of the lengths of those 4 tropical years--with respect to the steady year-length change-rate due to the lengthening of the day--will cancel eachother out.  So, a reasonable first-guess would be to say that the Length of the MTY varies at a rate corresponding to the general common overall downslope of the graphs.

 

It's said that the mean solar day, today, is about 1/400 of a second longer than it was in 1820.

 

An increase of 1/400 of second, in 200 years. If that's typical, that amounts to change n day-length of a second every 80,000 years.  In 1000 years, that would be 1/80 of a second, in the length of a mean solar day.

 

That would amount to about 1/5 of a second change in the length of a year, in 1000 years.

 

Karl specifically referred to the needed change in Y after 1000 years, so he's referring to a change of about 5 seconds in the length of the year.

 

Admittedly that's only a guess, but it's probably reasonably close.

 

Is Karl saying that a need to change Y by 1/5 of a second, from 365.24217 to 365.24222, is going to cause some kind of big accuracy problem for the calendar of 3017 A.D.?

 

If so, then Karl needs to be a lot more specific about the nature of that problem, and exactly, in detail, how it comes about. A numerical calculation of some error resulting from that problem, whatever its nature, would be much more convincing than a mere alleging of a problem.

 

Michael Ossipoff

 

 

 

 

On Wed, Feb 1, 2017 at 11:21 AM, Karl Palmen <[hidden email]> wrote:

Dear Michael and Calendar People

 

Here Michael addresses the problem I found with his definition of year-round accuracy, which involves only the comparison of year lengths.

 

From: East Carolina University Calendar discussion List [mailto:[hidden email]] On Behalf Of Michael Ossipoff
Sent: 31 January 2017 17:34
To: [hidden email]
Subject: Re: Displacement Calculation & RE: Jitter Calculation RE: ....

 

Here's the part that I said I'd reply to later, and I'll reply to it here:


This definition would be OK for one arithmetic calendar with one mean year, but ceases to be OK for a series of arithmetic calendars with different mean years as proposed by Michael. This is because when one changes from one calendar to the next a one-off change of displacement is made. 

Changing the value of Y, as might be done in about a millennium, or a few millennia, doesn't change D. It changes only Y.

But, as, over time, the Y value in use becomes less accurate, due to gradual change in the length of the chosen reference-year (such as the MTY), then that results in a cumulative error in D.

So, when the calendar, after a millennium or a few millennia, accumulates an unacceptable (.4 days?) amount of error (as determined by astronomical observations), then, when changing Y to the correct current value of the reference-year, the value of D could, likewise, be reset to correct the accumulated error resulting from the incorrect Y values.

Maybe you (Karl) have another way you'd like to do the longterm corrections. Fine. You have your way, I have my way, and the people 1000 years, or several thousand years, from now will no doubt have their way. If you think they need your way, and if you think that they won't come up with it on their own, then you need to give them good instructions, and store it in a time-capsule. But you'll have to be a lot more specific and clear about it than you have been.

If you're going to say that there's a problem, then you need to be a lot more specific and clear about what you think the problem is.




We discussed this last year. These changes, if not chosen carefully, can lead to a calendar with accurate mean year lengths to wander with respect the reference year, because drifts are not corrected properly.

That's why I suggested changing Y & D when the astronomical observations indicate that the calendar's center of oscillation is off by .4 day (or whatever other amount is considered unacceptable).

 

KARL REPLIES: The problem is that if one defines year-round accuracy (or any other kind of accuracy) by only comparing year lengths (such as calendar mean year (Y) and tropical year lengths), one is not going see what effect any change in the D is going to have on the accuracy.

 

I don’t understand what Michael means by “the calendar’s centre of oscillation” , but I don’t care what method is used to decide on the change of D, provided it improves the accuracy of the calendar. A measure of accuracy that only compares year lengths cannot show any such improvement of accuracy.

 

If one uses desired dates DDs for measuring accuracy, one can choose the DDs to be exactly the same every year even after Y changes. There must be no substantial change in DDs over a period of time that accuracy is measured, because such a change would allow cheating on accuracy. Note that the DDs are being used only for measuring accuracy and do not form part of the leap year rules (or first day of year rules) of any calendar.

 

Karl

 

16(06(05

 

 

 

 

 


Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: Year-Round Accuracy RE: Displacement Calculation & RE: Jitter Calculation RE: ....

Karl Palmen
In reply to this post by Karl Palmen

Dear Michael and Calendar People

 

Michael is resolving this criticism by restricting the types of calendars that his measure of year-round accuracy applies to. This restriction is to a type of minimum displacement calendar, which Michael describes below.

 

Then the criticism of Michael’s definition of year-round accuracy is that it applies only to a certain type of minimum-displacement calendar and so making comparison with other calendars impossible, such as the type of calendar Walter Ziobro has been suggesting or a minimum displacement calendar that changes Y & D in a different manner than described below.

 

Because of this restriction, it does not deserve to be called “year-round accuracy”, but perhaps “year-round mean year accuracy”, which describes it well.

 

Karl

 

16(06(06

 

From: East Carolina University Calendar discussion List [mailto:[hidden email]] On Behalf Of Michael Ossipoff
Sent: 02 February 2017 18:24
To: [hidden email]
Subject: Re: Year-Round Accuracy RE: Displacement Calculation & RE: Jitter Calculation RE: ....

 

 

 

On Thu, Feb 2, 2017 at 11:24 AM, Karl Palmen <[hidden email]> wrote:

Dear Michael and Calendar People

 

It is obvious to me that Michael has not understood what I’ve been attempting to express at all.

 

We definitely agree on that.

 

 So I’ll make one more attempt with a simple example possibly exaggerated for clarity.

 

Ok.
 

 

Consider two calendars whose calendar mean years (Y) are exactly equal every year. The Y values are changed at exactly the same year to exactly the same new value. The only difference is that the first calendar never changes the displacement (D) whenever Y is changed and the second calendar always reduces the displacement by 0.2 day, whenever Y is changed. Then after five Y changes the second calendar will be running one day ahead of the first calendar (on average if they are leap week calendars).

 

Now we’d like to know which one of these two calendars is the more accurate.

 

Between the years when Y is changed, Y has been kept constant, while the reference-year that Y represents has been continually changing. The annual changes in D have been calculated by using that un-updated value of Y. So, at the end of that period, when Y is updated, D should be corrected too.

One of your 2 calendars doesn't do that correction of D..That calendar is leaving an error that gets bigger and bigger.

You other calendar has been reducing D by 0.2 day whenever Y is changed. If +0.2 is the amount by which astronomical observations say that D is off, at that time, then that calendar is the one that is the more accurate one.

 

Any method based only on comparing year lengths will give exactly the same accuracy to both calendars and so will not indicate which one of the two calendars is more accurate.

 

Yes, it's necessary to correct D, when it gets too far off.

 

Of course. I never denied that it's necessary to correct D (in addition to correcting Y) when D differs too much from what it would be if the actual reference year, instead of the constant Y, were used for each annual change in D,

At that time, D should be changed to the value that it would have if the actual value of the reference year, instead of the constant Y, were used when calculating each annual change in Y.

But, nevertheless, for any SEL, the calendar will be more accurate, in the way that I've been repeating, if the the length of the reference year is closer to the length of the tropical year measured at that SEL.

Could it be that I still don't know what you're talking about?

Yes, that's entirely possible!
 

 

I made another criticism of Michael’s definition of year-round accuracy and it is that it will not indicate the year-round accuracy of a calendar of the type Walter Ziobro  is suggesting and not show that such a calendar has better year-round accuracy.

 

When I spoke of "year-round accuracy", I merely meant the average closeness of the reference-year's length to the lengths of the tropical years measured at the various SELs.

As for the accuracy-benefit of that, I'll repeat something that I said above:

"But, nevertheless, for any SEL, the calendar will be more accurate, in the way that I've been repeating, if the the length of the reference year is closer to the length of the tropical year measured at that SEL."

I've repeated several times, the way in which that calendar is more accurate for that SEL, if its reference-year is close to the tropical year measured at that SEL. I'm not going to repeat that again.

 

I never claimed that that definition indicates the accuracy of a type of calendar that I don't know about. There are many kinds of calendars. I spoke about the choice of Y only for the Minimum-Displacement Calendar.

If we have different kinds of cars, their parts might not be interchangeable. The constant "Y" was intended solely for the Minimum-Displacement Calendar. Other calendars will have different rules, and that's as it should be.

Michael Ossipoff




 

Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: Year-Round Accuracy RE: Displacement Calculation & RE: Jitter Calculation RE: ....

Michael Ossipoff


On Fri, Feb 3, 2017 at 7:49 AM, Karl Palmen <[hidden email]> wrote:

Dear Michael and Calendar People

 

Michael is resolving this criticism by restricting the types of calendars that his measure of year-round accuracy applies to.



Not only that, but "year-round accuracy" didn't refer to the accuracy of calendars at all !

(Dick Hanke likes to use lots of exclamation-points. I'm going to borrow that practice from him!)

I've explained to you, several times, what I meant by "year-round accuracy". I'll say this just one more time. After this, I won't repeat it for you again:

When I spoke of the "year-round accuracy" of a value for Y (but not for a calendar), I was referring to its overall closeness to the lengths of the various tropical years, measured at the various points of the ecliptic. Period (full stop).

I was using "year-round accuracy", with the meaning stated above, in reference to choices of the value of the variable "Y", for use in the Minimum-Displacement Calendar.

The use of the expression "year-round accuracy" for that purpose, led to Karl's misunderstanding. His misunderstanding initially, therefore, is understandable and justifiable. His continual expression of the same misunderstanding is no longer justifiable.


