Additional universe information patch completed

[Rei]

Also, might I add, I just tried this:

http://www.nineplanets.org/
-------------------------------------
An error occured while loading http://www.nineplanets.org/:

Could not connect to host http://www.nineplanets.org
-------------------------------------

Needless to say, that never happens with a 'verbose' feature.

Also, if SIMBAD has a database interface, and you really wanted, I could code a couple queries in that get the information for display. I have experience in Postgres and Oracle.

[Evil Dr Ganymede]

Fridger:
For example, would you like to join in our conversation about how, given the mass, orbit, and star details of an extrasolar planet, to get as realistic of a radius as possible, since the current gas giant formula doesn't accurately represent large gas giants or those in orbits close to stars? My last proposed formula was to use the distance from the star and its luminosity to approximate how much radiation the giant is getting, and use that as a logarithmic scalar (since heat can "inflate" gas giants) to a formula that approaches a radius of 80,000km as the mass of the gas giant approaches infinity. The whole formula would be curve-fit to match our solar system's giants and the known extrasolar radii. Certainly, without knowing how oblate the planet is, what it's chemical composition is like, and other factors, there's no way that we can be perfect. But we can do our best to be as close as possible, now can't we?

Check out the papers by Adam Burrows. Um.... *rummage, rummage*
(the trick is to find a preprint paper that you don't have to bloody well pay a fortune for. Damn those journals!). Aha.

The Theory of Brown Dwarfs and Extrasolar Giant Planets, Burrows et al., Reviews of Modern Physics, volume 73, 719-765 (2001).
URL: http://zenith.as.arizona.edu/~burrows/p ... -final.pdf

It has formulae linking age, luminosity, temperature, and mass of brown dwarfs. Dunno if you can apply it to superjovians, though it does have graphs showing how the temperature evolves with time. That might be elaborated on in:

"Theory of Giant Planets," (Burrows A., W.B. Hubbard and J.I. Lunine), Annu. Rev. Astron. Astrophys., 40, 103 (2002).

But I can't look at that because of the journal access policies. :/

[stonedyak]

On to my next project: enabling a new class of object that, now that we have masses, can "free-float" around, being tugged on by the gravity of the N nearest/heaviest objects.

Have you seen the 'Mostly Harmless' project? (http://mostlyharmless.sourceforge.net)Its a game in progress based on Celestia that implements gravity for spaceships etc.. It might be worth having a look before you start on this.

I've written a really nice looking terrain generator before, and I plan to adapt it to Celestia planet surfaces

Wow, that'd be great! This feature is requested quite often, but the general response is 'maybe in a couple of years' , since it will take quite a bit of work to do.

[Rei]

Ah, thankyou thankyou thankyou! That's just what I need. For age, we at least have an upper bound (I did program star lifespan tables into the infopatch), but at least that's something. It looks like this would translate best into a table sort of like star temperatures had been doing, and I had star lifespans and masses do. Thanks, I'll try and read over this next weekend, and take it into account.

This looks just perfect!

[t00fri]

Fridger:

Here's a google image search for "celestia"

http://images.google.com/images?q=celes ... 59-1&hl=en

Notice that almost all of the images that haven't been cropped not only have information on, but have verbose information on. People like this feature.

I also had a look and found only a couple or so of celestia images with verbose=ON. This is statistically irrelevant as you should know. Some people might just be unaware how to switch the info off;-)

I for my part, never use the verbose info, since I know very well that much of the given info has not yet been cross-checked. I never use info of any kind that I cannot rely on. I simply cannot afford to do this!
Don't forget Celestia is still young, but already exists for a number of platforms and the number of developers is quite small...Be assured, we get to that!

You don't take out/weaken features that people like with software, that's generally a bad design philosophy.

This may be a valid rule for producing commercial software. Celestia is NOT.

It's not a lack of knowing about the URL feature, it's that people want to use Celestia, not a web browser. If they wanted to use a web browser, they'd pull up a web browser and go straight to what they wanted. Often times, websites are slow (last night, I couldn't connect to nineplanets at all). They block your screen, which would otherwise be filled with beautiful Celestia views. They're inconvenient. There are dozens of reasons why that wouldn't be wanted.

Are you seriously denying that sophisticated software, notably commercial one, is mostly using WEB-browsers as the interface for their integrated help system?? If people would not like this, commercial firms would not use this interface in so many cases!

