Constellations and star's name related...

[ElChristou]

I was checking a few constellations with a friend and his son, when I noticed a few anomalies:

Auriga: what is the source of such representation? (a triangle) Why 6 stars are used to define this shape (a triangle!) in asterism.dat?

Why Capella (?± Aur) don't display it's name beside being defined correctly in starnames.dat (HIP 24608)?

[Toti]

Why 6 stars are used to define this shape (a triangle!) in asterism.dat?

One is included twice ("Eta Aur") and "To Aur" (?) is never found in the database and thus ignored, so we end up with 3+2 points, two of them in the "same" place. This is the triangle that's being drawn.

[ElChristou]

Why 6 stars are used to define this shape (a triangle!) in asterism.dat?

One is included twice ("Eta Aur") and "To Aur" (?) is never found in the database and thus ignored, so we end up with 3+2 points, two of them in the "same" place. This is the triangle that's being drawn.

Fine, but...

I'll do a check of the file to see if there is more things like this; also seems some present asterisms are really not the most common... As I don't want to lose my time, any interest if I do a revision?

[t00fri]

Why 6 stars are used to define this shape (a triangle!) in asterism.dat?

One is included twice ("Eta Aur") and "To Aur" (?) is never found in the database and thus ignored, so we end up with 3+2 points, two of them in the "same" place. This is the triangle that's being drawn.

Fine, but...

I'll do a check of the file to see if there is more things like this; also seems some present asterisms are really not the most common... As I don't want to lose my time, any interest if I do a revision?

I have done the asterisms a long long time ago (6 years...) and spend quite a bit of research as to their shapes. There is clearly not a scientific algorithm to build asterisms. The largest differences are of national origin (which again is no surprise). I have collected quite an amount of different asterism sets, but can't find them anymore...

Bye Fridger

[ajtribick]

The reason the name Capella is not displayed is because of Fridger's binary files, but this is really related to an underlying deficiency of the .stc file parser which has forced Fridger into a sub-optimal solution.

In spectbins.stc, Capella (=ALF Aur) is redefined:

Code: Select all

 Barycenter 24608 "ALF Aur"
 {
 RA        79.172065
 Dec       45.999031
 Distance  42.200414
 }


This wipes out the other associations to the star. If Fridger had defined the barycenter without the name part, the star names would get transferred with the HIP number. Alternatively what could be done is to use the full name definition from starnames.dat, however this would in effect force Fridger to track all changes to starnames.dat and update the binary star files if necessary, which is clearly not ideal.

Unfortunately what prevents this is the following bug, which I have brought up for quite a while now but is still present: catalogue references are not handled by the code that resolves the OrbitBarycenter property. This means that Fridger has to use ugly code like the following (from visualbins.stc) to get the binaries to work:

Code: Select all

 Barycenter 2237 "HIP 2237"
 {
 RA         7.088207
 Dec      -20.334776
 Distance 105.181230
 }

This associates the NAME "HIP 2237" with the star HIP 2237 (and also inserts it into the name list displayed after you press ENTER), which enables the barycenter to be referred to in the star definitions. This is clearly an ugly hack which is necessitated by the failure of OrbitBarycenter to resolve catalogue numbers.

[ElChristou]

...I have done the asterisms a long long time ago (6 years...) and spend quite a bit of research as to their shapes. There is clearly not a scientific algorithm to build asterisms. The largest differences are of national origin (which again is no surprise). I have collected quite an amount of different asterism sets, but can't find them anymore...

So YOU are the responsible for the Auriga monstrosity? (kidding)

Seriously, after digging a bit, indeed not easy to find some sources... Anyway I think there is a few improvements to do (Auriga is a good example), sometimes the names can be centered better etc...

Can I proceed or better not touch?

[ElChristou]

Tx Chaos for the explication. That's a bit a pity to loose the name of such a famous star... Well, if nothing can be done, patience...

[Toti]

Fine, but...

I am not sure that I get the meaning of that emoticon. Didn't you ask for the reason of the triangular shape with 6 points?

Seriously, after digging a bit, indeed not easy to find some sources... Anyway I think there is a few improvements to do (Auriga is a good example)

Fridger's constellations are taken from some file in this page:
http://www.skymap.com/constellations.htm

, sometimes the names can be centered better etc...

