Available for download - Asteroid Maker 1.6

[bdm]

Summary
Version 1.6 fixes a similar bug with the asteroid periods. Occasionally the period would be incorrect.

This bug did not affect users who are not using Hilda and Trojan orbits.

Discussion
What happened here with these two bugs is a result of leftover code that I have hopefully now removed completely. If this is true, the only cells that should reference column S on the Calculations sheet are the cells in column T of the same sheet.

An early version of Asteroid Maker used a random value to determing the orbit class for the asteroid, and then I changed the way that the orbits were determined so I could include Hilda-class orbits. To facilitate this change I made a small change to the way orbits were determined, inserted an extra cell that translated the random value to a code for orbit class (MB, TI, TO, HI, HO), and then used the orbit class code throughout. I overlooked two columns when making this change.

The crucial bit here is the inclusion of a new method of determining the orbit class. With Trojan orbits, I simply generated a random number, and made the lowest values one class of Trojan and the highest value the other. When I included Hilda-class orbits, this method would not work efficiently, so I moved one class of Trojan from the top of the range to the bottom, so all special orbit classes were generated with low numbers. This change facilitates the future inclusion of new orbit classes.

The mechanism for the bug should now be obvious. The unchanged cells used the old formula for orbit class whereas all other cells used the new formula. When a sufficiently high random number was generated for the orbit class, it fell into the old range for orbit class, and output an incorrect value. Similar incorrect values were output for the reverse case but these were harder to spot.

Now the orbit class code is used throughout, this bug should no longer occur.

The affected fields for this class of bug were:
* Mean Longitude (fixed in 1.5)
* Period (fixed in 1.6)

[ANDREA]

Well, the 1.6 release solved my problems, and here are 400 nice asteroids around Tatooine:



Why Tatooine? Well, it's a nice texture, moreover it's light in color, so darker asteroids are shown clearly on its surface...
To be sincere, it's as alternate texture of Sol/Pluto.
I know it's an astronomically impossible environment, but sometimes the public needs "beautiful" scenes, that can take up their attention, before attempting to speak of science.
I use this method to wake up people, if they are not particularly interested on the matter, and IMHO it works.
Bye and thank you, bdm

Andrea

[Cham]

Fun !

What is your FPS with this scene ?

[ANDREA]

Fun ! What is your FPS with this scene ?

I've just checked it, and obviously it's VEEERY slow when I show all the 400 asteroids:
it's 8-9 fps.
Not fantastic, I know, but nevertheless it's very nice to see the chaotic movements of asteroids and their moons (many moons, I forgot to say), even more nice if I choose a slow (500-1000) time increase.
On January I'll have my new 6800GT 256 MB Nvidia card, and then things will go definitely better, I'm sure.
Bye, Cham

Andrea

[bdm]

So that's what Pluto looked like before the debris coalesced into a single moon.

[symaski62]

Hank,

The spreadsheet is in three "sheets". SLK only supports a single sheet.

When saving each sheet individually in SLK format, Excel 2000 generates an error message saying that features may be lost.

At any rate, they're temporarily available at
http://www.lepp.cornell.edu/~seb/celestia/files/asteroid-maker-slk.zip (900KB)

^^ Openoffice.org 2.0 finale (SYLK) file.slk

[piellepi]

Hi bdm!
I downloaded version 1.6 of your add-on but there's something wrong in the xls file.
When I run it, Excel 2003 (running in Windows XP Professional) opens a messagebox saying that (I translate from italian...) there's an "error in the file with possible data loss".
Clicking OK on the messagebox, four sheets open regularly, but if I close Excel, without doing anything, Excel says that the file has changed and then asks me to save the new version: so it's evident that there was effectively a data loss somewhere and I cannot expect that the program works perfectly...

I hope this can help you!
Bye
Pierluigi

[ANDREA]

Hello bdm, sorry but I have another problem: I wish to make all the 400 asteroids and their moon of a reddisch dark color, e.g. like Amalthea, so I used this moon's color, that is 063 045 035
Just to make things a bit differentiated, I slightly modified the numbers, about 5% plus and minus from the Amalthes's ones, and checked results.
WOW! Nothing has changed, asteroids are various shades of grey, yellow, cream, pink and so on, but absolutely no one with Amalthea's dark red color.
Where am I mistaking?
I need that color, because I'll put the asteroids around Geonosis, together with Obi-Wan Kenobi Starfighter and HyperRing, as in StarWars II.
Can you help me, please?
Thank you

Andrea

[bdm]

Buongiorno,

piellepi - I believe Excel has that warning whenever the contents of a spreadsheet has changed. The Asteroid Maker is full of RAND() function calls, and RAND() will calculate a new value whenever it is called. This means anything that will cause Excel to recalculate will cause the contents of the spreadsheet to change. I think it's safe to ignore the warning but I want to experiment a bit to see if that is the case. Just make sure you haven't made any real changes before ignoring that warning!

ANDREA - I think the "color" setting is the colour that a body will have when it is so small that Celestia renders it as a few pixels. It won't change the colour of a body when it is visible as a disk, because the rendering of that disk comes from the texture attribute and not the color attribute.

[ANDREA]

ANDREA- I think the "color" setting is the colour that a body will have when it is so small that Celestia renders it as a few pixels. It won't change the colour of a body when it is visible as a disk, because the rendering of that disk comes from the texture attribute and not the color attribute.