 

This restriction is to a type of minimum displacement calendar, which Michael describes below.


It's a restriction to reference to the choice of Y-values for the Minimum-Displacement Calendar, and how well that Y-value matches the lengths of the tropical years defined with respect to the various points of the ecliptic.

 

 

Then the criticism of Michael’s definition of year-round accuracy is that it applies only to a certain type of minimum-displacement calendar


It's even better than that: It applies only to the choice of the value of the variable "Y" for the Minimum-Displacement Calendar. It refers to the accuracy of Y, in terms of how well it matches, overall, the lengths of the tropical years defined with respect to the various points of the ecliptic.

Forgive me for evaluating choices of Y for the Minimum-Displacement Calendar :^)


 

and so making comparison with other calendars impossible



So Karl thinks that comparison of Minimum-Displacement to other calendars is made impossible by the fact that I evaluate choices of values for the value of Y for the Minimum-Displacement Calendar.

:^)

I've already made it abundantly clear that my standard for choosing Y was not intended for comparing the Minimum-Displacement Calendar to other calendars.

If Karl wants to compare the accuracy of Minimum-Displacement to that of other calendars, then there are measures for that.

Karl said that he prefers the measure of displacement that I've called "D1". It's displacement that's constant throughout a displacement-year, and which changes abruptly at the end of a displacement-year.

A good measure of a calendar's accuracy is the maximum amount by which it is displaced, between its periodic displacement-extremes, with respect to ecliptic longitudes (as represented by the progress of some chosen mean-year--I use the MTY).

We've discussed two measures of displacement, which I've called (for brevity) D1 & D2. I've re-stated, above in this post, what I mean by D1.

Because D1 is the measure of displaceent that Karl said that he prefers, and because it's the one by which I evaluate Minimulm-Displacment's accuracy, and because it matches the mechanism of Minimum-Displacement, then D1 would be a fair choice for comparing Minimum-Displacement to other calendars.

By D1, Minimlum-Displacement's displacement-range is 7 days.

Karl: Use that to compare The Minimum-Displacement Calendar's accuracy to that of other calendars.

(In fairness, only compare it to other leapweek calendars, because leapweek calendars have more maximum periodic displacement than leapday calendars do.)




 

, such as the type of calendar Walter Ziobro has been suggesting or a minimum displacement calendar that changes Y & D in a different manner than described below.


If you want to compare the accuracy of the Minimum-Dispacement Calendar to the accuracy of Walter Ziobro's calendar, then find out the maximum periodic displacement-range of Walter's calendar. If it's greater than 7, then it's less accurate than Minimum-Displacement. If it's less than 7, then it's more accurate than Minimum-Displacment.


 

 

Because of this restriction, it does not deserve to be called “year-round accuracy”


As I said, I've quit using that name for that (rough) standard for choosing Y for the Minimum-Displacement Calendar.



 

, but perhaps “year-round mean year accuracy”, which describes it well.


Call it what you want, but I've made it clear from the start that it wasn't intended for comparing the accuracies of different calendars.

Michael Ossipoff

 

Karl

 

16(06(06

 

From: East Carolina University Calendar discussion List [mailto:[hidden email]] On Behalf Of Michael Ossipoff
Sent: 02 February 2017 18:24
To: [hidden email]
Subject: Re: Year-Round Accuracy RE: Displacement Calculation & RE: Jitter Calculation RE: ....

 

 

 

On Thu, Feb 2, 2017 at 11:24 AM, Karl Palmen <[hidden email]> wrote:

Dear Michael and Calendar People

 

It is obvious to me that Michael has not understood what I’ve been attempting to express at all.

 

We definitely agree on that.

 

 So I’ll make one more attempt with a simple example possibly exaggerated for clarity.

 

Ok.
 

 

Consider two calendars whose calendar mean years (Y) are exactly equal every year. The Y values are changed at exactly the same year to exactly the same new value. The only difference is that the first calendar never changes the displacement (D) whenever Y is changed and the second calendar always reduces the displacement by 0.2 day, whenever Y is changed. Then after five Y changes the second calendar will be running one day ahead of the first calendar (on average if they are leap week calendars).

 

Now we’d like to know which one of these two calendars is the more accurate.

 

Between the years when Y is changed, Y has been kept constant, while the reference-year that Y represents has been continually changing. The annual changes in D have been calculated by using that un-updated value of Y. So, at the end of that period, when Y is updated, D should be corrected too.

One of your 2 calendars doesn't do that correction of D..That calendar is leaving an error that gets bigger and bigger.

You other calendar has been reducing D by 0.2 day whenever Y is changed. If +0.2 is the amount by which astronomical observations say that D is off, at that time, then that calendar is the one that is the more accurate one.

 

Any method based only on comparing year lengths will give exactly the same accuracy to both calendars and so will not indicate which one of the two calendars is more accurate.

 

Yes, it's necessary to correct D, when it gets too far off.

 

Of course. I never denied that it's necessary to correct D (in addition to correcting Y) when D differs too much from what it would be if the actual reference year, instead of the constant Y, were used for each annual change in D,

At that time, D should be changed to the value that it would have if the actual value of the reference year, instead of the constant Y, were used when calculating each annual change in Y.

But, nevertheless, for any SEL, the calendar will be more accurate, in the way that I've been repeating, if the the length of the reference year is closer to the length of the tropical year measured at that SEL.

Could it be that I still don't know what you're talking about?

Yes, that's entirely possible!
 

 

I made another criticism of Michael’s definition of year-round accuracy and it is that it will not indicate the year-round accuracy of a calendar of the type Walter Ziobro  is suggesting and not show that such a calendar has better year-round accuracy.

 

When I spoke of "year-round accuracy", I merely meant the average closeness of the reference-year's length to the lengths of the tropical years measured at the various SELs.

As for the accuracy-benefit of that, I'll repeat something that I said above:

"But, nevertheless, for any SEL, the calendar will be more accurate, in the way that I've been repeating, if the the length of the reference year is closer to the length of the tropical year measured at that SEL."

I've repeated several times, the way in which that calendar is more accurate for that SEL, if its reference-year is close to the tropical year measured at that SEL. I'm not going to repeat that again.

 

I never claimed that that definition indicates the accuracy of a type of calendar that I don't know about. There are many kinds of calendars. I spoke about the choice of Y only for the Minimum-Displacement Calendar.

If we have different kinds of cars, their parts might not be interchangeable. The constant "Y" was intended solely for the Minimum-Displacement Calendar. Other calendars will have different rules, and that's as it should be.

Michael Ossipoff




 


Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: Year-Round Accuracy RE: Displacement Calculation & RE: Jitter Calculation RE: ....

Michael Ossipoff

Oops! I said:

"(Dick Hanke likes to use lots of exclamation-points. I'm going to borrow that practice from him!)"

I meant Dick Henry!.

Michael Ossipoff
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: Year-Round Accuracy RE: Displacement Calculation & RE: Jitter Calculation RE: ....

Michael Ossipoff
In reply to this post by Michael Ossipoff
I've just noticed an error that might be the cause of at least some of Karl's objections, regarding my statements about the choice of the Y value for Minimum-Displacement.

At least in some instances, I spoke of a drift of the "center of oscillation" of the date of a SEL's calendar-date.

I was referring to the oscillation of a SEL's date, over a leapyear-period. (the duration between leapyears)

So here's what I mean, and would have said, were not for that error:

For the short-term, during which the length of a tropical year (for an individual SEL, or the MTA) doesn't change:

The date of an SEL will always oscillate in the calendar, over every leapyear period, from one leapyear to the next. The "center of oscillation" of the date of that SEL is the midpoint of the extremes of its dates, over a leapyear-period.

If Y differs from the length of the tropical year defined for a particular SEL, and to the extent that it does, then the center of oscillation of the date of that SEL will drift in the calendar. .Of course a choice of Y is  going to favor some SEL, by reducing or eliminating the drift of the center of oscillation of its calendar-date.  ...or else Y can be chosen so that, by some measure of central tendency, Y is reasonably close to all of the SELs' tropical-year lengths, and doesn't favor one SEL over another.

I choose the latter. That's why I use the length of the MTY as Y.

As the length of the reference-year varies, over the millennia:

If Y has been chosen as the length of the June solstice tropical year, to avoid calendar-date drift of the center of oscillation of the calendar-date of the June solstice, then, because the length of the June solstice year changes very little over the next 9000 years or so, then Y will remain relatively close to the June solstice tropical year, and the center of oscillation of the date of the June solstice will remain fairly close to its initial value (the value that it has now, just before the adoption of the new calendar).

But, with that choice of Y, of course, even initially (before the length of the June solstice year changes any), of course Y's initial difference from the tropical-year lengths for the other SELs will be large, and so the center of oscillation of the calendar-date for some SEL other than the June solstice will have a high drift-rate, even initially.

Of course, as the length of the reference-year changes, over the millennia. the difference between Y, and the tropical-year lengths of the various SELs will change too--positively for some SELs, and negatively for ohers--and therefore so will the drift rates of the centers of oscillation of the dates for those SELs.

I choose the current length of the MTY as the initial value for Y, to, in some way try to reduce, over the whole eclliptic, the drift rate of the center of oscillation of the calenndar-date of any particular SEL.

I realize that calendar-date consistency for a particular SEL isn't what the D1 measure of calendar-displacement is about (that's the calendar-displacement measure that we agree on preferring). But I still prefer a Y-value that, in some sense, over the whole ecliptic, tries to minimize the drift-rate for the center of oscillation of the calendar-date of any particular SEL.

I emphasize that this isn't intended or offered as a way to compare the accuracy of two calendars.

