Couple of questions on various topics/feature requests

[HS]

The last few messages are referring, I believe to the add-on called h-alpha.dsc. It shows the distribution of ionized hydrogen and the text file that goes with it is signed Selden Ball and is dated 19 Sept 2003.

[Harry]

- add "background" scripts, i.e. scripts which run the whole time and independent of any normal scripts, performing some simple task (like e.g. pointing some antenna to earth) periodically.

They're usually called event handlers. You might take a look for example into javascript to see how they act (sure you know that already). They shouldn't be too hard to implement as they are part of common APIs nowadays.

What I wrote about aren't event handlers, but indeed background scripts - i.e. they run completely without events. Lua scripts in Celestia do have some kind of event handlers for handling keyboard input and the termination of scripts (think destructor).

I proceeded to implement what I had described, and it works now (though with some minor problems). The changes in the core are mostly trivial, and even lead to IMHO cleaner code. Referencing scripts from .ssc-files works, which requires that a pointer to the CelestiaCore is passed to the SolarSystemLoader. However the handling of paths is not yet correct. And allowing scripts to set the orientation of bodies is trivial too, but doesn't quite do what I'd like - now e.g. having the solar panels of hubble always point to the sun works, but using follow or sync still uses the original orientation (can't do anything about that). But it's a start, and a good part of the changes should be checked in simply because it's cleaner code than before.
Oh, and one of these days I should get a clue about Quaternions (rotations)

Harald

[t00fri]

I completely understand your viewpoint, but simply do not agree with it.

Interesting you bring up this point. I think links to external sources of information were a good thing to start Celestia with, but it's time for an internal database to take over, such as maxim has been working on. When a curious person is sent away from a program, to another program, they will stay at the second program (if it captures their curiosity) instead of returning to the first program. At least this is what happens with me.

Homebrewn databases can never keep up with the pro-level accuracy and completeness that astronomical data bases like the one in Strassbourg provide! And there is no need to care for regular updates...

If I want to really know numbers, I would always refer to sources I really can trust! This is never guaranteed, if anybody (except Grant;-)) does the job within our Celestia community...

... but we are back at toy robots, trucks and cranes, were I started.

How can you dismiss toys, so simply and completely, as being useless, non-educational things? You never learned anything from playing with toys? No mere toy ever encouraged you to learn more about something, such as why it did a certain thing, or how it worked, or anything else? Not even questions to your parents?

You are arguing entirely besides my point.

For a certain age group (kids) I value toys very highly, of course. But one should definitely not use the same learning tools at all stages of our life!

-- wooden toys, dolls, Legos and what not are perfect for Kindergarden times. They stimulate the kid's fantasy, no doubt.

-- these Kindergarden kids do not yet need Celestia, though! There is ample time for it at a later stage when wodden toys, dolls, Legos etc are out and their logical thinking starts to develop.

For these reasons, the world famous Jesuit schools use an entirely different teaching plan from usual schools:

The younger kids almost exclusively deal with subjects they are ingenously and intuitively good at: language training, music, arts,...

Only much later, towards the end of the high school period, they switch and concentrate on math and sciences in general.

Throughout, they do a lot of praying, of course;-)

-- For grown-ups and kids alike, too explicit but intrinsically primitive animations are rather fantasy killers!

Seeing the wheels turn of a Mars rover, for example, does not provide much non-trivial additional information to the observer. Whenever somebody's eyes notice wheels with a moving vehicle, our brain automatically adds the missing rotations by means of our fantasy! Do I really have to animate a plane's propeller rotation to make it "educative"?

Come on...

Why is it so hard to get these messages over to you??

-- ...and as a truly educational (and fun) enhancement of Celestia, I have proposed my modular "instrument concept" that admittedly is a lot of work, too.

I am just a little "allergic" towards calling things "educative" that essentially add to people's desire to play and have fun

Now you sound like a GRUP, from the old Star Trek series ... BONK BONK! Or maybe you are the Grinch, in disguise?

Such TV series are not part of my educational background. So I really do not understand the words you were using, sorry.

At the Star Trek age, I built my first cardboard telescope with a single 1 inch lens ($2!) and saw Jupiter's moons and Saturn's rings for the first time. At least as concerns myself, this provided vastly more "education" and enjoyment than watching StarTrek...

