stars.txt

[t00fri]

Sorry, but we established certain scientific standards in Celestia that are NOT up to you to modify or decide about. So the official distribution is to provide the public with a data base that is up to high scientific standards. That's the target whether you like it or not.

In fact I'm strongly in favor of high scientific standards for Celestia's standard databases. That isn't the issue. The issue is how to fix the immediate problems with the current version of the binary star database for the 1.5.0 release. If you can do it, that would great. If not, some other solution would be necessary. I never intended to suggest anything more than that.

OK Hank, then could you please "re"-interpret these statements of yours:

If it can't be completed in time, would it be better to leave the binary orbits out of 1.5.0 entirely?

As a worst case, the binary star work would need to be deferred to a later release.

Being a foreigner, I might well have been unable to translate these simple sentences?

Fridger, I really hope you won't remove the binary orbit data from CVS just because I've annoyed you. I certainly didn't intend to do so.

- Hank

Of course this is in nobody's interest. Certainly also not in mine after working on this stuff for ~ 1 whole year!

Bye Fridger

[Cham]

Just for completeness, I'm repeating here the list of conflicting binaries. This list doesn't include which star have a "wrong" greek name formulation :

Chi Dra (two times)
Eps Cet (two times)
51 Tau (two times)
9 Pup (two times)
HIP7580 and Kui 7
70 Tau and Fin 342
HIP2941 and ADS 520
HIP111528 and ADS 16098
81 Cnc and Fin 347Aa
Eta CrB (two times)
HIP85667 and SIG 2174
HIP87895 and HR 6697
HIP95995 and Gl 762.1
HIP98416 and Gl 773.3
HIP99376 and ADS 13461

The stars with the comment "(two times)" are confusing Celestia, since the name is exactly the same for both stars.

Personally, I simply commented out a star in each line (adding a # in the STC files).

Following Fridger's suggestion, it may be interesting to actually use all of these stars, even the doubles, but uniformizing their names, and adding "(spectroscopic)" and "(visual)". Something like this :

Code: Select all

"Chi Dra (visual)"
"Chi Dra (spectroscopic)"

Code: Select all

"HIP111528 (visual):ADS 16098 (visual)"
"HIP111528 (spectroscopic):ADS 16098 (spectroscopic)"

This way, it would completely solves the conflicts and the binaries could also show the uncertainties Fridger was talking about. And we don't have to decide which stars to remove from the database, since we keep all of them !

[t00fri]

Wow, Cham,

that sounds like a good idea that is both simple and very instructive/educational!

I'll definitely have a look into this possibility at my next free evening!

Bye Fridger

[hank]

OK Hank, then could you please "re"-interpret these statements of yours:

Gladly.

If it can't be completed in time, would it be better to leave the binary orbits out of 1.5.0 entirely?

That was mainly a rhetorical question. I thought the answer would be obvious. The only reason I could think of why the binary orbits database would be left out of Celestia was if you refused to allow it to be included (e.g. by removing it from CVS).

As a worst case, the binary star work would need to be deferred to a later release.

Obviously that's the worst case. I certainly hoped the worst case wouldn't happen.

Fridger, I really hope you won't remove the binary orbit data from CVS just because I've annoyed you. I certainly didn't intend to do so.

- Hank

Of course this is in nobody's interest. Certainly also not in mine after working on this stuff for ~ 1 whole year!

I'm glad we agree on that.

Fridger, I'm genuinely sorry to have annoyed you, even inadvertantly.

Martin,
...
I even promise that I'll try to be reasonably fast with providing the fixes.

So thankfully it seems this discussion wasn't completely useless. ;-)

But I DON'T have myself implicitly or explicitly
threatened by anybody that there might be a problem of having my binary orbit data prevail in the distribution, if I don't manage to observe some (unspecified) timeline for the fixes. Notably if that person is not authorized to make such far reaching statements.