Right now the names are placed at the center of the first "chain" (array limited by brackets in asterisms.dat) of every constellation. I have a patch that uses the geometrical center, which is computed only once.

It could use a weighting factor to move the label towards the most important/visible elements. But in the end it's a matter of personal taste, I think.

(The patch also handles fading of labels and lines in a more robust way than the code I sent you. I'll upload it as soon as I could fix my SF problems)

[ElChristou]

Fine, but...

I am not sure that I get the meaning of that emoticon. Didn't you ask for the reason of the triangular shape with 6 points?

Off course, but Toti, your response here don't enlighten me much about the reason of 6 entries when 3 are enough to draw a triangle!
(Ok, sometime I'm a noob, but... )

Seriously, after digging a bit, indeed not easy to find some sources... Anyway I think there is a few improvements to do (Auriga is a good example)

Fridger's constellations are taken from some file in this page:
http://www.skymap.com/constellations.htm

Tx will have a look!

, sometimes the names can be centered better etc...

Right now the names are placed at the center of the first "chain" (array limited by brackets in asterisms.dat) of every constellation. I have a patch that uses the geometrical center, which is computed only once.

It could use a weighting factor to move the label towards the most important/visible elements. But in the end it's a matter of personal taste, I think.

(The patch also handles fading of labels and lines in a more robust way than the code I sent you)

Hey, interesting! (could we have a look? )
Concerning the placement of the name I fear there is not a coded way to get a good result everywhere; sometimes the placement is not easy to get well; it's not really a question of taste, just of balance name vs general shape...

[Toti]

Fine, but...

I am not sure that I get the meaning of that emoticon. Didn't you ask for the reason of the triangular shape with 6 points?

Off course, but Toti, your response here don't enlighten me much about the reason of 6 entries when 3 are enough to draw a triangle!
(Ok, sometime I'm a noob, but... )

But I summarized the problem pretty well: as you tacitly implied, the file "asterisms.dat" has some flaws; there are redundant entries and at least one is not recognized by the star database. I think that you misinterpreted my post: of course nothing there tries to justify the 6 points for this triangle definition.

Hey, interesting! (could we have a look? )

I sent it to the developers list but it never reached there. I am having some problems with SF mail that I hope to fix ASAP.

[t00fri]

Fridger's constellations are taken from some file in this page:
http://www.skymap.com/constellations.htm

That is not correct, if I remember well. I certainly used that url for some ideas about other implementations of asterisms, but mine were largely derived from our asterisms that we made for XEphem before. As you know, I was involved with XEphem for 11.5 years before joining Celestia. XEphem is a standard 2D ephemerides program to be found in every scientific institute since many years.

If the asterism.dat file is identical to one file from skymap.com then some dev has exchanged it without me knowing. But nothing such is listed in CVS.

Bye Fridger

[ElChristou]

...but mine were largely derived from our asterisms that we made for XEphem before. As you know, I was involved with XEphem for 11.5 years before joining Celestia. XEphem is a standard 2D ephemerides program to be found in every scientific institute since many years....

It's what I was thinking, but taking Andromeda as example show a big difference with XEphem...

Fridger, you haven't told me; can I check the file in detail to see what can be improved?

[Toti]

Fridger's constellations are taken from some file in this page:
http://www.skymap.com/constellations.htm

That is not correct, if I remember well. I certainly used that url for some ideas about other implementations of asterisms, but mine were largely derived from our asterisms that we made for XEphem before.

I based my assertion in the comment of this file, stating your sources:
http://www.celestiaproject.net/~t00fri/asterisms-skywatch.dat
But now that you mention it, the internal data is different from the current asterisms.dat. But in both files there are several identical definitions, Auriga being one of them.

[t00fri]

Fridger's constellations are taken from some file in this page:
http://www.skymap.com/constellations.htm

That is not correct, if I remember well. I certainly used that url for some ideas about other implementations of asterisms, but mine were largely derived from our asterisms that we made for XEphem before.

I based my assertion in the comment of this file, stating your sources:
http://www.celestiaproject.net/~t00fri/asterisms-skywatch.dat
But now that you mention it, the internal data is different from the current asterisms.dat. But in both files there are several identical definitions, Auriga being one of them.

Toti,

that's really mostly incorrect.