Also, displaying Celestia's internal information will be misleading to people. For example, consider the radius. What radius does it report for, say, Amalthea? Amalthea is not a sphere. Talking with people, I've determined that Celestia is supposed to use the maximum radius. How do people know this?
Minormoons.ssc, a popular addon, uses the average. If people interpret the radius of Amalthea as being real, it'll give a volume over 3 times higher than its real volume. BTW, any clue why the radii on solarsys.ssc are all smaller than those that NASA lists as the maximum, except for one which is larger than the max? In my last mail to Chris, I suggested that I make it so that we store the 3 axes' max dimensions, and then have getRadius return the max, and a new function - getRadii - return values for all of them, so that we can display all of them.

Rei, everyone would be very happy if you were ready to invest your time to work towards improvements of radius definitions for nonspherical bodies, eliminate further bugs in parameter values etc. Just to make Celestia more reliable..

But, please......., approach the problem more gently, take your time, think about how your modifications would fit into the general design goals of Celestia etc.
In no way will Celestia be transformed "within a few weeks" into a spaceship simulator with which you can land on Olympus Mons to watch the sunset ....That's what is in your mind if I am correct?;-)

For example, would you like to join in our conversation about how, given the mass, orbit, and star details of an extrasolar planet, to get as realistic of a radius as possible, since the current gas giant formula doesn't accurately represent large gas giants or those in orbits close to stars? My last proposed formula was to use the distance from the star and its luminosity to approximate how much radiation the giant is getting, and use that as a logarithmic scalar (since heat can "inflate" gas giants) to a formula that approaches a radius of 80,000km as the mass of the gas giant approaches infinity. The whole formula would be curve-fit to match our solar system's giants and the known extrasolar radii. Certainly, without knowing how oblate the planet is, what it's chemical composition is like, and other factors, there's no way that we can be perfect. But we can do our best to be as close as possible, now can't we?

I would think that, given your background, you could contribute greatly to such a discussion.

Of course I could and others among us could as well. But some of us believe that it is a better policy to refer to globally accepted "standard data" rather than homebrewn ones, the errors of which are hard to assign...

Bye Fridger

[Rei]

Have you seen the 'Mostly Harmless' project? (http://mostlyharmless.sourceforge.net)Its a game in progress based on Celestia that implements gravity for spaceships etc.. It might be worth having a look before you start on this.

I'm actually writing an email to Marc right now. I'm noticing in their change log that one thing on the TODO is "reimplement gravity".

Concerning terrain generation, though, just the controversy there's been over an infopatch makes me wonder whether it's worth adapting my terrain generator to Celestia at all. I mean, if I were to make that, would there be a huge controversy over the fact that the data between the known points on bumpmaps is made up by a fractal terrain generation algorithm? And that on planets without bumpmaps, it's only made from generation parameters? As I mentioned before, it's really discouraging to think about. I may just finish the infopatch here (if they're even going to use it at all...) and then head over to MH to develop there if Marc wants. I work in graphics (yes, Fridger - do you want screenshots of programs that I've developed, since you decided to cast aspersions about that in your last email simply because the sysadmin didn't restore my graphics driver the last time that they restored this system, as if that's my fault?), and like to program, so I thought something like this would be good to work on...

For those who want to see what my terrain generator looked like, visit:

http://www.DaughtersofTiresias.org/imag ... shot01.jpg
http://www.DaughtersofTiresias.org/imag ... shot02.jpg
http://www.DaughtersofTiresias.org/imag ... shot03.jpg
http://www.DaughtersofTiresias.org/imag ... shot04.gif

(for those of you curious about the titles... the first 3 titles are from when I was in a wierd moon; the last one was when I was considering making a freeciv opengl terrain)

(about the water - that was due to a bug on the opengl driver for the laptop that I took the screenshots on; on all other systems, you don't get that "choppy" effect. It's actually just one polygon in the horizontal plane).

and, for Fridger, some screenshots of the prototype of the 3d engine that I wrote for work:

http://www.DaughtersofTiresias.org/images/brain/

It's a high performance voxel-based (not polygon based) 3d engine that doesn't use any acceleration and outperforms all of the voxel rendering engines that we've tried that use polygons on accelerated cards. And yes, I've written polygon engines too. And can get you screenshots of anything else that I mentioned to you that I've written.