Michael Ossipoff


On Fri, Feb 3, 2017 at 4:07 PM, Michael Ossipoff <[hidden email]> wrote:


On Fri, Feb 3, 2017 at 7:49 AM, Karl Palmen <[hidden email]> wrote:

Dear Michael and Calendar People

 

Michael is resolving this criticism by restricting the types of calendars that his measure of year-round accuracy applies to.



Not only that, but "year-round accuracy" didn't refer to the accuracy of calendars at all !

(Dick Hanke likes to use lots of exclamation-points. I'm going to borrow that practice from him!)

I've explained to you, several times, what I meant by "year-round accuracy". I'll say this just one more time. After this, I won't repeat it for you again:

When I spoke of the "year-round accuracy" of a value for Y (but not for a calendar), I was referring to its overall closeness to the lengths of the various tropical years, measured at the various points of the ecliptic. Period (full stop).

I was using "year-round accuracy", with the meaning stated above, in reference to choices of the value of the variable "Y", for use in the Minimum-Displacement Calendar.

The use of the expression "year-round accuracy" for that purpose, led to Karl's misunderstanding. His misunderstanding initially, therefore, is understandable and justifiable. His continual expression of the same misunderstanding is no longer justifiable.


 

This restriction is to a type of minimum displacement calendar, which Michael describes below.


It's a restriction to reference to the choice of Y-values for the Minimum-Displacement Calendar, and how well that Y-value matches the lengths of the tropical years defined with respect to the various points of the ecliptic.

 

 

Then the criticism of Michael’s definition of year-round accuracy is that it applies only to a certain type of minimum-displacement calendar


It's even better than that: It applies only to the choice of the value of the variable "Y" for the Minimum-Displacement Calendar. It refers to the accuracy of Y, in terms of how well it matches, overall, the lengths of the tropical years defined with respect to the various points of the ecliptic.

Forgive me for evaluating choices of Y for the Minimum-Displacement Calendar :^)


 

and so making comparison with other calendars impossible



So Karl thinks that comparison of Minimum-Displacement to other calendars is made impossible by the fact that I evaluate choices of values for the value of Y for the Minimum-Displacement Calendar.

:^)

I've already made it abundantly clear that my standard for choosing Y was not intended for comparing the Minimum-Displacement Calendar to other calendars.

If Karl wants to compare the accuracy of Minimum-Displacement to that of other calendars, then there are measures for that.

Karl said that he prefers the measure of displacement that I've called "D1". It's displacement that's constant throughout a displacement-year, and which changes abruptly at the end of a displacement-year.

A good measure of a calendar's accuracy is the maximum amount by which it is displaced, between its periodic displacement-extremes, with respect to ecliptic longitudes (as represented by the progress of some chosen mean-year--I use the MTY).

We've discussed two measures of displacement, which I've called (for brevity) D1 & D2. I've re-stated, above in this post, what I mean by D1.

Because D1 is the measure of displaceent that Karl said that he prefers, and because it's the one by which I evaluate Minimulm-Displacment's accuracy, and because it matches the mechanism of Minimum-Displacement, then D1 would be a fair choice for comparing Minimum-Displacement to other calendars.

By D1, Minimlum-Displacement's displacement-range is 7 days.

Karl: Use that to compare The Minimum-Displacement Calendar's accuracy to that of other calendars.

(In fairness, only compare it to other leapweek calendars, because leapweek calendars have more maximum periodic displacement than leapday calendars do.)




 

, such as the type of calendar Walter Ziobro has been suggesting or a minimum displacement calendar that changes Y & D in a different manner than described below.


If you want to compare the accuracy of the Minimum-Dispacement Calendar to the accuracy of Walter Ziobro's calendar, then find out the maximum periodic displacement-range of Walter's calendar. If it's greater than 7, then it's less accurate than Minimum-Displacement. If it's less than 7, then it's more accurate than Minimum-Displacment.


 

 

Because of this restriction, it does not deserve to be called “year-round accuracy”


As I said, I've quit using that name for that (rough) standard for choosing Y for the Minimum-Displacement Calendar.



 

, but perhaps “year-round mean year accuracy”, which describes it well.


Call it what you want, but I've made it clear from the start that it wasn't intended for comparing the accuracies of different calendars.

Michael Ossipoff

 

Karl

 

16(06(06

 

From: East Carolina University Calendar discussion List [mailto:[hidden email]] On Behalf Of Michael Ossipoff
Sent: 02 February 2017 18:24
To: [hidden email]
Subject: Re: Year-Round Accuracy RE: Displacement Calculation & RE: Jitter Calculation RE: ....

 

 

 

On Thu, Feb 2, 2017 at 11:24 AM, Karl Palmen <[hidden email]> wrote:

Dear Michael and Calendar People

 

It is obvious to me that Michael has not understood what I’ve been attempting to express at all.

 

We definitely agree on that.

 

 So I’ll make one more attempt with a simple example possibly exaggerated for clarity.

 

Ok.
 

 

Consider two calendars whose calendar mean years (Y) are exactly equal every year. The Y values are changed at exactly the same year to exactly the same new value. The only difference is that the first calendar never changes the displacement (D) whenever Y is changed and the second calendar always reduces the displacement by 0.2 day, whenever Y is changed. Then after five Y changes the second calendar will be running one day ahead of the first calendar (on average if they are leap week calendars).

 

Now we’d like to know which one of these two calendars is the more accurate.

 

Between the years when Y is changed, Y has been kept constant, while the reference-year that Y represents has been continually changing. The annual changes in D have been calculated by using that un-updated value of Y. So, at the end of that period, when Y is updated, D should be corrected too.

One of your 2 calendars doesn't do that correction of D..That calendar is leaving an error that gets bigger and bigger.

You other calendar has been reducing D by 0.2 day whenever Y is changed. If +0.2 is the amount by which astronomical observations say that D is off, at that time, then that calendar is the one that is the more accurate one.

 

Any method based only on comparing year lengths will give exactly the same accuracy to both calendars and so will not indicate which one of the two calendars is more accurate.

 

Yes, it's necessary to correct D, when it gets too far off.

 

Of course. I never denied that it's necessary to correct D (in addition to correcting Y) when D differs too much from what it would be if the actual reference year, instead of the constant Y, were used for each annual change in D,

At that time, D should be changed to the value that it would have if the actual value of the reference year, instead of the constant Y, were used when calculating each annual change in Y.

But, nevertheless, for any SEL, the calendar will be more accurate, in the way that I've been repeating, if the the length of the reference year is closer to the length of the tropical year measured at that SEL.

Could it be that I still don't know what you're talking about?

Yes, that's entirely possible!
 

 

I made another criticism of Michael’s definition of year-round accuracy and it is that it will not indicate the year-round accuracy of a calendar of the type Walter Ziobro  is suggesting and not show that such a calendar has better year-round accuracy.

 

When I spoke of "year-round accuracy", I merely meant the average closeness of the reference-year's length to the lengths of the tropical years measured at the various SELs.

As for the accuracy-benefit of that, I'll repeat something that I said above:

"But, nevertheless, for any SEL, the calendar will be more accurate, in the way that I've been repeating, if the the length of the reference year is closer to the length of the tropical year measured at that SEL."

I've repeated several times, the way in which that calendar is more accurate for that SEL, if its reference-year is close to the tropical year measured at that SEL. I'm not going to repeat that again.

 

I never claimed that that definition indicates the accuracy of a type of calendar that I don't know about. There are many kinds of calendars. I spoke about the choice of Y only for the Minimum-Displacement Calendar.

If we have different kinds of cars, their parts might not be interchangeable. The constant "Y" was intended solely for the Minimum-Displacement Calendar. Other calendars will have different rules, and that's as it should be.

Michael Ossipoff




 



Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: Year-Round Accuracy RE: Displacement Calculation & RE: Jitter Calculation RE: ....

Michael Ossipoff

I said:

"I've just noticed an error that might be the cause of at least some of Karl's objections, regarding my statements about the choice of the Y value for Minimum-Displacement."

The error I was referring to was:

I neglected to say exactly what I was referring to the oscillation of.

As I clarified in my previous post to this thread, I was referring to the oscillation of a SEL's calendar-date, from one leapyear to the next.

Michael Ossipoff

On Fri, Feb 3, 2017 at 11:11 PM, Michael Ossipoff <[hidden email]> wrote:
I've just noticed an error that might be the cause of at least some of Karl's objections, regarding my statements about the choice of the Y value for Minimum-Displacement.

At least in some instances, I spoke of a drift of the "center of oscillation" of the date of a SEL's calendar-date.

I was referring to the oscillation of a SEL's date, over a leapyear-period. (the duration between leapyears)

So here's what I mean, and would have said, were not for that error:

For the short-term, during which the length of a tropical year (for an individual SEL, or the MTA) doesn't change:

The date of an SEL will always oscillate in the calendar, over every leapyear period, from one leapyear to the next. The "center of oscillation" of the date of that SEL is the midpoint of the extremes of its dates, over a leapyear-period.

If Y differs from the length of the tropical year defined for a particular SEL, and to the extent that it does, then the center of oscillation of the date of that SEL will drift in the calendar. .Of course a choice of Y is  going to favor some SEL, by reducing or eliminating the drift of the center of oscillation of its calendar-date.  ...or else Y can be chosen so that, by some measure of central tendency, Y is reasonably close to all of the SELs' tropical-year lengths, and doesn't favor one SEL over another.

I choose the latter. That's why I use the length of the MTY as Y.

As the length of the reference-year varies, over the millennia:

If Y has been chosen as the length of the June solstice tropical year, to avoid calendar-date drift of the center of oscillation of the calendar-date of the June solstice, then, because the length of the June solstice year changes very little over the next 9000 years or so, then Y will remain relatively close to the June solstice tropical year, and the center of oscillation of the date of the June solstice will remain fairly close to its initial value (the value that it has now, just before the adoption of the new calendar).

