[Kstars-devel] Drawing comet tail patch

[Charles Peng]

Hi Akarsh,

On Sun, Apr 19, 2009 at 07:48:05PM +0530, Akarsh Simha wrote:
> For the time being, you could try computing the derivative numerically
> within KSPlanetBase::findPosition(), where you can ask where the comet
> will be some time back and compute the tangent to the trajectory.

I haven't thought of such use of findPosition(). I think I need to
understand the classes better. :)

But another question arises that a comet at its apogee has a much lower
angular velocity relative to the Earth than at its perigee, so if I
compute the tangent using a constant time interval, I may result in a
great variation of errors. ( Is this variation of errors significant? )
If findPosition() one second ago always return a different position, I
think the variation is tolerable. But I guess it will return the very
same position if the corresponding comet is very far. ( Haven't tested )

I think of a possible solution, if solving this problem is necessary.
RFC: geometrically increase the time interval, ie: 1s, 2s, 4s, ..., to
get an acceptable time interval fast. The constant needs to be tuned or
maybe a Fibonacci series does the job better.

> Maybe you'd want to look at Cartes du Ciel's source-code
> [http://sourceforge.net/projects/skychart/] and see what he does.

Oh, Cartes du Ciel is really doing a great job. I'll read his code this
week. Getting busy again now...

Regards,
Charles