- Karen

[Rei]

Fridger:

Lets look at non-cropped images for the first 10 pages of images (an arbitrary number), and see what we find:

None: 0
Terse: 9
Verbose: 6
Can't tell, but either Terse or Verbose: 2

Of the 17 identifiable non-cropped Celestia images, not a single one had turned information from the default option of "terse" to off. Of the 15 for which I could tell whether it was terse or verbose, 2/5 of those people had taken the time to turn it *on*. But, this is the feature that you want to make always *off*. I have trouble understanding this. Say "it's not statistically significant" if you want. But where else will you go to see what users are using?

I mean, of course you have the right to do whatever you want with it - it's your code. But even in open-source free software, you still don't want to upset your users. Do you think that Linux got to where it is by taking out features that people liked? Or any open-source software? Again, it's your call, I just have trouble understanding why you'd remove an option which people like, which is just that - an option.

Concerning web browsers, did you see me complain about using a web browser for help? I think that it's *good* that Celestia has a web browser for help. But, you'll notice that you don't see web browsers coming up for much apart from help. That's what tooltips and info boxes are for. That's what we have here. I did UI for a surveilance project at Rockwell-Collins. If I had made our customers pull up a web browser every time they wanted to know what modulation a signal was or what the peak signal strength of a certain frequency was, they'd have been throwing things at us.

Concerning capabilities, why not give Celestia the *capability* to let people do whatever they want? Why remove capabilities that are given to you for free, by people who want to contributute, for which, if you'll read this message board, several people have been wanting? That's what I really don't understand.

Concerning your conversation about standard data, tell me: What is the "standard data" for the radius of HD 12661c? What is the "standard data" for the radius of HD 179949b? What is the "standard data" for the radius of HD 190228b?
(I could go on through 95% of the extrasolars).

The answer: There is no standard data. And that is the key: we need this information - try drawing a sphere without a radius. But it doesn't exist. And the question becomes, is it better to make up arbitrary values, or to have a systematic method of calculating it as closely as possible?

[t00fri]

I work in graphics (yes, Fridger - do you want screenshots of programs that I've developed, since you decided to cast aspersions about that in your last email simply because the sysadmin didn't restore my graphics driver the last time that they restored this system, as if that's my fault?), and like to program, so I thought something like this would be good to work on...
- Karen

But since you are pulling all this out of the email privacy (not my fault), this argument does not apply to your home machine, where you are your own sysadmin;-) are'nt you?

I am not doubting anything, Karen. I just try (in vain?) to moderate you to get down to a "reasonable pace", to make you start thinking about how to fit your interesting ideas into the existing framework without radically changing it. That you try to think about the excellent teamwork we are having here so far and what you could contribute without sabotaging this spirit...

Bye Fridger

[t00fri]

Fridger:

Lets look at non-cropped images for the first 10 pages of images (an arbitrary number), and see what we find:

None: 0
Terse: 9
Verbose: 6
Can't tell, but either Terse or Verbose: 2

Karen:

Please, I accept that you tell me you like to be as verbose or terse or whatever you want;-)

But don't use 2 or 6 or 9 events to "prove" anything statistically to me!

We have about 900 registered users here. Why don't you make a poll and ask everyone, how often they switch on the /most verbose =terse?/ info deliberately and watch the results???
Find out, how many people are regularly looking at the temperature of a particular star to which they just chose to travel?

Those people, who are really interested in those values also want to rely on them, scientists for example. They will certainly get these values from OTHER sources than Celestia. You bet.

Also after your patch had been incorporated;-)

Bye Fridger

My guess is you will be surprised.
Bye Fridger

[Paolo]

To Rei

I want to spend some words about the terrain question.
Your implementation of a terrain generator looks very beautiful but seems not adequate to support the very large scale variations needed by Celestia.
I think that you can find more inspiration on these two websites:

- http://home.comcast.net/~s-p-oneil/
- http://drtypo.free.fr/

On the first I think that you will find the source code too. It manages LOD very well.
Regarding the second Dr Typo was involved in Celestia discussions some time ago, but is a while that I don't see new posts from him. It uses bump image as source for the elevation data.

Merging these projects with Celestia is extremely complicated, a very-very efficent LOD mechanism must be implemented to guarantee a sufficent framerate. Probably the whole Celestia rendering engine will need to be redesigned and rewritten.

Bye - Paolo

[selden]

Karen,

A quibble (which I do a lot of ) -- Celestia does not have an HTML help system. It only has a "display plain-text file" help system.

