[Kstars-devel] Compressing the binary data

James Bowlin bowlin at mindspring.com
Sat Jul 12 00:48:11 CEST 2008


On Fri July 11 2008, madpenguin8 at yahoo.com wrote:
> Why not split it up into several files containing different
>  magnitudes? Users may never need magnitudes that are more than the
> magnitude limit of there equipment. Plus you wouldn't have bulky
> files, just a few of reasonable size. Just my $0.02 anyway. :)

The plan is to have several different files available for different
limiting magnitudes.  The 74 Meg file is the smallest (for now).  The 
next file will be five (or ten?) times bigger and then the next file 
will be ten times bigger again.  Expect some *very* bulky files to 
become available.

Early on in the design process we were thinking of splitting the data
up into 512 smaller files, but the split would have been by location in
the sky, not by magnitude.  Putting it all in one file made the code 
simpler and probably faster.  Splitting up the stars by magnitude would
seriously degrade performance.

There are longer range plans (i.e. probably not this summer) to use
multiple spatial meshes for displaying stars based on magnitude and
zoom level.  This should improve performance especially at intermediate
zoom levels.  But for now, the plan is to have all the stars in just
two files: shallowstars.dat and deepstars.dat with the vast majority
(> 98%) in deepstars.dat.  This is not optimal but we hope it will give
us a quick path to greatly extend the star catalog.  

-- 
Peace, James


More information about the Kstars-devel mailing list