Bye Fridger

[Harry]

I hate to interrupt your discussion about what Celestia should and what it should not be

I've made some performance measurements with background-scripts, by writing down the FPS for varying number of bodies (and respectively background-scripts) and with different type of scripts. I tried to check the results, they should be ok (except for some rounding etc.).

Baseline is the initial scene of Celestia after start (Io), with ~40 FPS (25 ms per frame) on my system.
Then I've added a ssc file with 10/100/1000 definitions of a planet (very simple, all the same). Time per frame is
10 planets: 25,84 ms
100 planets: 30,77 ms
1000 planets: 84,03 ms
That's about 0.06-0.08 ms additional rendering time per frame and new planet.

Now the same with a trivial script, which only waits in a loop. This measures the overhead for managing and activating the scripts.
10 planets/scripts: 25,71 ms
100 planets/scripts: 33,33 ms
1000 planets/scripts: 100 ms
(about 0.07 to 0.085 ms more per frame and planet+script)

And finally a script which performs a useful operation, in this case getting the current position of two objects, computing the orientation pointing from the first to the second and then setting the orientation of the first object to the computed orientation multiplied by a fixed rotation.
In short. make the first object point to the second.

10 planets/scripts: 27,17 ms
100 planets/scripts: 45,05 ms
1000 planets/scripts: 222,22 ms
(about 0.2 ms more per frame and planet+script)


So the time to manage and activate a script is small compared to the overhead already implied by simply adding planets. The time needed to actually do something useful in a script is significant, but not prohibitive - we are talking about 150-200% additional overhead per planet/object which actually has a background script, (though for a still simple script). 10 simple background scripts would add 2 ms to each rendering cycle (on my slow system) - and you don't have to update the orientation every cycle, maybe every second cycle or every 0.1s would be enough.

I am still thinking about how to specify background-scripts. Currently I use a new property for bodies in .ssc-files, but it probably is better to introduce a new file-extension (".celb" perhaps?) and let Celestia run all files with that extension in extras as background scripts. These scripts would then run forever (unless they fail or end) and be allowed some special rights, like simple modifciations to objects (stuff like star-magnitude, orientation - I prefer not to let normal scripts do that, so you don't have to worry about whether everything is still as it should be after you ran some obscure script).

IMHO performance is "good enough" (and most importantly it has no impact when you don't specify background scripts), the code is simple and mostly done, so are there any objections against inclusion? (Chris?)

As I already wrote there are some problems when setting the orientation of e.g. Hubble, because this doesn't influence your frame of reference. Apart from this, are there any problems to expect when dynamically modifying the properties of bodies?

Harald

[don]

Come on...

Why is it so hard to get these messages over to you??

Easy Fridger ... don't pop your cork, okay?

(Sorry for all the caps, didn't want to mess with turning bold on.)

I'm 49 years old and STILL love to PLAY and have FUN ... especially when I learn new things. People are different. You and I are like night and day. This is OKAY, as it brings two different viewpoints to the table. The only point I was trying to make is that people learn things in different ways. There is no ONE RIGHT WAY.

MY preferred method of learning is by SEEING it done, and then COPYING. I do not learn well by READING. And, I happen to LIKE animation ... as "primitive" as you might think it is. I also prefer to WATCH Carl Sagan (Cosmos) and Stephen Hawking (various shows) and others ... as opposed to reading long, boring, scientific papers. This is ME, and there are many others in the world just like me.

At the Star Trek age, I built my first cardboard telescope...

Kewl! I helped my uncle rebuild his Chevy engine, built slot cars and go-karts from scratch, fixed electric devices (radios, tape recorders, motors), refinished a 14 foot fiberglass rowboat, helped my dad and his friends build us an A-frame summer home, played guitar, AND other things (1966 - 1967). And, believe it or not, I still found time to have some real FUN and watch StarTrek.

Cheers,

-Don G.

[granthutchison]

I'm 49 years old ...

Gasp.
Don, I thought you were old. Does this mean that ... gulp ... I'm old, too?

Grant

[don]

Grant,



Would it be better if I said my body feels like 80 (OLD) but my mind feels like 20 <grin>? I guess they kinda split the difference and call it a close "age".

As for you being "old", it all depends on how you feel, regardless of your clock age!

Cheers,

-Don G.

[Bob Hegwood]

Sorry Don... Sorry Grant...

Gotcha both beat by a few years. Mr. Goyette? Go have a look at my updated
website, and you'll see what I look like now.

Just thought that you and Selden would get a kick out of seeing what the old geezer actually looks like.