Celestia does spawn a browser, but that's in order to acess astronomical databases for information about the selected object. That is not at all the same as help for Celestia itself.

There is a version of Celestia's User's Guide that's in HTML format, but that's completely separate and not directly readable from Celestia as it is right now. It's easy for someone to define an object which links to it, but, again, that's not quite the same. Also, the User's Guide is a manual. It's not designed to provide help (e.g. it isn't even indexed).

(Personally, I detest HTML help systems. I much prefer plain-text. I want to be able to quickly search all of the help and doc files for the information that I want. HTML files are usually broken up into tiny pieces that are only searchable individually and/or you're dependant on the author's choice of what to index.)


I want to make another comment, this time with regard to surface texture generation, but that'll take a while to compose, so I'll just post this for now.

[t00fri]

Karen,
...
A quibble (which I do a lot of )
...

Come on Selden;-)

Celestia and its community would really be different without your dedication!

I definitely share the "special acknowledgement" Chris put into Celestia's README for you...

Bye bye Fridger

[Rei]

To Paolo:

Oh, indeed, the engine that I wrote would need modification. It wasn't designed for a landscape that reaches to the horizon, and was only designed for specific terrain sets. But, one of the great things about fractal terrain generation is its adaptability to all sorts of environments. I could whip up terrains with it that look like those links (actually, even better) in just a few hours of tweaking. Getting the terrains to extend to landscape-view instead of a block view would take about a day. The biggest amount of work involved would be getting it to load and unload parts at proper times, and to get a smooth transition from the bumpmapped sphere (distance) to a polygon-mesh terrain (close).

Fractal terrain generation is really neat in how it works. For the basic unmodified algorithm (I modified mine), you take "known points" (or make up some points) that form a grid. Then, you double the resolution of the grid. You make the new points be at the "halfway height" of its neighboring known points. But, not exactly - you add a random element to it. Then, you double the resolution again, randomize again, etc, until you have a vivid mountainscape. There are all sorts of things you can do to modify the process - in my "rugged jungle" images, I determined whether there were trees by how extreme the slope was, and changed the color and light properties, and added some random height noise to simulate a canopy. I made the landscape appear very rugged by using my random-tweaking-value tend to be more extreme when the altitude is higher. For moonscapes, you place craters of random sizes along the scene, with a transition algorithm, in addition to scattering surface rocks. Etc. By tweaking generation parameters, you can make any sort of world that you want.

[Paolo]

To Rei and t00fri

I think that the discussion about verbosity of displayed objects info is degenerating a bit. I feel guilty about this so I want to say something to bring a peaceful environment.

Karen I think that the point of view of Fridger about statistic is correct. The defense of your point of view cannot rely on this small and so not representative google set of data.

My point of view regarding more infos about objects was not related to verbosity. One of the others feature requests that I will submit soon will be about the HUD reorganization, so space on the screen is precious.

My request was intended to be more simple. I would like to see more object informations in separate dialog box. But dialog boxes are ugly, covers the background. Something of semi-transparent should be better (BTW that was one of the reasons of by my UI implementation proposal).

Using a web link in the context menu is the actual solution. Fridger I knew this feature very well. Perhaps I'm not representative but I do not use Celestia when I navigate on Internet. First I have a phone modem connection end is not very cheap so time is precious. Second Celestia manages very bad the multithreading under windows platform, so it eats almost practically every available CPU resource. Under these conditions performing other activities becomes irritating slow (I have Win XP an Athlon XP 1800, 256 MB RAM and an ATI Radeon 8500 it is not the best available but ...).

Fridger I agree that on a third party website on the internet (if available and reachable) you can find more reliable info. But my request was intented as a compromise. A little bit more info on objects excerpted from these reliable sources contained in plain text files and so always available in Celestia as said by Karen.

This argument is very huge and I have a lot of ideas regarding more info about objects. Sooner or later I will bring them together in my website.
Meanwhile please find a compromise to bring a positive discussion.

Bye - Paolo

[Rei]

Karen,

There is a version of Celestia's User's Guide that's in HTML format, but that's completely separate and not directly readable from Celestia as it is right now. It's easy for someone to define an object which links to it, but, again, that's not quite the same. Also, the User's Guide is a manual. It's not designed to provide help (e.g. it isn't even indexed).

