Equatorial grid display bug over the time

Report bugs, bug fixes and workarounds here.
Avatar
Topic author
jogad
Posts: 458
Joined: 17.09.2008
With us: 11 years 2 months
Location: Paris France

Equatorial grid display bug over the time

Post #1by jogad » 26.11.2009, 20:08

Hello,

In the new version of Celestia there is a precession of the earth.
I wanted to demonstrate the movement of the north celestial pole among the stars.

Here is the nowadays situation.
now.jpg
We are looking in the direction of the north celestial pole.
And of course it is in the Earth's axis. :D

And now in the year 11608
later.jpg
The Earth's axis has shifted in respect of the stars
but the equatorial grid has not moved with it. :(

I think this is wrong. Is it? :?:

chris
Site Admin
Posts: 4211
Joined: 28.01.2002
With us: 17 years 10 months
Location: Seattle, Washington, USA

Re: Equatorial grid display bug over the time

Post #2by chris » 26.11.2009, 22:00

later.jpg
The Earth's axis has shifted in respect of the stars
but the equatorial grid has not moved with it. :(

I think this is wrong. Is it? :?:

It depends on which equatorial coordinate system the grid is meant to represent. In Celestia, it's the mean equator and equinox of J2000, i.e. the equatorial coordinate system is frozen at 12:00 January 1, 2000. One could also have a grid depicting a coordinate system based on the 'mean equator of date'--this would move as the Earth precessed. Yet another equatorial coordinate system is based on the 'true equator of date.' This is based on the instantaneous orienation of the Earth's equator, and it includes the shorter period effects of nutation as well as the slower motions due to precession.

I'd say that the issue is one of documentation: we need to specify precisely what is meant by 'equatorial grid.'

--Chris

Avatar
Cham M
Posts: 4325
Joined: 14.01.2004
Age: 55
With us: 15 years 11 months
Location: Montreal

Re: Equatorial grid display bug over the time

Post #3by Cham » 26.11.2009, 22:34

chris wrote:
later.jpg
The Earth's axis has shifted in respect of the stars
but the equatorial grid has not moved with it. :(

I think this is wrong. Is it? :?:

It depends on which equatorial coordinate system the grid is meant to represent. In Celestia, it's the mean equator and equinox of J2000, i.e. the equatorial coordinate system is frozen at 12:00 January 1, 2000.

I'm seeing this as a feature, instead of a "bug" : activating the J2000 equatorial grid, we could feel better the precession of the equinoxes, and show why it is important to take this effect into account in modern astronomy.

Of course, a moving equatorial grid could also be usefull.
"Well! I've often seen a cat without a grin", thought Alice; "but a grin without a cat! It's the most curious thing I ever saw in all my life!"

Avatar
Topic author
jogad
Posts: 458
Joined: 17.09.2008
With us: 11 years 2 months
Location: Paris France

Re: Equatorial grid display bug over the time

Post #4by jogad » 27.11.2009, 10:15

Hello,

Cham wrote:I'm seeing this as a feature, instead of a "bug" : activating the J2000 equatorial grid, we could feel better the precession of the equinoxes, and show why it is important to take this effect into account in modern astronomy.

I respect this argumentation because it's yours but I don't understand it. This sounds for me like saying :
"Look. We have the Earth in a uniform movement around a circular obit. As you can see the positions are false. This "feature" demonstrates the importance to take in account an elliptic and non-uniform movement." :roll:

To feel better the precession, I'd rather see the pole of the equatorial grid, or the intersection between its equator and the ecliptic, to move among the stars.
This will be the direct visualization of the phenomenon.

But the most important reason why a moving grid seems a better choice to me is the following:
The grids are used to identify but also to make measurements.

And serious problems are to worry with this frozen grid: :x

- If one tries to determine the seasons by the passage of the sun at the equinox point represented by the intersection of ecliptic and equatorial grids, the result will be even more wrong when we depart from the year 2000.

- Similarly, if we try to measure the duration of the year on a significant period of time. We will get the sidereal year (which nobody cares) instead of the tropical year which is the one that brought back the season on the same dates on calendars.

One might think that a fixed grid would be an advantage as regards the coordinates of the stars. Coordinates measured with the grid will correspond to the catalog (J2000).
But in reality, for calculations of stars positions, we have always to convert the catalog coordinates into the actual ones.
So the most useful information are current coordinates corresponding to a moving grid.
And that grid will show how the coordinates change.

I agree that the decision to display a fixed grid is a choice and not a bug.
But I do not see what are the advantages of a fixed grid over a moving one. Except perhaps to point at the errors it generates.
I probably missed something. :oops: But what?

Avatar
t00fri
Developer
Posts: 8772
Joined: 29.03.2002
Age: 17
With us: 17 years 8 months
Location: Hamburg, Germany

Re: Equatorial grid display bug over the time

Post #5by t00fri » 27.11.2009, 10:58

Jogad,

Chris' explanation was perfectly correct and exhaustive from a physics point of view. The infinite variety of pysically correct coordinate systems and the transformations between them are not always intuitive, though.

While personally you may well prefer a time-dependent frame, it is a fact that today virtually all catalogs of stars and other celestial objects refer to the J2000 frames. They are standard frames and that's why Celestia also offers them as grid displays! The scientific catalogs just provide the initial data required to precess the objects' positions to some desired time using the laws of physics along with certain proper motion parameters..

Taking into account that in printed media storage space is highly limited, this traditional method of archiving has no alternative, really. Except perhaps that the older catalogs used the J1950 equinox instead of J2000.

So for any fixed-time equatorial frame some well-defined calculational work by the user is implicit to precess the archived data to the time of interest. An invaluable advantage of the J2000 frames is of course that they represent a standardized, accurate reference for comparing positions of celestial objects!

Since with Celestia, we have many more possibilities for displaying celestial objects than in case of printed media, I also recommend to include an option for precessing our standard J2000 coordinate grids to the actual time of observation. As Cham already emphasized, this would be an instructive feature.

Fridger
Image

Avatar
Topic author
jogad
Posts: 458
Joined: 17.09.2008
With us: 11 years 2 months
Location: Paris France

Re: Equatorial grid display bug over the time

Post #6by jogad » 27.11.2009, 11:54

Fridger,

I agree with you on all points.

It is true that a time-dependent frame would have been more helpful for my needs.
I never questioned the need to store objects with standard coordinates from the catalogs.
But the internal storage is one thing and how we display data is a matter of choice.
For me the "bug" was particularly in this election.

I agree especially your final recommendation to offer the possibility to choose the display mode of the frame.

Maybe for a next release?

Thank you

Avatar
t00fri
Developer
Posts: 8772
Joined: 29.03.2002
Age: 17
With us: 17 years 8 months
Location: Hamburg, Germany

Re: Equatorial grid display bug over the time

Post #7by t00fri » 27.11.2009, 13:15

Jogad,

let me add some second thoughts about time-dependent grids:

mainly for reasons of saving CPU time, Celestia's stars and deep-sky objects do NOT (yet) incorporate precession! Hence, a fixed-time J2000 set of grids appears much more appropriate in most cases of interest.

Another issue is that any precession calculation for times that are MUCH different from the actual times of celestial position measurement are considerably uncertain! The reason is that the required precessional parameters (<=> velocity vectors) have been observed only over comparatively short times.

It would not correspond well to Celestia's science-based design philosophy to emphasize displays involving huge uncertain extrapolations in time.


Fridger
Image


Return to “Bugs”

Who is online