Just click on the flag that says "Meet the Geezer!"

Take care, Bob <snicker>

[maxim]

I completely understand your viewpoint, but simply do not agree with it.

Interesting you bring up this point. I think links to external sources of information were a good thing to start Celestia with, but it's time for an internal database to take over, such as maxim has been working on. When a curious person is sent away from a program, to another program, they will stay at the second program (if it captures their curiosity) instead of returning to the first program. At least this is what happens with me.

Homebrewn databases can never keep up with the pro-level accuracy and completeness that astronomical data bases like the one in Strassbourg provide! And there is no need to care for regular updates...

In fact I do not try do build up a complete internal database - that would be a too huge task thought.

What I'm actually doing with my info-plugin, is setting up a kind of intermediate set of real basic infos that one might not start an internet session for. Especially when having dial-on connections or being 'on the road'. Of course these also always link to the original online databases provided by default, so you will have only one click inbetween here.

The concept means also that you can use this approach for providing your own notes for yourself here - some remarks you want to remember about an object you are currently examining for example. Having Apache and mySQL running on my home network, I've started this kind of use with some php-pages providing such notepad forms. There are other kinds of usages for a php-aided Celestia. Thought this is not the typical computer setup in this community, I won't populate those page sets. It`s up to everyone to have his/her own ideas.

maxim

[Toti]

But to me it seems we are quite far away from a practicable implementation of a consistent set of animations that

-- performs sufficiently well
-- tracks the actual maneuvering data as function of time accurately
-- for all satelites/space vehicles

Well, Orbiter can mix VSOP 87, advanced 3D graphics, high resolution textures and very complex physics simulation with mesh animation. And it does run very smoothly on my fairly antique PC (as it does Celestia). In a similar way, I've heard that some Celestia users are obtaining a large number of fps, even with a lot of addons and big textures installed. So it seems that there is enough computational room for additional features. Moreover, I think that the proposed implementation is not that resource-hungry, so we should expect a reasonable performance, even in older computers.

The required extended xyz files must be huge! Because when you want to incorporate such things like movements according to the actual mission logs, the resolution in time must be extremely high!

Probably that's why Chris was asking for the nature of the orientation file:

* If it is homogeneously spaced (ie. two consecutive entries are separated by a constant delta-time) then certainly won't be very useful because of the resolution needed for short orientation changes:

Code: Select all

time         angles
0             ang0
0.5           ang1
1             ang2
1.5           ang3
2             ang4
...


Note that if some interesting rotation is done between 1.75 and 1.8 we do not find out about it.

* If it is keyframed (ie. any important change has its own entry and the intermediate positions are not actually present but interpolated) it will allow fast changes without becoming a huge file (saving access time and memory):

Code: Select all

time          angles
0              ang0
2.5            ang1
58             ang2
58.05          ang3
58.07          ang4
58.2           ang5
200.5          ang6
204.3          ang7
...


Note that with this approach, only significative rotation CHANGES are included, and small/fast variations are properly descripted without enlarging the file.

I strongly suspect that given a huge non-keyframed, high resolution orientation file, it can be processed to obtain a smaller, keyframed one that can be comfortably used in Celestia.

On the performance side: as Don G. already said, there is no need of keeping rotation and animation data for every spacecraft in memory all the time.
The input info can be splitted in 3 files: one (usual) 'spacecraft.xyz' file for position info, another file with rotation data and the third one with the "animation" script commands.
Load the .xyz for every spacecraft at the startup, since we only need to render small white dots for every craft at the correct positions. When a spacecraft gets close enough to the viewer, load the orientation and animation data and update the 3D object according to the present time.
This way, the bulkier orientation and animation data will use memory space only one craft at a time, saving a lot of resources.

Seeing the wheels turn of a Mars rover, for example, does not provide much non-trivial additional information to the observer. Whenever somebody's eyes notice wheels with a moving vehicle, our brain automatically adds the missing rotations by means of our fantasy! Do I really have to animate a plane's propeller rotation to make it "educative"?

Well, this is an extreme example. Our brains can indeed join the "missing parts" in common experience related phenomena, like wheel motion, but the intrincated stages in some spacecraft unfolding/deploying routines are more difficult to conceive.
Moreover, I have listed some other uses to the animation implementation that are obviously much more worth while than wheels moving.