But, with that choice of Y, of course, even initially (before the length of the June solstice year changes any), of course Y's initial difference from the tropical-year lengths for the other SELs will be large, and so the center of oscillation of the calendar-date for some SEL other than the June solstice will have a high drift-rate, even initially.

Of course, as the length of the reference-year changes, over the millennia. the difference between Y, and the tropical-year lengths of the various SELs will change too--positively for some SELs, and negatively for ohers--and therefore so will the drift rates of the centers of oscillation of the dates for those SELs.

I choose the current length of the MTY as the initial value for Y, to, in some way try to reduce, over the whole eclliptic, the drift rate of the center of oscillation of the calenndar-date of any particular SEL.

I realize that calendar-date consistency for a particular SEL isn't what the D1 measure of calendar-displacement is about (that's the calendar-displacement measure that we agree on preferring). But I still prefer a Y-value that, in some sense, over the whole ecliptic, tries to minimize the drift-rate for the center of oscillation of the calendar-date of any particular SEL.

I emphasize that this isn't intended or offered as a way to compare the accuracy of two calendars.

Michael Ossipoff


On Fri, Feb 3, 2017 at 4:07 PM, Michael Ossipoff <[hidden email]> wrote:


On Fri, Feb 3, 2017 at 7:49 AM, Karl Palmen <[hidden email]> wrote:

Dear Michael and Calendar People

 

Michael is resolving this criticism by restricting the types of calendars that his measure of year-round accuracy applies to.



Not only that, but "year-round accuracy" didn't refer to the accuracy of calendars at all !

(Dick Hanke likes to use lots of exclamation-points. I'm going to borrow that practice from him!)

I've explained to you, several times, what I meant by "year-round accuracy". I'll say this just one more time. After this, I won't repeat it for you again:

When I spoke of the "year-round accuracy" of a value for Y (but not for a calendar), I was referring to its overall closeness to the lengths of the various tropical years, measured at the various points of the ecliptic. Period (full stop).

I was using "year-round accuracy", with the meaning stated above, in reference to choices of the value of the variable "Y", for use in the Minimum-Displacement Calendar.

The use of the expression "year-round accuracy" for that purpose, led to Karl's misunderstanding. His misunderstanding initially, therefore, is understandable and justifiable. His continual expression of the same misunderstanding is no longer justifiable.


 

This restriction is to a type of minimum displacement calendar, which Michael describes below.


It's a restriction to reference to the choice of Y-values for the Minimum-Displacement Calendar, and how well that Y-value matches the lengths of the tropical years defined with respect to the various points of the ecliptic.

 

 

Then the criticism of Michael’s definition of year-round accuracy is that it applies only to a certain type of minimum-displacement calendar


It's even better than that: It applies only to the choice of the value of the variable "Y" for the Minimum-Displacement Calendar. It refers to the accuracy of Y, in terms of how well it matches, overall, the lengths of the tropical years defined with respect to the various points of the ecliptic.

Forgive me for evaluating choices of Y for the Minimum-Displacement Calendar :^)


 

and so making comparison with other calendars impossible



So Karl thinks that comparison of Minimum-Displacement to other calendars is made impossible by the fact that I evaluate choices of values for the value of Y for the Minimum-Displacement Calendar.

:^)

I've already made it abundantly clear that my standard for choosing Y was not intended for comparing the Minimum-Displacement Calendar to other calendars.

If Karl wants to compare the accuracy of Minimum-Displacement to that of other calendars, then there are measures for that.

Karl said that he prefers the measure of displacement that I've called "D1". It's displacement that's constant throughout a displacement-year, and which changes abruptly at the end of a displacement-year.

A good measure of a calendar's accuracy is the maximum amount by which it is displaced, between its periodic displacement-extremes, with respect to ecliptic longitudes (as represented by the progress of some chosen mean-year--I use the MTY).

We've discussed two measures of displacement, which I've called (for brevity) D1 & D2. I've re-stated, above in this post, what I mean by D1.

Because D1 is the measure of displaceent that Karl said that he prefers, and because it's the one by which I evaluate Minimulm-Displacment's accuracy, and because it matches the mechanism of Minimum-Displacement, then D1 would be a fair choice for comparing Minimum-Displacement to other calendars.

By D1, Minimlum-Displacement's displacement-range is 7 days.

Karl: Use that to compare The Minimum-Displacement Calendar's accuracy to that of other calendars.

(In fairness, only compare it to other leapweek calendars, because leapweek calendars have more maximum periodic displacement than leapday calendars do.)




 

, such as the type of calendar Walter Ziobro has been suggesting or a minimum displacement calendar that changes Y & D in a different manner than described below.


If you want to compare the accuracy of the Minimum-Dispacement Calendar to the accuracy of Walter Ziobro's calendar, then find out the maximum periodic displacement-range of Walter's calendar. If it's greater than 7, then it's less accurate than Minimum-Displacement. If it's less than 7, then it's more accurate than Minimum-Displacment.


 

 

Because of this restriction, it does not deserve to be called “year-round accuracy”


As I said, I've quit using that name for that (rough) standard for choosing Y for the Minimum-Displacement Calendar.



 

, but perhaps “year-round mean year accuracy”, which describes it well.


Call it what you want, but I've made it clear from the start that it wasn't intended for comparing the accuracies of different calendars.

Michael Ossipoff

 

Karl

 

16(06(06

 

From: East Carolina University Calendar discussion List [mailto:[hidden email]] On Behalf Of Michael Ossipoff
Sent: 02 February 2017 18:24
To: [hidden email]
Subject: Re: Year-Round Accuracy RE: Displacement Calculation & RE: Jitter Calculation RE: ....

 

 

 

On Thu, Feb 2, 2017 at 11:24 AM, Karl Palmen <[hidden email]> wrote:

Dear Michael and Calendar People

 

It is obvious to me that Michael has not understood what I’ve been attempting to express at all.

 

We definitely agree on that.

 

 So I’ll make one more attempt with a simple example possibly exaggerated for clarity.

 

Ok.
 

 

Consider two calendars whose calendar mean years (Y) are exactly equal every year. The Y values are changed at exactly the same year to exactly the same new value. The only difference is that the first calendar never changes the displacement (D) whenever Y is changed and the second calendar always reduces the displacement by 0.2 day, whenever Y is changed. Then after five Y changes the second calendar will be running one day ahead of the first calendar (on average if they are leap week calendars).

 

Now we’d like to know which one of these two calendars is the more accurate.

 

Between the years when Y is changed, Y has been kept constant, while the reference-year that Y represents has been continually changing. The annual changes in D have been calculated by using that un-updated value of Y. So, at the end of that period, when Y is updated, D should be corrected too.

One of your 2 calendars doesn't do that correction of D..That calendar is leaving an error that gets bigger and bigger.

You other calendar has been reducing D by 0.2 day whenever Y is changed. If +0.2 is the amount by which astronomical observations say that D is off, at that time, then that calendar is the one that is the more accurate one.

 

Any method based only on comparing year lengths will give exactly the same accuracy to both calendars and so will not indicate which one of the two calendars is more accurate.

 

Yes, it's necessary to correct D, when it gets too far off.

 

Of course. I never denied that it's necessary to correct D (in addition to correcting Y) when D differs too much from what it would be if the actual reference year, instead of the constant Y, were used for each annual change in D,

At that time, D should be changed to the value that it would have if the actual value of the reference year, instead of the constant Y, were used when calculating each annual change in Y.

But, nevertheless, for any SEL, the calendar will be more accurate, in the way that I've been repeating, if the the length of the reference year is closer to the length of the tropical year measured at that SEL.

Could it be that I still don't know what you're talking about?

Yes, that's entirely possible!
 

 

I made another criticism of Michael’s definition of year-round accuracy and it is that it will not indicate the year-round accuracy of a calendar of the type Walter Ziobro  is suggesting and not show that such a calendar has better year-round accuracy.

 

When I spoke of "year-round accuracy", I merely meant the average closeness of the reference-year's length to the lengths of the tropical years measured at the various SELs.

As for the accuracy-benefit of that, I'll repeat something that I said above:

"But, nevertheless, for any SEL, the calendar will be more accurate, in the way that I've been repeating, if the the length of the reference year is closer to the length of the tropical year measured at that SEL."

I've repeated several times, the way in which that calendar is more accurate for that SEL, if its reference-year is close to the tropical year measured at that SEL. I'm not going to repeat that again.

 

I never claimed that that definition indicates the accuracy of a type of calendar that I don't know about. There are many kinds of calendars. I spoke about the choice of Y only for the Minimum-Displacement Calendar.

If we have different kinds of cars, their parts might not be interchangeable. The constant "Y" was intended solely for the Minimum-Displacement Calendar. Other calendars will have different rules, and that's as it should be.

Michael Ossipoff




 




Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: Year-Round Accuracy RE: Displacement Calculation & RE: Jitter Calculation RE: ....

Karl Palmen
Dear Michael and Calendar People

My objection about Michael's definition of 'year-round accuracy' is that it applies only to a narrow range of minimum-displacement calendars in which Y & D are changed in a certain manner. This objection remains regardless of the details this certain manner and the ideas underlying it. One should be able to compare a wide range of calendars for 'year-round accuracy' to see which one is more accurate.

I have difficulty understand Michael, because he is not precise about the ideas he mentions such as displacement and oscillation. So to continue discussion, I suggest Michael clarify each of his ideas afresh thinking carefully about each one.

