Celestia Forums

Firstly, I learnt (ironically from the horoscope page) that 2001 KX76 is to be named Ixion. In mythology, Zeus caught him looking at his gal and sentenced him to roll around on a flaming wheel. Don't know whether you knew this already, but...

And mainly, how about someone designing a Ptolemic Universe? With the Earth in the centre, followed by the Sun and the planets (no Pluto of course) and finally stars orbiting at the edge? It might mess things up with two Sol systems in the same universe, but it's just an idea.

This Ptolemaic universe sounds like a fun project-- but very ambitious! Technically it would involve replacing the entire universe with a new one, but I suppose that since the Ptolemaic universe wasn't very big, you could just stash it off in intergalactic space somewhere.

To do it right, you'd want to give everything a custom orbit written in C, to get all those epicycles and equants in. The default treatment of stars and planets in Celestia couldn't be used. Does Celestia allow stars to orbit planets? (Or could a "sun" be modified enough to resemble Earth, and vice versa?) I wonder if there's a way to hack in a representation of the crystal spheres.

It would take a lot of historical research, too.

For more fun, it could be extended into Dante's cosmology, with Mount Purgatory, a cutaway view of Hell, and angelic spheres surrounding the Empyrean.

Matt McIrvin wrote:For more fun, it could be extended into Dante's cosmology, with Mount Purgatory, a cutaway view of Hell, and angelic spheres surrounding the Empyrean.

I might be able to pull this one off...Where would I get data on this?

Rassilon wrote:

Matt McIrvin wrote:For more fun, it could be extended into Dante's cosmology, with Mount Purgatory, a cutaway view of Hell, and angelic spheres surrounding the Empyrean.

I might be able to pull this one off...Where would I get data on this?

Well, for the Dante part, you'd want to go straight to the source, the Divine Comedy... most editions have helpful diagrams as well, especially for the Inferno, which is a rather complicated Hell.

For quantitative data on the Ptolemaic universe, I'm not sure... maybe there are annotated translations of the Almagest available somewhere. I remember seeing some nice diagrams and explanations of Ptolemy's epicycles and equants in a book called Constructing the Universe by David Layzer; it was part of the Scientific American Library. But it didn't have the numbers.
[/i]

Well if theres some data on the net, that could help profoundly...But otherwise I'll check in with my library for those books...Thanks.

Hmm... sounds like this'll turn out quite interesting...
As for the stars orbiting the Earth, they could just be moons with a star texture (for the beginning, at least). And tell people not to look at them up close.

I'd love to see a 3D Ptolomeic Solar System.... the idea of the epicycles is strangely appealing. But didn't Ptolemy think that all the stars were on a single fixed sphere ? If so things would get REALLY weird if you tried to zoom up close... you could end of going off the edge of the Universe !

It would be nice to be able to (ab)use celestia as an atlas
of models of the universe, flat earth, aristotle, ptolemy, dante...

I want to see the cosmos from the outside! a black ball hovering
in divine light, with the constellations appearing mirrored from
without!

just, how would you want the crystal spheres rendered - they are
transparent, right? apart from that, we could start with just dumb
circular motions around earth in the center (aristotle) but that wouldn't
allow for retrograde motion and therefore the planets could not be shown
in their correct positions as seen from earth.

To implement epicycles it appears we would have to descend into the
sourcecode: Apparently, the orbits are encoded as kepler ellipses
(e.g. for mercury):

EllipticalOrbit {
Period 0.2408
SemiMajorAxis 0.3871
Eccentricity 0.2056
Inclination 7.0049
AscendingNode 48.33167
LongOfPericenter 77.456
MeanLongitude 252.251
}

Incidentally, this makes me wonder, how can celestia be accurate
with kepler orbits?? mercury's perihelion rotation??
i would expect any planetarium software to employ (unintuitive
but accurate) series expansions;


anyway, we could come up with a different kind of orbit, as in

EpiCyclicalOrbit {
Period ...
Radius ....
EpiPeriod ...
EpiRadius ...
EpiEpiPeriod ...
EpiEpiRadius
etc.
}

and then we would want keystroke 'O' to display the epicycles.
also, the opaque fixed stars sphere is probably not achievable
without tampering with the program, since it appears you cannot
be 'inside' objects at the moment.