I doubt that the required extensive information is available for all missions that are implemented in Celestia

This is also of my concern (see my second post, and Jack's answer to it). But after a really quick search I found some interesting (non systematic) data available.
Nevertheless, even in the case that some missions' information was not available, animation support will still be worth while, in the same manner that Celestia's superb rendering capabilities can't be fully used in some bodies because of lack of data (real/complete surface texture, correctly modelled orbits, etc.)

Bye

[don]

"Meet the Geezer!"

Awwwww, Katelyn sure is a CUTIE!

Bob, you look about as much like an "Old Geezer" as I do. ... A little grayer maybe.

Since you *think* you know what I look like, I just had to put a quick web page together for you, and other curiosity seekers <snicker> ...

http://s94992504.onlinehome.us/celestia ... s/Don.html

Enjoy!

-Don G.

[Bob Hegwood]

Since you *think* you know what I look like, I just had to put a quick web page together for you, and other curiosity seekers

<Snicker> hell...

You got me again Mr. Goyette. You ARE the MAN!

Take care, Bob

[don]

Toti,

Excellent! Love your idea.

For the keyframe file example, the "missing" frames could be extracted from the xyz file, IF it is provided by the same person (so the two files match time and position).

Think it would be good to keep this discussion going until we can actually formulate a plan of what would be necessary, both internal and external to Celestia, to make it work -- within Chris' stated limits ... "A script callback interface is the limit of what I'm willing to support. Keyframe files (like .xyz files with orientation) and animation information in cmod files ...".

Cheers,

-Don G.

[don]

You got me again...

Glad to hear it. Hope you had a good laugh Bob. My wife just about split a gut laughing when I showed the page to her.

[Bob Hegwood]

Yeah, my wife was giving me that quizzical look again when I started choking on the lemonade I was drinking.

I don't really think that she believes I'm working with Celestia when I'm in here.

At any rate, it was great fun Don.

Thanks loads for that entertainment.

-Bob

[chris]

* If it is keyframed (ie. any important change has its own entry and the intermediate positions are not actually present but extrapolated) it will allow fast changes without becoming a huge file (saving access time and memory):

Code: Select all

time          angles
0              ang0
2.5            ang1
58             ang2
58.05          ang3
58.07          ang4
58.2           ang5
200.5          ang6
204.3          ang7
...


Note that with this approach, only significative rotation CHANGES are included, and small/fast variations are properly descripted without enlarging the file.

I strongly suspect that given a huge non-keyframed, high resolution orientation file, it can be processed to obtain a smaller, keyframed one that can be comfortably used in Celestia.

.xyz files are already stored as keyframes. Some people on the forum have done work to convert Horizons trajectories with a uniform step size into .xyz files with a variable step size. It would be straightforward to do something similar with orientation keyframes.

So, I'll write the code . . . who's going to track down the data?

--Chris

[Toti]

Thank you, Don.
Yes, there is a lot of interesting issues on this feature proposal. The attitude data availability is one of the most conspicuous...
I have made a Google search for "attitude data":

Here is attitude data for SOHO:
http://sohowww.nascom.nasa.gov/data/ancillary/attitude/definitive/
The format is explained here:
http://sohowww.nascom.nasa.gov/data/ancillary/#attitude
Resolution: 1 sample/10 minutes
There is also a full resolution set as obtained directly from the spacecraft attitude registers from 1 sample/second (roll) up to 10 samples/second (pitch & yaw)!

A link to Cassini attitude info in various formats:
http://www-pw.physics.uiowa.edu/~tfa/cassatt.html

Attitude info for ACE:
http://www.srl.caltech.edu/cgi-bin/dib/rundibviewancil/ACE/ASC/DATA/ancillary?ACE_ANCIL.HDF!hdfref;tag=1962,ref=3,s=0
Data format is explained here:
http://www.srl.caltech.edu/ACE/ASC/orbit.html

A much more comprehensive volume of this kind of data can be useful for scripting animations and mission schedules, etc:
http://nssdc.gsfc.nasa.gov/planetary/galileo_probe_time.html

Bye

[don]

Are any of our famous data-manipulators hanging out here? Grant, Jack, others?

Can this data be used in the manner Chris described?

Thanks!

-Don G.