Karl

16(06(11 till noon


________________________________
From: East Carolina University Calendar discussion List [[hidden email]] on behalf of Michael Ossipoff [[hidden email]]
Sent: 05 February 2017 20:21
To: [hidden email]
Subject: Re: Year-Round Accuracy RE: Displacement Calculation & RE: Jitter Calculation RE: ....


I said:

"I've just noticed an error that might be the cause of at least some of Karl's objections, regarding my statements about the choice of the Y value for Minimum-Displacement."

The error I was referring to was:

I neglected to say exactly what I was referring to the oscillation of.

As I clarified in my previous post to this thread, I was referring to the oscillation of a SEL's calendar-date, from one leapyear to the next.

Michael Ossipoff

On Fri, Feb 3, 2017 at 11:11 PM, Michael Ossipoff <[hidden email]<mailto:[hidden email]>> wrote:
I've just noticed an error that might be the cause of at least some of Karl's objections, regarding my statements about the choice of the Y value for Minimum-Displacement.

At least in some instances, I spoke of a drift of the "center of oscillation" of the date of a SEL's calendar-date.

I was referring to the oscillation of a SEL's date, over a leapyear-period. (the duration between leapyears)

So here's what I mean, and would have said, were not for that error:

For the short-term, during which the length of a tropical year (for an individual SEL, or the MTA) doesn't change:

The date of an SEL will always oscillate in the calendar, over every leapyear period, from one leapyear to the next. The "center of oscillation" of the date of that SEL is the midpoint of the extremes of its dates, over a leapyear-period.

If Y differs from the length of the tropical year defined for a particular SEL, and to the extent that it does, then the center of oscillation of the date of that SEL will drift in the calendar. .Of course a choice of Y is  going to favor some SEL, by reducing or eliminating the drift of the center of oscillation of its calendar-date.  ...or else Y can be chosen so that, by some measure of central tendency, Y is reasonably close to all of the SELs' tropical-year lengths, and doesn't favor one SEL over another.

I choose the latter. That's why I use the length of the MTY as Y.

As the length of the reference-year varies, over the millennia:

If Y has been chosen as the length of the June solstice tropical year, to avoid calendar-date drift of the center of oscillation of the calendar-date of the June solstice, then, because the length of the June solstice year changes very little over the next 9000 years or so, then Y will remain relatively close to the June solstice tropical year, and the center of oscillation of the date of the June solstice will remain fairly close to its initial value (the value that it has now, just before the adoption of the new calendar).

But, with that choice of Y, of course, even initially (before the length of the June solstice year changes any), of course Y's initial difference from the tropical-year lengths for the other SELs will be large, and so the center of oscillation of the calendar-date for some SEL other than the June solstice will have a high drift-rate, even initially.

Of course, as the length of the reference-year changes, over the millennia. the difference between Y, and the tropical-year lengths of the various SELs will change too--positively for some SELs, and negatively for ohers--and therefore so will the drift rates of the centers of oscillation of the dates for those SELs.

I choose the current length of the MTY as the initial value for Y, to, in some way try to reduce, over the whole eclliptic, the drift rate of the center of oscillation of the calenndar-date of any particular SEL.

I realize that calendar-date consistency for a particular SEL isn't what the D1 measure of calendar-displacement is about (that's the calendar-displacement measure that we agree on preferring). But I still prefer a Y-value that, in some sense, over the whole ecliptic, tries to minimize the drift-rate for the center of oscillation of the calendar-date of any particular SEL.

I emphasize that this isn't intended or offered as a way to compare the accuracy of two calendars.

Michael Ossipoff


On Fri, Feb 3, 2017 at 4:07 PM, Michael Ossipoff <[hidden email]<mailto:[hidden email]>> wrote:


On Fri, Feb 3, 2017 at 7:49 AM, Karl Palmen <[hidden email]<mailto:[hidden email]>> wrote:
Dear Michael and Calendar People

Michael is resolving this criticism by restricting the types of calendars that his measure of year-round accuracy applies to.


Not only that, but "year-round accuracy" didn't refer to the accuracy of calendars at all !

(Dick Hanke likes to use lots of exclamation-points. I'm going to borrow that practice from him!)

I've explained to you, several times, what I meant by "year-round accuracy". I'll say this just one more time. After this, I won't repeat it for you again:

When I spoke of the "year-round accuracy" of a value for Y (but not for a calendar), I was referring to its overall closeness to the lengths of the various tropical years, measured at the various points of the ecliptic. Period (full stop).

I was using "year-round accuracy", with the meaning stated above, in reference to choices of the value of the variable "Y", for use in the Minimum-Displacement Calendar.

The use of the expression "year-round accuracy" for that purpose, led to Karl's misunderstanding. His misunderstanding initially, therefore, is understandable and justifiable. His continual expression of the same misunderstanding is no longer justifiable.



This restriction is to a type of minimum displacement calendar, which Michael describes below.

It's a restriction to reference to the choice of Y-values for the Minimum-Displacement Calendar, and how well that Y-value matches the lengths of the tropical years defined with respect to the various points of the ecliptic.



Then the criticism of Michael’s definition of year-round accuracy is that it applies only to a certain type of minimum-displacement calendar

It's even better than that: It applies only to the choice of the value of the variable "Y" for the Minimum-Displacement Calendar. It refers to the accuracy of Y, in terms of how well it matches, overall, the lengths of the tropical years defined with respect to the various points of the ecliptic.

Forgive me for evaluating choices of Y for the Minimum-Displacement Calendar :^)



and so making comparison with other calendars impossible


So Karl thinks that comparison of Minimum-Displacement to other calendars is made impossible by the fact that I evaluate choices of values for the value of Y for the Minimum-Displacement Calendar.

:^)

I've already made it abundantly clear that my standard for choosing Y was not intended for comparing the Minimum-Displacement Calendar to other calendars.

If Karl wants to compare the accuracy of Minimum-Displacement to that of other calendars, then there are measures for that.

Karl said that he prefers the measure of displacement that I've called "D1". It's displacement that's constant throughout a displacement-year, and which changes abruptly at the end of a displacement-year.

A good measure of a calendar's accuracy is the maximum amount by which it is displaced, between its periodic displacement-extremes, with respect to ecliptic longitudes (as represented by the progress of some chosen mean-year--I use the MTY).

We've discussed two measures of displacement, which I've called (for brevity) D1 & D2. I've re-stated, above in this post, what I mean by D1.

Because D1 is the measure of displaceent that Karl said that he prefers, and because it's the one by which I evaluate Minimulm-Displacment's accuracy, and because it matches the mechanism of Minimum-Displacement, then D1 would be a fair choice for comparing Minimum-Displacement to other calendars.

By D1, Minimlum-Displacement's displacement-range is 7 days.

Karl: Use that to compare The Minimum-Displacement Calendar's accuracy to that of other calendars.

(In fairness, only compare it to other leapweek calendars, because leapweek calendars have more maximum periodic displacement than leapday calendars do.)





, such as the type of calendar Walter Ziobro has been suggesting or a minimum displacement calendar that changes Y & D in a different manner than described below.

If you want to compare the accuracy of the Minimum-Dispacement Calendar to the accuracy of Walter Ziobro's calendar, then find out the maximum periodic displacement-range of Walter's calendar. If it's greater than 7, then it's less accurate than Minimum-Displacement. If it's less than 7, then it's more accurate than Minimum-Displacment.




Because of this restriction, it does not deserve to be called “year-round accuracy”

As I said, I've quit using that name for that (rough) standard for choosing Y for the Minimum-Displacement Calendar.




, but perhaps “year-round mean year accuracy”, which describes it well.

Call it what you want, but I've made it clear from the start that it wasn't intended for comparing the accuracies of different calendars.

Michael Ossipoff

Karl

16(06(06

From: East Carolina University Calendar discussion List [mailto:[hidden email]<mailto:[hidden email]>] On Behalf Of Michael Ossipoff
Sent: 02 February 2017 18:24
To: [hidden email]<mailto:[hidden email]>
Subject: Re: Year-Round Accuracy RE: Displacement Calculation & RE: Jitter Calculation RE: ....



On Thu, Feb 2, 2017 at 11:24 AM, Karl Palmen <[hidden email]<mailto:[hidden email]>> wrote:
Dear Michael and Calendar People

It is obvious to me that Michael has not understood what I’ve been attempting to express at all.

We definitely agree on that.

 So I’ll make one more attempt with a simple example possibly exaggerated for clarity.

Ok.


Consider two calendars whose calendar mean years (Y) are exactly equal every year. The Y values are changed at exactly the same year to exactly the same new value. The only difference is that the first calendar never changes the displacement (D) whenever Y is changed and the second calendar always reduces the displacement by 0.2 day, whenever Y is changed. Then after five Y changes the second calendar will be running one day ahead of the first calendar (on average if they are leap week calendars).

Now we’d like to know which one of these two calendars is the more accurate.

Between the years when Y is changed, Y has been kept constant, while the reference-year that Y represents has been continually changing. The annual changes in D have been calculated by using that un-updated value of Y. So, at the end of that period, when Y is updated, D should be corrected too.
One of your 2 calendars doesn't do that correction of D..That calendar is leaving an error that gets bigger and bigger.
You other calendar has been reducing D by 0.2 day whenever Y is changed. If +0.2 is the amount by which astronomical observations say that D is off, at that time, then that calendar is the one that is the more accurate one.

Any method based only on comparing year lengths will give exactly the same accuracy to both calendars and so will not indicate which one of the two calendars is more accurate.

Yes, it's necessary to correct D, when it gets too far off.

