Reconciling nearstars.stc with visualbins and specbins

[fsgregs]

Hi. I've noticed that the cfg file for 1.4.0pre9 still indicates that the binary stars referenced in visualbins and specbins sometimes conflict with the same stars listed in the nearstars.stc file. I can confirm that it is still a problem. Unfortunately, we can't get to see the binary stars orbiting each other until we fix the problem.

Is anyone already working on this, or is this something you all would like me to tackle? I can manually do it (it will take a long time ) by going to every binary set in visualbins and specbins, identifying those stars that have a nearstars component in the wrong place, and remove the star from the nearstars data set. It will be a lot of work, but I can do it for the good of the program if no one else is doing so?

If there is an easier way to reconcile the three files, let me know. For example, it would be much simplier to just not use nearstars.stc. Is that an option?

Frank

[t00fri]

Hi. I've noticed that the cfg file for 1.4.0pre9 still indicates that the binary stars referenced in visualbins and specbins sometimes conflict with the same stars listed in the nearstars.stc file. I can confirm that it is still a problem. Unfortunately, we can't get to see the binary stars orbiting each other until we fix the problem.

Is anyone already working on this, or is this something you all would like me to tackle? I can manually do it (it will take a long time ) by going to every binary set in visualbins and specbins, identifying those stars that have a nearstars component in the wrong place, and remove the star from the nearstars data set. It will be a lot of work, but I can do it for the good of the program if no one else is doing so?

If there is an easier way to reconcile the three files, let me know. For example, it would be much simplier to just not use nearstars.stc. Is that an option?

Frank

Frank,

what binary stars are you precisely referring to in your "confirmation of the problem"? Were you referring to celestia.cfg of 1.4.0pre9? What is written there? The one at CVS is 3 months old...

I thought I did eliminate all the overlaps between my visualbins and spectbins files and nearstars that had been located by Grant some time ago.

So could you become more specific please??


Bye Fridger

[fsgregs]

Fridger:

You are correct. When I use stars.dat, + nearstars.stc and visualbins.stc and specbins.stc files, I see no conflicts, at least with the first 10 or so binary pairs.

MY problem is that I am using the new 1 million stars.dat file for 1.4.0. Therein lies the problem. I tried the first five binary star sets in visualbins.stc list and found all five in conflict with the stars1M.dat file in that three stars appear ... the two binaries in the visualbins file and the one specified in the stars.dat file. The stars1M.dat star just sits there annoyingly!!! close by ... while your stars orbit their barycenters.

I really love the 1M and even 2M stars.dat file set and would like to keep using it, without messing up your orbiting binaries. Is there any reasonable way to edit the dat file to remove all of the stars mentioned in visualbins and specbins, or does it have to be done laboriously by hand ... one star at a time? If the latter is the case, just how many binaries are in visualbins and specbins? What kind of work am I looking at?

Frank

[selden]

Frank,

The problem is in the 1M and 2M stars.dat files.

You need to contact Pascal and try to get him to fix the files.

[t00fri]

erased

[fsgregs]

Fridger:

Crap! ....

I Did not mean to open a "can of worms" as they say in America.

Please don't remove visualbins and specbins from the distribution. We all benefit from seeing those great binaries orbiting each other. Your work is too good to be lost to the community.

I don't even know if there is still any conflict between nearstars and your files. As I mentioned, I found the problem only when using the 1 million stars.dat file, which I assume includes some of the stars in nearstars.

Perhaps Chris simply included an older version of the celestia.cfg file when packaging up 1.4.0pre9 by mistake.

I'm sure he did not mean to snub your efforts. Contact him and see if you can resolve this without removing binary stars from Celestia.

Respectfully,


Frank

[t00fri]

Frank,

OK, I checked that the celestia.cfg file in 1.4.0pre9 is apparently (hence possibly erroneously) identical to the one of 1.4.0pre7, where indeed there were still ~10 overlapping multiple star systems with nearstars.stc. After I had carefully eliminated these conflicting stars from my 2 data files, we had agreed that visualbins and spectbins would be part of 1.4.0 final.

So I erase my above post for now and first check with Chris whether his commenting out of my files in 1.4.0pre9 was indeed unintentional (this time)...

As to your request above, this is problematic, since I am not willing to cut out arbitraily some data from published scientific catalogs. The elimination of the few systems overlapping with nearstars.stc is to be considered as a mere workaround for now. It goes strongly against my general strategy of dealing with scientific data in Celestia. It is certainly not to be generalized...

Clearly, we do need some additional coding here that checks for such cases and replaces individual stars by the binary orbit systems if these exist.

Bye Fridger

[ajtribick]

Frank, can you confirm if the problematic binaries involve stars with a Tycho number ending in 0.

The 1M and 2M star databases include some companion stars, which were assigned fictitious catalogue numbers, which if I recall correctly were Tycho numbers that end in 0.

If this is the case then stars with "HIP" numbers < 1000000000 and above the end of the Hipparcos range should be removed from these databases for the 1.4.0 versions.

[fsgregs]

Dear Chaos:

Frank, can you confirm if the problematic binaries involve stars with a Tycho number ending in 0.

The 1M and 2M star databases include some companion stars, which were assigned fictitious catalogue numbers, which if I recall correctly were Tycho numbers that end in 0.

If this is the case then stars with "HIP" numbers < 1000000000 and above the end of the Hipparcos range should be removed from these databases for the 1.4.0 versions.

It appears you are correct. Of the first five binaries in visualbins.stc, four of them exhibit a 3rd nearby star when using the 1million stars.dat database. Those stars are named "TYC 6669-3355-0" or "TYC 7194-3355-0", etc. They do end in zero.

I would be willing to edit stars1M.dat to remove these fictitious stars, only I find I cannot open it. When I try using MS Notepad to open it, the screen freezes. When I use MS Wordpad to open it, I get a page full of scrambled characters. Apparently, the database is not in standard text format, so I cannot edit it.

Any suggestions would be appreciated.

Thanks

Frank

[Malenfant]

The million stars .dat file - like the normal stars.dat file - is in binary format. You can't open it with a text editor.

I think there's some software that converts between the two formats, but I can't recall where it is or how reliable it is.

I don't think the guy who did the 1M and 2M files is doing Celestia stuff anymore?

[buggs_moran]

I say this while treading very lightly here... How difficult would it be to parse a new dat file for a million (and 2M) star database? If it's computer time that's required, I'll donate it. I have extras not doing anything at the moment. Problem is that they are all Windows OS.... All we would need is a simple (and I say simple as a non programmer so don't hurt me too much) read/write program, right?

[t00fri]

I say this while treading very lightly here... How difficult would it be to parse a new dat file for a million (and 2M) star database? If it's computer time that's required, I'll donate it. I have extras not doing anything at the moment. Problem is that they are all Windows OS.... All we would need is a simple (and I say simple as a non programmer so don't hurt me too much) read/write program, right?

Have a look to my contributions in this thread

http://www.celestiaproject.net/forum/viewtopic ... sc&start=0

I spent a bit of thinking already about redoing the 2M stars via PERL code. Since on Pascal Hartmann's site it was precisely documented what they did to derive distance estimates for the stars, much of their work may be reused.

As to CPU time that's no big deal, at least not for my machine, 3.2GHz/3GB of (fast CL2 ) RAM...PERL is pretty fast. Extraction of my deepsky.dsc catalog with 10600 DSO's, including searching and combining data from /6/ other catalogs takes only a few seconds, for example.

Bye Fridger