Just so there's no future misunderstanding: I don't make threats, and I'm not authorized by anybody for anything I say here! :-)

- Hank

[t00fri]

There are a number of good reasons why Cham's proposal would have a high educational value:

In case of spectroscopic binaries, it is impossible to determine individually the semi-major axis a and the inclination of the orbit plane i (why?). However, the product of the semi-major axis and the sine of the inclination (i.e. a sin i) may be determined directly in linear units (e.g. kilometres). If either a or i can be determined by other means, as in the case of eclipsing binaries, a complete solution for the orbit can be found.

In case of visual binaries there are other pros and cons.

Being able to exhibit or compare the method-typical resulting uncertainties and also to visualize which aspects of the orbits these uncertainties dominantly influence, would be of great value.

So unless there is some kind of catch emerging, I am all for Cham's solution.

F.

[Cham]

So unless there is some kind of catch emerging, I am all for Cham's solution.

The only possible catch, is the name lenght. I suspect there's a limit. Adding "(spectroscopic)" may give a trouble (truncated name) in some cases. Maybe "(V)" and "(S)", but I much prefer the complete word, for clarity.

[hank]

Following Fridger's suggestion, it may be interesting to actually use all of these stars, even the doubles, but uniformizing their names, and adding "(spectroscopic)" and "(visual)". Something like this :

Code: Select all

"Chi Dra (visual)"
"Chi Dra (spectroscopic)"

Code: Select all

"HIP111528 (visual):ADS 16098 (visual)"
"HIP111528 (spectroscopic):ADS 16098 (spectroscopic)"

This way, it would completely solves the conflicts and the binaries could also show the uncertainties Fridger was talking about. And we don't have to decide which stars to remove from the database, since we keep all of them ! :wink:

But wouldn't this still cause duplicate stars to be displayed?

- Hank

[Cham]

But wouldn't this still cause duplicate stars to be displayed?

Yes, but there will be no conflict because of the names. These 15 binaries will display the "uncertainties", or "accuracy", or "reliability" of the orbits, according to the method used (visual or spectroscopic). The names will give a clear picture of the situation, at least.

In some sense, Fridger is right : it's an opportunity to display the "accuracy" of the method used (visual versus spectroscopy). Without the proper names, it would be confused (as it is, currently).

Of course, personaly, I would much prefer to not have duplicate binaries at all. But this is about an interesting compromise, if the names make the situation clear (including the barycenters).

[ElChristou]

Cham's solution seems very good, no need anymore of special rendering for those differences.

Now a quick noob question to learn something; a binary system is supposed to be constituted of 2 stars orbiting a barycenter within the ecliptic of the system, right? In the case of most discrepancy with stars from Hipparcos, we see that the stars are orbiting the barycenter each one in a different plane. Am I wrong of this is odd? (please don't be too hard with me... )

[Cham]

ElChristou,

The orbits of both stars should be in the same plane, or else, some physical laws may be violated (notably the conservation of angular momentum). This is a consequence of the fact that the gravitational force is "radial" (i.e. oriented toward the source of gravity).

[ElChristou]

ElChristou,

The orbits of both stars should be in the same plane, or else, some physical laws may be violated.

the binaries may be different in many aspects, depending of the technique used : The orbital plane may have a different inclination (or "tilt"), and also a different orientation (rotation of the "tilt" axis). The excentricity may even be different (circular vs elliptical orbits, for example), and the orbital size (semi-major axis) may also be different. There are few more parameters (position of a stars on its orbit, for example). Of course, if the uncertainties are small, the orbits from both techniques should be much alike, but some small variations may still remains, on each of these parameters. This is why the comparison visual vs spectroscopic is a good thing.

No, no, I don't want to discuss again the topic, both data are "official" so let's present them.

So the orbital plane may be different... ...another noob question then, is such system considered "stable"? I mean such configuration would not tend to align both planes with time?

[Cham]

So the orbital plane may be different... ...another noob question then, is such system considered "stable"? I mean such configuration would not tend to align both planes with time?

I'm not sure to understand the question.

In a simple binary (any system of two spherical stars), the motion is strictly confined in a plane. There's no other way. Both orbits are strictly aligned in the same plane. And that configuration is stable (ignoring the gravitational waves emission...).

Two stars alone can't have orbits with different inclination, for example. If you take two stars and throw them with any initial velocity and position, they will immediately follow some elliptical orbits in the same plane (defined by the initial velocities), assuming the velocities are not too high (lower than the "escape velocity").