Of course. I never denied that it's necessary to correct D (in addition to correcting Y) when D differs too much from what it would be if the actual reference year, instead of the constant Y, were used for each annual change in D,
At that time, D should be changed to the value that it would have if the actual value of the reference year, instead of the constant Y, were used when calculating each annual change in Y.
But, nevertheless, for any SEL, the calendar will be more accurate, in the way that I've been repeating, if the the length of the reference year is closer to the length of the tropical year measured at that SEL.
Could it be that I still don't know what you're talking about?
Yes, that's entirely possible!


I made another criticism of Michael’s definition of year-round accuracy and it is that it will not indicate the year-round accuracy of a calendar of the type Walter Ziobro  is suggesting and not show that such a calendar has better year-round accuracy.

When I spoke of "year-round accuracy", I merely meant the average closeness of the reference-year's length to the lengths of the tropical years measured at the various SELs.
As for the accuracy-benefit of that, I'll repeat something that I said above:

"But, nevertheless, for any SEL, the calendar will be more accurate, in the way that I've been repeating, if the the length of the reference year is closer to the length of the tropical year measured at that SEL."
I've repeated several times, the way in which that calendar is more accurate for that SEL, if its reference-year is close to the tropical year measured at that SEL. I'm not going to repeat that again.

I never claimed that that definition indicates the accuracy of a type of calendar that I don't know about. There are many kinds of calendars. I spoke about the choice of Y only for the Minimum-Displacement Calendar.
If we have different kinds of cars, their parts might not be interchangeable. The constant "Y" was intended solely for the Minimum-Displacement Calendar. Other calendars will have different rules, and that's as it should be.
Michael Ossipoff
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: Year-Round Accuracy RE: Displacement Calculation & RE: Jitter Calculation RE: ....

Michael Ossipoff


On Wed, Feb 8, 2017 at 4:50 AM, Karl Palmen <[hidden email]> wrote:
Dear Michael and Calendar People

My objection about Michael's definition of 'year-round accuracy'

I've repeated many times that I don't have a term "year-round accuracy". I don't have a definition for "year-round accuracy".

In a recent post, I told how I used that term, what I meant by it when I said it. At that time I said that I wasn't going to repeat that again (because I'd already repeated it so many times).

So, if Karl is still all confused, than i can't help him.

He needs to read one of the many posts in which I explained what I meant when I said "year-round accuracy" (a term that i no longer use or have a definition of).

But I've repeatedly told what I meant by "year-round accuracy", and have repeatedly clarified that i no longer use that term of have a definition for it.
 
is that it applies only to a narrow range of minimum-displacement calendars in which Y & D are changed in a certain manner. This objection remains regardless of the details this certain manner and the ideas underlying it.

No, it doesn't apply to anything, because I don't use that term or have a definition for it. However, I have amply explained what I meant by it at the time when I did say it.

I wanted a Y value that was an all-round good match for the lengths of the tropical-year-lengths defined for the various solar ecliptic longitudes. Period. That's it.


Karl is entirely making-up something to have a problem about.



One should be able to compare a wide range of calendars for 'year-round accuracy' to see which one is more accurate.

No, one shouldn't.

...because (as Ive' clarified for Karl many times) I only meant that I chose the MTY's length as the Y value of the Minimum-Displacement value, because the length of the MTY is a reasonable compromise among the lengths of the tropical-year-lengths defined for the various solar ecliptic longitudes.
 

I have difficulty understand Michael, because he is not precise about the ideas he mentions such as displacement and oscillation.

Oops! Karl forgot to specify a particular instance of the imprecision or lack of clarification he refers to..

That's Karl's typical vagueness.

I've defined my terms, and I've carefully clarified for Karl the intended meaning of everything that he asked about.

Because I trust Karl's word, then we can assume that he isn't going to continue this string of vague, unspecified, unsourced complaints that he calls a "discussion".

But can he keep that promise?

Again, thank you, Karl.

Michael Ossipooff

 

Karl

16(06(11 till noon


________________________________
From: East Carolina University Calendar discussion List [[hidden email]] on behalf of Michael Ossipoff [[hidden email]]
Sent: 05 February 2017 20:21
To: [hidden email]
Subject: Re: Year-Round Accuracy RE: Displacement Calculation & RE: Jitter Calculation RE: ....


I said:

"I've just noticed an error that might be the cause of at least some of Karl's objections, regarding my statements about the choice of the Y value for Minimum-Displacement."

The error I was referring to was:

I neglected to say exactly what I was referring to the oscillation of.

As I clarified in my previous post to this thread, I was referring to the oscillation of a SEL's calendar-date, from one leapyear to the next.

Michael Ossipoff

On Fri, Feb 3, 2017 at 11:11 PM, Michael Ossipoff <[hidden email]<mailto:[hidden email]>> wrote:
I've just noticed an error that might be the cause of at least some of Karl's objections, regarding my statements about the choice of the Y value for Minimum-Displacement.

At least in some instances, I spoke of a drift of the "center of oscillation" of the date of a SEL's calendar-date.

I was referring to the oscillation of a SEL's date, over a leapyear-period. (the duration between leapyears)

So here's what I mean, and would have said, were not for that error:

For the short-term, during which the length of a tropical year (for an individual SEL, or the MTA) doesn't change:

The date of an SEL will always oscillate in the calendar, over every leapyear period, from one leapyear to the next. The "center of oscillation" of the date of that SEL is the midpoint of the extremes of its dates, over a leapyear-period.

If Y differs from the length of the tropical year defined for a particular SEL, and to the extent that it does, then the center of oscillation of the date of that SEL will drift in the calendar. .Of course a choice of Y is  going to favor some SEL, by reducing or eliminating the drift of the center of oscillation of its calendar-date.  ...or else Y can be chosen so that, by some measure of central tendency, Y is reasonably close to all of the SELs' tropical-year lengths, and doesn't favor one SEL over another.

I choose the latter. That's why I use the length of the MTY as Y.

As the length of the reference-year varies, over the millennia:

If Y has been chosen as the length of the June solstice tropical year, to avoid calendar-date drift of the center of oscillation of the calendar-date of the June solstice, then, because the length of the June solstice year changes very little over the next 9000 years or so, then Y will remain relatively close to the June solstice tropical year, and the center of oscillation of the date of the June solstice will remain fairly close to its initial value (the value that it has now, just before the adoption of the new calendar).

But, with that choice of Y, of course, even initially (before the length of the June solstice year changes any), of course Y's initial difference from the tropical-year lengths for the other SELs will be large, and so the center of oscillation of the calendar-date for some SEL other than the June solstice will have a high drift-rate, even initially.

Of course, as the length of the reference-year changes, over the millennia. the difference between Y, and the tropical-year lengths of the various SELs will change too--positively for some SELs, and negatively for ohers--and therefore so will the drift rates of the centers of oscillation of the dates for those SELs.

I choose the current length of the MTY as the initial value for Y, to, in some way try to reduce, over the whole eclliptic, the drift rate of the center of oscillation of the calenndar-date of any particular SEL.

I realize that calendar-date consistency for a particular SEL isn't what the D1 measure of calendar-displacement is about (that's the calendar-displacement measure that we agree on preferring). But I still prefer a Y-value that, in some sense, over the whole ecliptic, tries to minimize the drift-rate for the center of oscillation of the calendar-date of any particular SEL.

I emphasize that this isn't intended or offered as a way to compare the accuracy of two calendars.

Michael Ossipoff


On Fri, Feb 3, 2017 at 4:07 PM, Michael Ossipoff <[hidden email]<mailto:[hidden email]>> wrote:


On Fri, Feb 3, 2017 at 7:49 AM, Karl Palmen <[hidden email]<mailto:[hidden email]>> wrote:
Dear Michael and Calendar People

Michael is resolving this criticism by restricting the types of calendars that his measure of year-round accuracy applies to.


Not only that, but "year-round accuracy" didn't refer to the accuracy of calendars at all !

(Dick Hanke likes to use lots of exclamation-points. I'm going to borrow that practice from him!)

I've explained to you, several times, what I meant by "year-round accuracy". I'll say this just one more time. After this, I won't repeat it for you again:

When I spoke of the "year-round accuracy" of a value for Y (but not for a calendar), I was referring to its overall closeness to the lengths of the various tropical years, measured at the various points of the ecliptic. Period (full stop).

I was using "year-round accuracy", with the meaning stated above, in reference to choices of the value of the variable "Y", for use in the Minimum-Displacement Calendar.

The use of the expression "year-round accuracy" for that purpose, led to Karl's misunderstanding. His misunderstanding initially, therefore, is understandable and justifiable. His continual expression of the same misunderstanding is no longer justifiable.



This restriction is to a type of minimum displacement calendar, which Michael describes below.

It's a restriction to reference to the choice of Y-values for the Minimum-Displacement Calendar, and how well that Y-value matches the lengths of the tropical years defined with respect to the various points of the ecliptic.



Then the criticism of Michael’s definition of year-round accuracy is that it applies only to a certain type of minimum-displacement calendar

It's even better than that: It applies only to the choice of the value of the variable "Y" for the Minimum-Displacement Calendar. It refers to the accuracy of Y, in terms of how well it matches, overall, the lengths of the tropical years defined with respect to the various points of the ecliptic.

Forgive me for evaluating choices of Y for the Minimum-Displacement Calendar :^)



and so making comparison with other calendars impossible


So Karl thinks that comparison of Minimum-Displacement to other calendars is made impossible by the fact that I evaluate choices of values for the value of Y for the Minimum-Displacement Calendar.

:^)

I've already made it abundantly clear that my standard for choosing Y was not intended for comparing the Minimum-Displacement Calendar to other calendars.

If Karl wants to compare the accuracy of Minimum-Displacement to that of other calendars, then there are measures for that.