DAB

I am reassured after checking that mercury's perihelion rotation is
in fact there. Indeed, all planets seem to have series expansions for
their orbits, for example earth goes into some kind of strange figure-8
orbit after year 55000 or so :-)
Why then the EllipticOrbit parameters? (As you can tell, I have not
bothered to look at the source, so far...)
Anyway, it seems to be related to the CustomOrbit keyword in
the .ssc file, but then the actual parameters would be hardcoded...

In this case we probably could add weird epicyclical orbits somehow.
Come to think of it, by introducing fictitious bodies with radius 0
we can do that anyway. The planets would end up being a moon of
a moon or so.

Anyway, this leaves the fixed stars sphere. I still suppose we
would "really" need to change the program for that...

DAB

The CustomOrbit keyword does refer to hardcoded orbital calculations in vsop87.cpp and customorbit.cpp . . . I've left the elliptical elements in solarsys.ssc even though they're overridden by the custom orbits. You could modify customorbit.cpp and add whatever strange orbits you like. As for the fixed star sphere, it seems like you could create a custom star database where all stars have the same distance from the Earth and all brightness variations are due to differences in absolute magnitude.

--Chris

I tried to do it that way. I downloaded hip_main.dat of the hipparcos catalogue
(just using dummy files for the other files buildstardb requires).

I reformatted the catalogue using the folowing perl script:

#! /usr/bin/perl
while($in=<>){
if($in=~/^(H\|([^\|]+\|){4})([\-\d\. ]+)(\|([^\|]+\|){5})([\-\d\. ]+)(\|.*)\n/){
$a=$1;$mag=$3;$b=$4;$par=$6;$c=$7;
if($mag<7){print $a.$mag.$b."9999.99".$c."\n";}}
}

piping hip_main.dat into this, I get a catalogue of all stars with magnitude<7
(about 15000) reformatted to parallax 9999.99

This corresponds to a sphere radius of 0.3ly. Note that to get a smaller sphere
I would need to tamper with buildstardb (i.e. changing the format of hip_main.dat,
as there is no space for parallaxes>=10000.

Anyway, I get a very cool fixed stars sphere I can zoom out of, and setting
"follow sync earth" I can watch it rotate once a day.

Also note that this way, the constellation lines are not rendered when inside
the sphere (because of the 'dont-draw-lines-when-too-close-to-star' feature)

If you can't be bothered to mess with the perl script but would like to have a look,
I can put the resulting stars.dat (1.2M) on a server somewhere.
(also more economical to download than the 53M catalogue...)


cheers,

dab.

dab wrote:
If you can't be bothered to mess with the perl script but would like to have a look,
I can put the resulting stars.dat (1.2M) on a server somewhere.
(also more economical to download than the 53M catalogue...)


cheers,

dab.

Very cool!
Mail it to me at claurel@shatters.net and I'll put it on shatters.net.

--Chris

Chris, I don't think this deserves to be put on shatters.net so far,
it's just a first half-assed attempt at a geocentric universe

anyway, I have gone a bit further now. I have made a replacement
data/ directory for geocentricity. You can download it here:

http://flaez.ch/scratch/ptolemy.tar.gz

besides my stars.dat it also contains buildptolemy.cpp - a modified
buildstardb.cpp that reads the original hipparcos catalogue and
instead of sol adds a star "Earth" with magnitude 32.5 which yields
a star of a radius of approx 6000km.

the modified solarsys.ssc accounts for that and inserts a planet "Sol"
with texture gstar.jpg instead. The planets have circular orbits & I have
shrunk everything by a factor 100 (except for the moon which is
shrunk by a factor 10). Of course here histoorical research into how
big people thought the spheres actually were would begin, but
this is just a stab at getting the general picture, so far. Does anybody
else want to help me with this?

I think this is about as far as I can go without modifying the celestia program.
But I am not sure this is the right way, because eventually I would want the
sun (now a planet) to be the source of light, and not earth.

I have modified the program a wee bit introducing a new spectral class
in order to get the earth.png texture for earth. but since the program thinks
this is such an unimportant star (magnitude 32.5!) it stops rendering earth
when I zoom out further than the moon.


dab