Stars in database, outside Milky Way ?

[Cham]

I've found a bunch of stars which I believe are part of the new stars.dat file (to be confirmed), and which lies very far away outside the Milky Way :

HD 191738
HD 195046
HD 196379

Can someone confirm this ? Is this a bug or what ?

[ajtribick]

These stars ought to have been automatically rejected from the final dataset for the reason of low parallax and parallax error > parallax.

I'll take another look at the generation code.

Thanks for the report.

---

EDIT: found a typo in the rejection code, fixed it.

[LordFerret]

This may or may not be true, I'll let more qualified minds determine it..... but when I went and visited each of these stars in my Celestia install, I found them all within the Milky Way. There were even additional stars around each of them, stars which are in Pascal Hartman's 2.5 mil star database (which I have loaded). If you would like I can post screenshots.

[ajtribick]

Yes these stars are in the OLD version of stars.dat, which was based on the original version of the Hipparcos catalogue (which I will refer to as HIP1). The 2.5 million star database is also based on HIP1.

Code: Select all

star        HIP1 parallax   HIP2 parallax
-----------------------------------------
HD 191738     1.60 ± 0.88     0.03 ± 0.68
HD 195046     0.81 ± 0.70     0.03 ± 0.66
HD 196379     0.45 ± 0.47     0.03 ± 0.25

As you can see, the more careful data reduction has resulted in these stars ending up with very dubious parallaxes. The HIP1 measurements had very high errors on them anyway.

In general the new reduction (HIP2) is much better, e.g. the distance to Beta Centauri now agrees with the determination from binary star orbits.