Mostly Harmless 0.2.0 System Generator

[marc]

Here it is, another milestone.

I've finished modifying stargen and celestia so that they respectivly output to, and read from a MySQL database. The database is now a complete storage replacement for stc files, stars.dat and starnames.dat and an alternative storage method for ssc files.

This means that I can place a randomly generated solar systems around every star. My first version that used ssc files would use 10gig for about 100,000 stars, this database version only uses about a tenth of that.

The generator also add a few extras to base output of stargen:
Ring systems. (for now all jovians have rings, I will make them less frequent later)
Large jovians are converted to brown dwarfs.
Small objects are assigned asteroid meshes (Jestrs).
Oblateness is calculated for jovians.
Textures are assigned according to hydrosphere and cloud coverage,

By default only the inner 2000 systems are generated. There is a script that generates systems for all the stars which takes about 6 hours.

I realise that many stars should not have any planets, later after a bit of research I will do some serious culling. I'm open to any suggestions on what type of stars should not have planets.

Thanks to Rassilon, Don Edwards, jestr, and Praesepe for most of the textures.
If I have missed anyone let me know.

Unlike the SQL addon, the installation is independent of Celestia and will install to its own program directory. I have not changed any of the registry code yet so the options you set in Celestia will be read by MH and visa-versa.

I have only been able to upload the lores version which is 60Mb. The highres version is about 300Mb so I probably won't release it until Mostly Harmless matures a bit more.

There are a few orbital cities around earth which I created with anim8or, the textures for these are a WIP.

For more information on what I've done and what I plan to do have a look at the changelog.
http://mostlyharmless.sourceforge.net/m ... c.php?t=29

And here are the credits, (still a work in progress)
http://mostlyharmless.sourceforge.net/m ... c.php?t=32

and of course a link to the download.
http://prdownloads.sourceforge.net/most ... e?download

Now the next stage is ditching the existing user interface and making my own one using Plib, wish me luck!

[Psykotik]

Hello Marc,

Is that an update to the sql addon 0.1.4, or is it something totally new ?

[selden]

Marc,

Luck!



Highly luminous stars seem unlikely to have large planets, if any at all: their accretion discs would have been blown away quickly by the light pressure. Systems around stars of other types that formed too near highly luminous stars probably would have suffered the same fate. "Rapid" evaporation of planetary discs has been seen in several nearby gas clouds.

[marc]

Hello Marc,

Is that an update to the sql addon 0.1.4, or is it something totally new ?

Both really, All the sql features are still there with the addition of randomly generated planets and moons which are also stored in the database.
The base data for the generated systems comes from stargen.
http://www.eldacur.com/~brons/NerdCorne ... arGen.html

I have added several tables that store the equvalent of ssc data, and have modified stargen so that it will push its output straight into these tables.
In order for celestia to read this data I made a new class which reads the systems from the database instead of the usual ssc files.
I've done this system generator before but it was based on gzipped ssc files which were too in-efficient with storage space, and difficult to edit once created.

Selden.
How near and how luminous would a bright star need to be to have an effect on an accretion disk? 20ly for a III luminosity class? Even if I only had one or two examples I could make some up rough formula. I imagine that light pressure would be directly proportional to the luminosity and it would decrease with the square of the distance.
Using such a formula I could give all the stars a 'no planets' and a 'reduced planets' range then do a series of deletes according to these. I've already played with range queries and they a fairly fast, so processing every star and the planets in range of each shouldn't take too long. Hopefully there will be something left for the alien species .
Do you have any links to papers that I could have a look at?

Marc.

[selden]

crud. I just lost my editing session.

briefly, See the AAS protoplanetary sessions.

e.g.

http://www.aas.org/publications/baas/v3 ... 2/S120.htm

http://www.aas.org/publications/baas/v3 ... 2/S240.htm

There probably are other sessions, too.

[maxim]

My first version that used ssc files would use 10gig for about 100,000 stars, this database version only uses about a tenth of that.

That's still a huge amount of data. About 10kb per star. I would expect about 32 Bytes per star. So you still can reduce by factor 300.

maxim

[marc]

This is including the data to define all the solar systems (about 50 bodies per star). The star table by itself only uses 60 bytes per star.

[marc]

Thanks for the links Selden

[maxim]

This is including the data to define all the solar systems (about 50 bodies per star). The star table by itself only uses 60 bytes per star.

Ah, ok, I hadn't got that. A very populated universe

maxim

[brunetto_64]

59,6 megs!!!! I will oblige to buy the adsl line!!!!!

[marc]

The lores already uses 512x256 textures. If I can reduce jestrs asteroid meshes by 25% then I can halve the size of the lores installer.
Does anyone know of any command line tools that can reduce the poly count of 3ds files?
I have had a quick look but havn't had much luck.

[jestr]

marc,what's the names of the meshes you are using,I can try to edit them,how many polys would you like ?Jestr

[marc]

Thanks Jestr.
I'm using all of them I hope , but I don't think that it is worth the effort of editing them. Ideally I would like to be able to reduce all the meshes (space stations ships etc) using a script at build time. I will do a similar thing with the textures (using imagemagik). Then we only have to maintain high res/high poly copies of everything, which is less work in the long run.