Karl said that he prefers the measure of displacement that I've called "D1". It's displacement that's constant throughout a displacement-year, and which changes abruptly at the end of a displacement-year.

A good measure of a calendar's accuracy is the maximum amount by which it is displaced, between its periodic displacement-extremes, with respect to ecliptic longitudes (as represented by the progress of some chosen mean-year--I use the MTY).

We've discussed two measures of displacement, which I've called (for brevity) D1 & D2. I've re-stated, above in this post, what I mean by D1.

Because D1 is the measure of displaceent that Karl said that he prefers, and because it's the one by which I evaluate Minimulm-Displacment's accuracy, and because it matches the mechanism of Minimum-Displacement, then D1 would be a fair choice for comparing Minimum-Displacement to other calendars.

By D1, Minimlum-Displacement's displacement-range is 7 days.

Karl: Use that to compare The Minimum-Displacement Calendar's accuracy to that of other calendars.

(In fairness, only compare it to other leapweek calendars, because leapweek calendars have more maximum periodic displacement than leapday calendars do.)





, such as the type of calendar Walter Ziobro has been suggesting or a minimum displacement calendar that changes Y & D in a different manner than described below.

If you want to compare the accuracy of the Minimum-Dispacement Calendar to the accuracy of Walter Ziobro's calendar, then find out the maximum periodic displacement-range of Walter's calendar. If it's greater than 7, then it's less accurate than Minimum-Displacement. If it's less than 7, then it's more accurate than Minimum-Displacment.




Because of this restriction, it does not deserve to be called “year-round accuracy”

As I said, I've quit using that name for that (rough) standard for choosing Y for the Minimum-Displacement Calendar.




, but perhaps “year-round mean year accuracy”, which describes it well.

Call it what you want, but I've made it clear from the start that it wasn't intended for comparing the accuracies of different calendars.

Michael Ossipoff

Karl

16(06(06

From: East Carolina University Calendar discussion List [mailto:[hidden email]<mailto:[hidden email]>] On Behalf Of Michael Ossipoff
Sent: 02 February 2017 18:24
To: [hidden email]<mailto:[hidden email]>
Subject: Re: Year-Round Accuracy RE: Displacement Calculation & RE: Jitter Calculation RE: ....



On Thu, Feb 2, 2017 at 11:24 AM, Karl Palmen <[hidden email]<mailto:[hidden email]>> wrote:
Dear Michael and Calendar People

It is obvious to me that Michael has not understood what I’ve been attempting to express at all.

We definitely agree on that.

 So I’ll make one more attempt with a simple example possibly exaggerated for clarity.

Ok.


Consider two calendars whose calendar mean years (Y) are exactly equal every year. The Y values are changed at exactly the same year to exactly the same new value. The only difference is that the first calendar never changes the displacement (D) whenever Y is changed and the second calendar always reduces the displacement by 0.2 day, whenever Y is changed. Then after five Y changes the second calendar will be running one day ahead of the first calendar (on average if they are leap week calendars).

Now we’d like to know which one of these two calendars is the more accurate.

Between the years when Y is changed, Y has been kept constant, while the reference-year that Y represents has been continually changing. The annual changes in D have been calculated by using that un-updated value of Y. So, at the end of that period, when Y is updated, D should be corrected too.
One of your 2 calendars doesn't do that correction of D..That calendar is leaving an error that gets bigger and bigger.
You other calendar has been reducing D by 0.2 day whenever Y is changed. If +0.2 is the amount by which astronomical observations say that D is off, at that time, then that calendar is the one that is the more accurate one.

Any method based only on comparing year lengths will give exactly the same accuracy to both calendars and so will not indicate which one of the two calendars is more accurate.

Yes, it's necessary to correct D, when it gets too far off.

Of course. I never denied that it's necessary to correct D (in addition to correcting Y) when D differs too much from what it would be if the actual reference year, instead of the constant Y, were used for each annual change in D,
At that time, D should be changed to the value that it would have if the actual value of the reference year, instead of the constant Y, were used when calculating each annual change in Y.
But, nevertheless, for any SEL, the calendar will be more accurate, in the way that I've been repeating, if the the length of the reference year is closer to the length of the tropical year measured at that SEL.
Could it be that I still don't know what you're talking about?
Yes, that's entirely possible!


I made another criticism of Michael’s definition of year-round accuracy and it is that it will not indicate the year-round accuracy of a calendar of the type Walter Ziobro  is suggesting and not show that such a calendar has better year-round accuracy.

When I spoke of "year-round accuracy", I merely meant the average closeness of the reference-year's length to the lengths of the tropical years measured at the various SELs.
As for the accuracy-benefit of that, I'll repeat something that I said above:

"But, nevertheless, for any SEL, the calendar will be more accurate, in the way that I've been repeating, if the the length of the reference year is closer to the length of the tropical year measured at that SEL."
I've repeated several times, the way in which that calendar is more accurate for that SEL, if its reference-year is close to the tropical year measured at that SEL. I'm not going to repeat that again.

I never claimed that that definition indicates the accuracy of a type of calendar that I don't know about. There are many kinds of calendars. I spoke about the choice of Y only for the Minimum-Displacement Calendar.
If we have different kinds of cars, their parts might not be interchangeable. The constant "Y" was intended solely for the Minimum-Displacement Calendar. Other calendars will have different rules, and that's as it should be.
Michael Ossipoff

Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: Year-Round Accuracy RE: Displacement Calculation & RE: Jitter Calculation RE: ....

Karl Palmen

Dear Michael and Calendar People

 

I found Michael’s statement that his definition of ‘year-round accuracy’ applies only to a calendar mean year (Y) rather than a calendar and so is not subject to my criticism.

 

Karl

 

16(06(14

 

From: East Carolina University Calendar discussion List [mailto:[hidden email]] On Behalf Of Michael Ossipoff
Sent: 08 February 2017 18:31
To: [hidden email]
Subject: Re: Year-Round Accuracy RE: Displacement Calculation & RE: Jitter Calculation RE: ....

 

 

 

On Wed, Feb 8, 2017 at 4:50 AM, Karl Palmen <[hidden email]> wrote:

Dear Michael and Calendar People

My objection about Michael's definition of 'year-round accuracy'

 

I've repeated many times that I don't have a term "year-round accuracy". I don't have a definition for "year-round accuracy".

In a recent post, I told how I used that term, what I meant by it when I said it. At that time I said that I wasn't going to repeat that again (because I'd already repeated it so many times).

So, if Karl is still all confused, than i can't help him.

He needs to read one of the many posts in which I explained what I meant when I said "year-round accuracy" (a term that i no longer use or have a definition of).

But I've repeatedly told what I meant by "year-round accuracy", and have repeatedly clarified that i no longer use that term of have a definition for it.

 

is that it applies only to a narrow range of minimum-displacement calendars in which Y & D are changed in a certain manner. This objection remains regardless of the details this certain manner and the ideas underlying it.

 

No, it doesn't apply to anything, because I don't use that term or have a definition for it. However, I have amply explained what I meant by it at the time when I did say it.

I wanted a Y value that was an all-round good match for the lengths of the tropical-year-lengths defined for the various solar ecliptic longitudes. Period. That's it.

 

Karl is entirely making-up something to have a problem about.

 

One should be able to compare a wide range of calendars for 'year-round accuracy' to see which one is more accurate.

 

No, one shouldn't.

...because (as Ive' clarified for Karl many times) I only meant that I chose the MTY's length as the Y value of the Minimum-Displacement value, because the length of the MTY is a reasonable compromise among the lengths of the tropical-year-lengths defined for the various solar ecliptic longitudes.

 


I have difficulty understand Michael, because he is not precise about the ideas he mentions such as displacement and oscillation.

 

Oops! Karl forgot to specify a particular instance of the imprecision or lack of clarification he refers to..

That's Karl's typical vagueness.

 

I've defined my terms, and I've carefully clarified for Karl the intended meaning of everything that he asked about.

Because I trust Karl's word, then we can assume that he isn't going to continue this string of vague, unspecified, unsourced complaints that he calls a "discussion".

But can he keep that promise?

 

Again, thank you, Karl.

Michael Ossipooff


 


Karl

16(06(11 till noon


________________________________
From: East Carolina University Calendar discussion List [[hidden email]] on behalf of Michael Ossipoff [[hidden email]]
Sent: 05 February 2017 20:21
To: [hidden email]
Subject: Re: Year-Round Accuracy RE: Displacement Calculation & RE: Jitter Calculation RE: ....


I said:

"I've just noticed an error that might be the cause of at least some of Karl's objections, regarding my statements about the choice of the Y value for Minimum-Displacement."

The error I was referring to was:

I neglected to say exactly what I was referring to the oscillation of.

As I clarified in my previous post to this thread, I was referring to the oscillation of a SEL's calendar-date, from one leapyear to the next.

Michael Ossipoff

On Fri, Feb 3, 2017 at 11:11 PM, Michael Ossipoff <[hidden email]<mailto:[hidden email]>> wrote:
I've just noticed an error that might be the cause of at least some of Karl's objections, regarding my statements about the choice of the Y value for Minimum-Displacement.

At least in some instances, I spoke of a drift of the "center of oscillation" of the date of a SEL's calendar-date.

I was referring to the oscillation of a SEL's date, over a leapyear-period. (the duration between leapyears)

So here's what I mean, and would have said, were not for that error:

For the short-term, during which the length of a tropical year (for an individual SEL, or the MTA) doesn't change:

The date of an SEL will always oscillate in the calendar, over every leapyear period, from one leapyear to the next. The "center of oscillation" of the date of that SEL is the midpoint of the extremes of its dates, over a leapyear-period.