At the time, I did quite a substantial research in various directions as to whether there would be a semi-objective prescription for asterisms. So to start, I downloaded and Celestia-converted by Perlscript quite a number of asterism data files from various sources. I had > 5 different asterism.dat files for comparison (and http://www.celestiaproject.net/~t00fri/asterisms-skywatch.dat was one of them), which we discussed among forum members and devs.

Today I found some of these files from skymap.com that I had converted with this perlscript in 2002:
http://www.celestiaproject.net/~t00fri/asterism.pl

here they are for playing...
http://www.celestiaproject.net/~t00fri/asteris ... oschka.dat
http://www.celestiaproject.net/~t00fri/asterisms-rey.dat
http://www.celestiaproject.net/~t00fri/asterisms-rukl.dat
http://www.celestiaproject.net/~t00fri/asterisms-skywatch.dat

Moreover there are always a number of naturally identical definitions, simply because they were either more or less super popular or rather unique in layout. This does not mean at all that they were copied from some source.

Here is the diff -u between the actual CVS asterism.dat and what you discovered in my scatters.net home directory, i.e. asterisms-skywatch.dat. They are quite different as one can easily see from here:

http://www.celestiaproject.net/~t00fri/images/ ... m-diff.txt

Bye Fridger

[t00fri]

...but mine were largely derived from our asterisms that we made for XEphem before. As you know, I was involved with XEphem for 11.5 years before joining Celestia. XEphem is a standard 2D ephemerides program to be found in every scientific institute since many years....

It's what I was thinking, but taking Andromeda as example show a big difference with XEphem...

Fridger, you haven't told me; can I check the file in detail to see what can be improved?

Yes Andromeda I modified quite a bit, I remember that. There were quite a few further changes relative to XEphem.

There are many asterism schemes that are very "baroque" in their layout, using lots of faint stars, which I find horrible. On the other hand in XEphem we were too "spartanic" at times... so the point is to find the best compromise. One important constraint is NOT to invoke any stars weaker than a certain accepted apparent magnitude that is CLEARLY accessible in naked eye observations.


But what do you call "improve" ?

We are not talking about a nice piece of graphics here, but about a DIFFICULT task to satisfy the largely different prejudices of many people from different countries, where different illustrations of the constellations are shown since childhood! When I made the asterisms for Celestia, I had 11.5 years of XEphem experience on my back

Do whatever you think is REALLY necessary.


Bye Fridger

[ElChristou]

But finally, as there is no "official" asterisms, can we say that the more subjective is the better?

As example, here one for Canis Major:


Perso I think it looks good for a dog, but what from a pro point of view?
(but BTW, are pro still use asterisms?)

[ElChristou]

...Do whatever you think is REALLY necessary...

There is no REAL necessity to change the actual ones, only cosmetic improvement can be done (better position of name and perhaps more expressive asterism like the above example)...

[ElChristou]

...We are not talking about a nice piece of graphics here, but about a DIFFICULT task to satisfy the largely different prejudices of many people from different countries, where different illustrations of the constellations are shown since childhood!...

Ouch, that sound impossible to me!
Or we need some localized files adapting the asterisms depending on the user nationality! THAT would be original! (what soft will do such thing? )

[Toti]

Fridger,

I did that diff before posting my last message above. That's why I wrote:

But now that you mention it, the internal data is different from the current asterisms.dat.

And Christophe's initial post was concerned with the problems found in e.g. Auriga, so that's the domain that I found most relevant to this discussion from the beginning:

But in both files there are several identical definitions, Auriga being one of them.

This does not mean at all that they were copied from some source.

I vaguely recall having read a long time ago about your problems with the extraction of constellation representations, so I never intended to imply this. In my second post, "based" would have been a more precise wording than "taken", perhaps?.

[t00fri]

But finally, as there is no "official" asterisms, can we say that the more subjective is the better?

As example, here one for Canis Major:


Perso I think it looks good for a dog, but what from a pro point of view?
(but BTW, are pro still use asterisms?)

Christophe,

CMa is perhaps a good example. If you play with some of my converted asterism files referred to above, you'll find some VERY detailed constellation figures. But the prize is that pretty weak stars are involved.

I am sure you can find some weak stars that will allow you to draw a CMa asterism such that one can even tell the sex of that celestial dog

You can also have a look at the special asterism display in Stellarium. ...

I don't use any asterisms. I know the stars in and out since I built my first telescope at the age of 11 or so. Therefore that asterism subject really doesn't excite me too much anymore.

Bye Fridger