[t00fri]

Martin,

thanks for you illustrative image (email). How about abreviating the two categories a bit. E.g.

HIP 7580A vis
HIP 7580A spc

Bye Fridger

[Cham]

How about abreviating the two categories a bit. E.g.

HIP 7580A vis
HIP 7580A spc

I think it's better.

[t00fri]

How about abreviating the two categories a bit. E.g.

HIP 7580A vis
HIP 7580A spc

I think it's better.

I'll do a preview tonight and post it here. So you all can try.

Also note: I previously eliminated already some doublets from the catalogs, also in a rather arbitrary manner. To offer the two ~ complete catalogs is now possible due to this addition of (vis) and (spc), respectively. This seems kind of attractive. What do you think? It's easy to get these former stars back in, too.

Another perhaps simpler option would be to assign one of the two sets, e.g. the visual one, as the default set without any added specs. Only for the second set there would then be a (spc) qualifier for distinction.

I am not sure which is better, the symmetrical one (vis) and (spc), or the unsymmetrical one. The latter would be perhaps preferable for Newbies who have no idea about (vis) and (spc) methodologies...

Bye Fridger

[Cham]

The unsymmetrical is attractive too. I think this solution is a very good compromise.

Just beware of the hip number in front of the name of some objects. It may conflict with the same number in the other file.

[hank]

I previously eliminated already some doublets from the catalogs, also in a rather arbitrary manner. To offer the two ~ complete catalogs is now possible due to this addition of (vis) and (spc), respectively. This seems kind of attractive. What do you think? It's easy to get these former stars back in, too.

Another perhaps simpler option would be to assign one of the two sets, e.g. the visual one, as the default set without any added specs. Only for the second set there would then be a (spc) qualifier for distinction.

I am not sure which is better, the symmetrical one (vis) and (spc), or the unsymmetrical one. The latter would be perhaps preferable for Newbies who have no idea about (vis) and (spc) methodologies...

I agree that the unsymmetrical naming scheme might be slightly better for users who are unaware of the meaning of the suffixes.

Do I understand correctly that a binary included in both datasets would be loaded twice (with names distinguished by suffix) and would be displayed as four stars? I think that might also be a problem for some users.

- Hank

[ElChristou]

Do I understand correctly that a binary included in both datasets would be loaded twice (with names distinguished by suffix) and would be displayed as four stars? I think that might also be a problem for some users.

But Hank, seems you have miss the root of the problem which is:

How do you chose a source or another?

The only way to eliminate the doubloons would be to change the sources, and for this you should be able to contest the calculations made by the original author...

(Now if you want to do the job... )

[hank]

But Hank, seems you have miss the root of the problem which is:

How do you chose a source or another?

The only way to eliminate the doubloons would be to change the sources, and for this you should be able to contest the calculations made by the original author...

(Now if you want to do the job... :wink:)

I realize that in some educational and scientific contexts it would be useful to display both versions of a binary system simultaneously. But I suspect most users most of the time would prefer to see only a single version. They would not care which methodology was used. Although the methodologies may differ as to the orbital elements, they do not differ as to the number of stars. If duplicate stars based on both methodologies are displayed, then the number of stars seen would be twice what would be seen in reality. I think that showing the correct number of stars would probably be preferred by most users.

- Hank

[t00fri]

Actually, I also had a second look and I tend to agree that on the long run, we should find a better way to support an alternative or simultaneous display of such binary orbit data from different methods.

While the naming distinction is fine, it is disturbing that one always sees more than the correct number of stars in orbit.

I think we will have to continue being creative as concerns possibilities to intuitively visualizing measurement uncertainties in Celestia! In my opinion it is a really important challenge to come up with a neat solution here.

Clearly this double display of 'spc' and 'vis' based orbits is interesting for astronomers, since it is really not clear which orbit aspects are most affected by particular parameter uncertainties.

On the other hand for less serious applications this is clearly a suboptimal choice.

So I propose I will arbitrarily kick out the doubled systems determined with the spectroscopic method.

Bye Fridger