[Toti]

A few more links:

This one contains NEAR attitude data under the form of quaternions:
http://pdssbn.astro.umd.edu/NEARdb/spice/
The above sets contain also the data format definitions. A little additional info:
http://www.astro.umd.edu/~stefmcl/near_spice.html

The site http://wwwpds.wustl.edu/missions/mgs/spice.html seems to have a lot of attitude info for various mars, venus, earth and moon spacecraft, including mission events and instrument's orientation records (so scripting movement of cameras, magnetometers and other science hardware can be possible!)
Some information pages linked from that site:

2001 Space Oddysey:
ftp://naif.jpl.nasa.gov/pub/naif/pds/data/ody-m-spice-6-v1.0/odsp_1000/data/ck
Info on format (quaternions):
ftp://naif.jpl.nasa.gov/pub/naif/pds/data/ody-m-spice-6-v1.0/odsp_1000/data/ck/ckinfo.txt

MGS:
ftp://naif.jpl.nasa.gov/pub/naif/pds/data/mgs-m-spice-6-v1.0/mgsp_1000/data/ck/
Info on format (quaternions):
ftp://naif.jpl.nasa.gov/pub/naif/pds/data/mgs-m-spice-6-v1.0/mgsp_1000/data/ck/ckinfo.txt

Lunar Prospector: there is attitude information on the 'geometry' entries under each of the listed files on this page:
http://wwwpds.wustl.edu/missions/lunarp/level0.html
(There is also data format description available)

Another link to NEAR ancillary data:
http://pds-smallbodies.astro.umd.edu/NEARdb/spice/

A link explaining SPICE datasets (the 'overview' link points to a HIGHLY RECOMMENDED .pdf ):
http://pds-naif.jpl.nasa.gov/naif.html
The 'supported missions' link is specially interesting:
http://pds-naif.jpl.nasa.gov/naifmsn.html

Here is a link to an explanation of SPICE tools subroutines (that can be used to extract and coordinate attitude info):
http://www-int.stsci.edu/~jordan/naif/spice_gloss

Please note that my search was really quick. I found a lot of references to scientific papers on orientation data processing. As I wrote above, there is also individual science instruments attitude information available. Perhaps not all of this info can be accessed online, and some may indeed need special e-mail requests. But summarizing, there seems to be much more information available (again, please read the last page of the cited .pdf).

What do you think?

[Guest]

HI everyone. I've not been keeping up with the forum this past week and I just discovered this thread.

I am using Celestia extensively in education, perhaps more than anyone else on the forum. I am teaching 160 high school students Astronomy, and Celestia is our window into space. To date, my students have spent over 14 hours in Celestia space, following the choreographed journeys that I have written for them as "Activities", and a few of the scripted journeys that have been put onto the forum. Emphatically ... absolutely .... without any question .... I can attest that my students love Celestia and where it has taken them. However, they are also teenagers living in a high tech, video game kind of world. As I watch them travel to various worlds, the places that they linger most on ... the places they really love to see ... are the places that have movement and action. Spacecraft drifting through space, asteroids tumbling silently on chaotic rotational movements, planets rotating on their axis, moons swirling around, pulsars spinning at high speed .... that is what really "turns them on" the most.

I would be delighted to see animation come to Celestia. Spacecraft with blinking lights or components rotating or opening would be just spectacular from an educational standpoint. Nebula gases swirling around or some kind of animated "impact" event would be so dramatic. Imagine seeing comet Shoemaker-Levy actually impact Jupiter, leaving a growing dark wound on the surface, or watch Spirit actually move across Gusev crater! I want to do an Activity showing how our moon formed from a possible collision with Orpheus. Imagine showing that collision as an animation! WOW! Volcanoes spewing lava on Io ... AWESOME! Lightning bolts on the back side of Jupiter flashing in the night ... COOL!

I do not consider such exciting animation as trivial or unimportant or childish. In fact, from my perspective, Celestia is the most powerful educational program I have ever seen. It brings the universe to LIFE. I will be showcasing it before 2,000 teachers in Viginia this summer and I want them all to use it ... to give their students the opportunity to go into space, not just sit in some boring seat in some boring classroom to discuss space and Astronomy. Virginia alone has over 500,000 kids that take some facet of Astronomy in either Earth Science or Astronomy courses. I would not be teaching Astronomy without it. I want to see hundreds of thousands of those kids, young adults and us oldies (I'm 56 for God sake!!!) around the world take the most visually stunning journey into space that we can give them. Animation will make it even more exciting for such an audience. I urge you all to work toward that goal as soon as possible.

Sorry if it sounds like me pontificating, but I have been hoping for animation for a long time, and I wanted to toss in my two cents here. PLEASE ... let's go for it.

Regards,

Frank