predicting stars with random star generator using a scientific method based on empirical facts.
It's probably worth noting that "predicting stars" is not Fridger's aim.
The aim (correct me if I'm wrong Fridger) is to produce a realistic "facsimile
" of the gross morphology of these deep space objects, such that the proportion and distribution of different classes of stars in the overall population of a certain type of galaxy, for example, reflects the reality we would observe. (e.g. Typically, what proportion of stars in a particular type of galaxy are red-dwarfs and how does the evidence suggest that they are usually distributed within that type of galaxy?)
So the aim is not to "predict" individual stars, but to reproduce their approximate distributions spatially according to the evidence we have. With this aim in mind, it is irrelevant where an individual star is situated exactly, or whether indeed it exists at all. What's important is that the entire population and distribution of that type of star within a galaxy, etc... is determined by the correct statistical technique for the circumstances, whether that be a modified gaussian distribution or some other distribution altogether, according to what we know about their distribution from observation and other evidence.
The hope is that by doing so the appearance from a distance of deep-space objects in Celestia will closely match their real-life appearance
At the end of the day, there are many other factors involved, such as how well do the star-shaders replicate the color and luminosity of each star and it's contribution to the whole, and how good is the quality of the monitor you display it on, etc, etc..., so some compromises to absolute accuracy are inevitable.
Nice summary! Thanks, CC.
One might add an important feature that needs to hold if everything has been done correctly:
After random generation of a certain population of stars, we may count the generated numbers of stars placed in bins of e.g. (absolute) visual magnitude M_V, V-I color index, mass, temperature, distance etc. The resulting number distributions like d N_stars / d M_V, ... must be identical to the respective experimental data we started from!
Here is a concrete example (taken from viewtopic.php?f=11&t=527
It also illustrates the important role played by the Luminosity Function d N_stars / d M_V
in the context of the random generation of globular cluster stars
Here are the measured data of the Luminosity function for two prominent globular clusters NGC 7078 (M 15) and NGC 6341 (M 92) along with a respective least-square fit done with Maple some time ago (blue curve).
[Click for a bigger image]
LF_gc.jpg [ 30.99 KiB | Viewed 893 times ]
After many thousand globular cluster stars were randomly generated with this probability function, here comes the mentioned crucial cross-check
Place the generated stars of the globular with absolute magnitudes between (M_V - w) and (M_V + w) in bins of suitable width w (e.g. 0.25), and plot the corresponding star counts you find in each bin versus M_V. The shape of the resulting distribution should be identical to the normalized Luminosity Function
Here is the amazing result from celestia.Sci
[Click for a bigger image]
LF_gc_stars.jpg [ 31.51 KiB | Viewed 893 times ]
As you see, the agreement of the actual bin distribution (dark blue) with the above fit to the measured Luminosity Function (bright green curve) is just perfect
as it should be...This check warrants that in each of the rendered globulars the brightness distribution of their thousands of gc stars follows precisely the measured Luminosity Function!