Hello bdm, this means that to change the asteroid colors I should have to modify the .cmod files?
If yes, what software can I use to do it?
Sorry for trouble, but as you see I'm very interested to your work and would like to obtain the results I need.
Bye and thank you

Andrea

[piellepi]

Hi!
I'm obviously sure that I have not done any change before quitting Excel!
I execute these steps:
- I run the xls file (double click on the icon)
- I press OK button in the messagebox
- I click the red "x" to close Excel
- now the message appears, telling that data has changed.

Now if I start a new copy and recalculate all (pressing F9, just after the message), in the sheet "Celestia" the cell A6 contains a "#VALUE" error , so as the cells A84, A162 and so on, up to cell A21144!!!. Other cells instead seem to be ok (without any error), apart from tons of "#BLANK" cells
Do you think it's possible to save the sheets in plane textual format, containing just formulas and data? In such a manner it could be easy to spot when and where there is a "loss of data", simply comparing the original file with the new one.
I'll investigate further...
Thanks
Pierluigi

[ANDREA]

I'll investigate further...Thanks Pierluigi

Hello Pierluigi, have you tried to download again the file?
Sometime files are corrupted, and in this case the error may be similar to the one you are obtaining.
My little (Euro) cent.
Bye

Andrea

[ANDREA]

Well, following bdm's suggestion, I modified the asteroids color, and here is the result, around Geonosis:



Hope you like it.
Bye and thank you, bdm, for this possibility.

Andrea

[bdm]

If the cell A6 and so forth on the Celestia sheet are the only cells affected, the error can be ignored. Here is a sample of what normally appears there:

Code: Select all

# Asteroid 1 - Mass: 4.4679e+20 kg.  Class: Normal.  Moons: 0.

It's only a comment, and won't affect the Celestia rendering.

However, it may be a good idea to investigate the cause of the issue further.

[ANDREA]

Well, another step for Geonosis and StarWars II:



Next will be Obi-Wan Kenobi Starfighter and HyperRing.
Bye

Andrea

[Malenfant]

Now what would be really cool is if you could make rings that actually WERE zillions of particles - though that'd probably screw up the FPS rate even on a computer a huge as IBM's Blue Gene/L .

Oh well, give it a few hundred years before we get that sort of thing on our desktops...

[buggs_moran]

Now what would be really cool is if you could make rings that actually WERE zillions of particles - though that'd probably screw up the FPS rate even on a computer a huge as IBM's Blue Gene/L .

Oh well, give it a few hundred years before we get that sort of thing on our desktops...

I don't know, you might not get the awful FPS you would think if you did it right. Think about the 2 million star Hipparcos database for celestia. With all magnitudes visible it can tax my computer, but it still works. Same with Don's new Earth textures, when I look at the whole globe with all the detail, FPS drops. However, if I move in to shuttle orbit distance, everything is smooth as silk. It would depend strongly on detail/visibilty tweaking for distances. You would only see thousands at one time if you were in the rings. This goes back to the posts on variable mesh level details at different distances...

[piellepi]

Hi bdm!
I found the error in the cell A6 and so on!
I am astonished!
In the formula, where you specify the exponential format ("0.0000e+00"), the letter "e" MUST BE CAPITALIZED! Try "0.0000E+00" and magically the error disappears! (for italian people, in particular for Andrea: non ho parole! )
I downloaded again the file, but obviously the file is exacly the same and the problem remains.
I investigated further and I found that when excel loads a file saved with a older version (for Excel 2003, other "Excels" are older ones...), it recalculates all the formulas: perhaps it finds some errors so it shows the messagebox...
The problem is now: is it possible to inhibit this recalculation in order to do it manually? In Options you can check "Manual" instead of "Automatic", but can you set this option OUTSIDE the file you have loaded, as a default? Remember that I want to recalculate manually the file after I load it!
Ciao!
Pierluigi

[bdm]

Hi bdm!
I found the error in the cell A6 and so on!
I am astonished!
In the formula, where you specify the exponential format ("0.0000e+00"), the letter "e" MUST BE CAPITALIZED! Try "0.0000E+00" and magically the error disappears! (for italian people, in particular for Andrea: non ho parole! )
I downloaded again the file, but obviously the file is exacly the same and the problem remains.
I investigated further and I found that when excel loads a file saved with a older version (for Excel 2003, other "Excels" are older ones...), it recalculates all the formulas: perhaps it finds some errors so it shows the messagebox...

Thanks for finding that. I have limited access to Excel so it would have been difficult for me to track down.

I developed the spreadsheet in OpenOffice, and it appears OpenOffice is more forgiving of scientific formatting than Excel is. The upcoming 1.7 release should have this change in it (it's not difficult to make).

Also coming in 1.7 will be support for Kuiper belts and limited help text attached to the most important cells on the Constants page. The help text will take a little time to write so it's not an imminent release.

The problem is now: is it possible to inhibit this recalculation in order to do it manually? In Options you can check "Manual" instead of "Automatic", but can you set this option OUTSIDE the file you have loaded, as a default? Remember that I want to recalculate manually the file after I load it!
Ciao!
Pierluigi

I don't know how to do this in spreadsheets in Excel format. I'm not sure it's the right thing to do because other people may prefer different settings.

[Guckytos]

Hi bdm,

just a small suggestion. Since you developed it in OpenOffice, could you perhaps also release it as OpenOffice Calc dokument?

So some errors from converting to Excel wouldn't pop up for people that have OpenOffice.

Just my 2 cents.

Otherwise, keep up the good work.

Regards,

Guckytos