If Y differs from the length of the tropical year defined for a particular SEL, and to the extent that it does, then the center of oscillation of the date of that SEL will drift in the calendar. .Of course a choice of Y is  going to favor some SEL, by reducing or eliminating the drift of the center of oscillation of its calendar-date.  ...or else Y can be chosen so that, by some measure of central tendency, Y is reasonably close to all of the SELs' tropical-year lengths, and doesn't favor one SEL over another.

I choose the latter. That's why I use the length of the MTY as Y.

As the length of the reference-year varies, over the millennia:

If Y has been chosen as the length of the June solstice tropical year, to avoid calendar-date drift of the center of oscillation of the calendar-date of the June solstice, then, because the length of the June solstice year changes very little over the next 9000 years or so, then Y will remain relatively close to the June solstice tropical year, and the center of oscillation of the date of the June solstice will remain fairly close to its initial value (the value that it has now, just before the adoption of the new calendar).

But, with that choice of Y, of course, even initially (before the length of the June solstice year changes any), of course Y's initial difference from the tropical-year lengths for the other SELs will be large, and so the center of oscillation of the calendar-date for some SEL other than the June solstice will have a high drift-rate, even initially.

Of course, as the length of the reference-year changes, over the millennia. the difference between Y, and the tropical-year lengths of the various SELs will change too--positively for some SELs, and negatively for ohers--and therefore so will the drift rates of the centers of oscillation of the dates for those SELs.

I choose the current length of the MTY as the initial value for Y, to, in some way try to reduce, over the whole eclliptic, the drift rate of the center of oscillation of the calenndar-date of any particular SEL.

I realize that calendar-date consistency for a particular SEL isn't what the D1 measure of calendar-displacement is about (that's the calendar-displacement measure that we agree on preferring). But I still prefer a Y-value that, in some sense, over the whole ecliptic, tries to minimize the drift-rate for the center of oscillation of the calendar-date of any particular SEL.

I emphasize that this isn't intended or offered as a way to compare the accuracy of two calendars.

Michael Ossipoff


On Fri, Feb 3, 2017 at 4:07 PM, Michael Ossipoff <[hidden email]<mailto:[hidden email]>> wrote:



On Fri, Feb 3, 2017 at 7:49 AM, Karl Palmen <[hidden email]<mailto:[hidden email]>> wrote:
Dear Michael and Calendar People

Michael is resolving this criticism by restricting the types of calendars that his measure of year-round accuracy applies to.


Not only that, but "year-round accuracy" didn't refer to the accuracy of calendars at all !

(Dick Hanke likes to use lots of exclamation-points. I'm going to borrow that practice from him!)

I've explained to you, several times, what I meant by "year-round accuracy". I'll say this just one more time. After this, I won't repeat it for you again:

When I spoke of the "year-round accuracy" of a value for Y (but not for a calendar), I was referring to its overall closeness to the lengths of the various tropical years, measured at the various points of the ecliptic. Period (full stop).

I was using "year-round accuracy", with the meaning stated above, in reference to choices of the value of the variable "Y", for use in the Minimum-Displacement Calendar.

The use of the expression "year-round accuracy" for that purpose, led to Karl's misunderstanding. His misunderstanding initially, therefore, is understandable and justifiable. His continual expression of the same misunderstanding is no longer justifiable.



This restriction is to a type of minimum displacement calendar, which Michael describes below.

It's a restriction to reference to the choice of Y-values for the Minimum-Displacement Calendar, and how well that Y-value matches the lengths of the tropical years defined with respect to the various points of the ecliptic.



Then the criticism of Michael’s definition of year-round accuracy is that it applies only to a certain type of minimum-displacement calendar

It's even better than that: It applies only to the choice of the value of the variable "Y" for the Minimum-Displacement Calendar. It refers to the accuracy of Y, in terms of how well it matches, overall, the lengths of the tropical years defined with respect to the various points of the ecliptic.

Forgive me for evaluating choices of Y for the Minimum-Displacement Calendar :^)



and so making comparison with other calendars impossible


So Karl thinks that comparison of Minimum-Displacement to other calendars is made impossible by the fact that I evaluate choices of values for the value of Y for the Minimum-Displacement Calendar.

:^)

I've already made it abundantly clear that my standard for choosing Y was not intended for comparing the Minimum-Displacement Calendar to other calendars.

If Karl wants to compare the accuracy of Minimum-Displacement to that of other calendars, then there are measures for that.

Karl said that he prefers the measure of displacement that I've called "D1". It's displacement that's constant throughout a displacement-year, and which changes abruptly at the end of a displacement-year.

A good measure of a calendar's accuracy is the maximum amount by which it is displaced, between its periodic displacement-extremes, with respect to ecliptic longitudes (as represented by the progress of some chosen mean-year--I use the MTY).

We've discussed two measures of displacement, which I've called (for brevity) D1 & D2. I've re-stated, above in this post, what I mean by D1.

Because D1 is the measure of displaceent that Karl said that he prefers, and because it's the one by which I evaluate Minimulm-Displacment's accuracy, and because it matches the mechanism of Minimum-Displacement, then D1 would be a fair choice for comparing Minimum-Displacement to other calendars.

By D1, Minimlum-Displacement's displacement-range is 7 days.

Karl: Use that to compare The Minimum-Displacement Calendar's accuracy to that of other calendars.

(In fairness, only compare it to other leapweek calendars, because leapweek calendars have more maximum periodic displacement than leapday calendars do.)





, such as the type of calendar Walter Ziobro has been suggesting or a minimum displacement calendar that changes Y & D in a different manner than described below.

If you want to compare the accuracy of the Minimum-Dispacement Calendar to the accuracy of Walter Ziobro's calendar, then find out the maximum periodic displacement-range of Walter's calendar. If it's greater than 7, then it's less accurate than Minimum-Displacement. If it's less than 7, then it's more accurate than Minimum-Displacment.




Because of this restriction, it does not deserve to be called “year-round accuracy”

As I said, I've quit using that name for that (rough) standard for choosing Y for the Minimum-Displacement Calendar.




, but perhaps “year-round mean year accuracy”, which describes it well.

Call it what you want, but I've made it clear from the start that it wasn't intended for comparing the accuracies of different calendars.

Michael Ossipoff

Karl

16(06(06

From: East Carolina University Calendar discussion List [mailto:[hidden email]<mailto:[hidden email]>] On Behalf Of Michael Ossipoff
Sent: 02 February 2017 18:24
To: [hidden email]<mailto:[hidden email]>
Subject: Re: Year-Round Accuracy RE: Displacement Calculation & RE: Jitter Calculation RE: ....


On Thu, Feb 2, 2017 at 11:24 AM, Karl Palmen <[hidden email]<mailto:[hidden email]>> wrote:
Dear Michael and Calendar People

It is obvious to me that Michael has not understood what I’ve been attempting to express at all.

We definitely agree on that.

 So I’ll make one more attempt with a simple example possibly exaggerated for clarity.

Ok.


Consider two calendars whose calendar mean years (Y) are exactly equal every year. The Y values are changed at exactly the same year to exactly the same new value. The only difference is that the first calendar never changes the displacement (D) whenever Y is changed and the second calendar always reduces the displacement by 0.2 day, whenever Y is changed. Then after five Y changes the second calendar will be running one day ahead of the first calendar (on average if they are leap week calendars).

Now we’d like to know which one of these two calendars is the more accurate.

Between the years when Y is changed, Y has been kept constant, while the reference-year that Y represents has been continually changing. The annual changes in D have been calculated by using that un-updated value of Y. So, at the end of that period, when Y is updated, D should be corrected too.
One of your 2 calendars doesn't do that correction of D..That calendar is leaving an error that gets bigger and bigger.
You other calendar has been reducing D by 0.2 day whenever Y is changed. If +0.2 is the amount by which astronomical observations say that D is off, at that time, then that calendar is the one that is the more accurate one.

Any method based only on comparing year lengths will give exactly the same accuracy to both calendars and so will not indicate which one of the two calendars is more accurate.

Yes, it's necessary to correct D, when it gets too far off.

Of course. I never denied that it's necessary to correct D (in addition to correcting Y) when D differs too much from what it would be if the actual reference year, instead of the constant Y, were used for each annual change in D,
At that time, D should be changed to the value that it would have if the actual value of the reference year, instead of the constant Y, were used when calculating each annual change in Y.
But, nevertheless, for any SEL, the calendar will be more accurate, in the way that I've been repeating, if the the length of the reference year is closer to the length of the tropical year measured at that SEL.
Could it be that I still don't know what you're talking about?
Yes, that's entirely possible!


I made another criticism of Michael’s definition of year-round accuracy and it is that it will not indicate the year-round accuracy of a calendar of the type Walter Ziobro  is suggesting and not show that such a calendar has better year-round accuracy.

When I spoke of "year-round accuracy", I merely meant the average closeness of the reference-year's length to the lengths of the tropical years measured at the various SELs.
As for the accuracy-benefit of that, I'll repeat something that I said above:

"But, nevertheless, for any SEL, the calendar will be more accurate, in the way that I've been repeating, if the the length of the reference year is closer to the length of the tropical year measured at that SEL."
I've repeated several times, the way in which that calendar is more accurate for that SEL, if its reference-year is close to the tropical year measured at that SEL. I'm not going to repeat that again.

I never claimed that that definition indicates the accuracy of a type of calendar that I don't know about. There are many kinds of calendars. I spoke about the choice of Y only for the Minimum-Displacement Calendar.
If we have different kinds of cars, their parts might not be interchangeable. The constant "Y" was intended solely for the Minimum-Displacement Calendar. Other calendars will have different rules, and that's as it should be.
Michael Ossipoff

 

Loading...