That's what I was referring to. You go to "help"->"celestia handbook", and it loads in a browser window. My for not checking to see whether it was, in fact, a web browser and not a standard help browser that it was loaded in.

[Rei]

I work in graphics (yes, Fridger - do you want screenshots of programs that I've developed, since you decided to cast aspersions about that in your last email simply because the sysadmin didn't restore my graphics driver the last time that they restored this system, as if that's my fault?), and like to program, so I thought something like this would be good to work on...
- Karen

But since you are pulling all this out of the email privacy (not my fault), this argument does not apply to your home machine, where you are your own sysadmin;-) are'nt you?

Actually, apart from the fact that I don't usually code graphical work at home nowadays, my home machine was accelerated, and my argument - that the ability to change polygon resolution would be a nice thing - thus was completely reasonable. And concerning email, seing as you were pulling things out of Chris's private emails to you to send to me, and that you had moved this conversation to the message board yourself...

So, is there interest in holding a poll about who uses information-off, information-terse, and information-verbose, who would like to see them removed, who prefers using a web browser rather have the information come up on their screen, etc? Certainly I would think that google's random selection of pictures of Celestia (for which I would gather, say, the results for 100 images if you'd prefer a more representative sample) available is at least as good as the limited subset of people who come to forums if not better, but I would welcome such a poll to know what those who haven't been taking part think.

[Paolo]

To Paolo:

Fractal terrain generation is really neat in how it works. For the basic unmodified algorithm (I modified mine), you take "known points" (or make up some points) that form a grid. Then, you double the resolution of the grid. You make the new points be at the "halfway height" of its neighboring known points. But, not exactly - you add a random element to it. Then, you double the resolution again, randomize again, etc, until you have a vivid mountainscape.

Ok, we know all of this, is almost trivial. But you never mentioned LOD management that is the real critical point. Obviously you cannot multiply the dataset splitting and randomizing the grid without taking care of the memory usage. I am interested in how you manage this point.

The fly simulators that show very large landscapes to bring up sustained framerates have highly optimized rendering engines wit a lot of tricks regarding LOD management and advanced z-buffering tecniques.
Moreover if you implement this in Celestia you have to remeber that Celestia performs a lot of other tasks each cicle.

Bye - Paolo

[selden]

That's what I was referring to. You go to "help"->"celestia handbook", and it loads in a browser window. My for not checking to see whether it was, in fact, a web browser and not a standard help browser that it was loaded in.

Which version of Celestia are you using on what platform?
I'm guessing you're using a recent prerelease on Linux.
That feature is not in the Windows version. Since Christophe (who uses Mandrake) was the one responsible for translating the manual from Word DOC format into HTML, I suspect it's something he sneaked in

[t00fri]

I work in graphics (yes, Fridger - do you want screenshots of programs that I've developed, since you decided to cast aspersions about that in your last email simply because the sysadmin didn't restore my graphics driver the last time that they restored this system, as if that's my fault?), and like to program, so I thought something like this would be good to work on...
- Karen

But since you are pulling all this out of the email privacy (not my fault), this argument does not apply to your home machine, where you are your own sysadmin;-) are'nt you?

Actually, apart from the fact that I don't usually code graphical work at home nowadays, my home machine was accelerated, and my argument - that the ability to change polygon resolution would be a nice thing - thus was completely reasonable.

No, not at all! Before I'd be writing long posts about bad graphical performance of Celestia in a forum with 900 registered users (!) as a 3d graphics expert, I would sure know how to check first what was going on before exposing myself!

We are all making mistakes. Continuously. I am pulling this out for another reason. It illustrates very well that with a little bit of thinking and applying common sense before acting, most of these "blown up" discussions would have been superfluous...

Bye Fridger

[Rei]

The key to extending it to a landscape mode is tesselating the grid more when it's closer to you; thus, the big adaptation would be to make it use not a single grid, but a series of grids of different resolutions, the more distant ones lower res and the closer ones higher res.

I'm not sure what sort of advanced z-buffering techniques you could do in OpenGL, since OpenGL generally does z-buffering internally. However, there are things that the programmer can do; I already had it do a degree of backface culling, which could be augmented by culling of faces that are obscurred. Given the performance of it on that ancient laptop, however, I don't think there will be too much of a problem. If there is, I could switch from reliance on polygon resolution for all details to texture and bump map resolution to some extent. Also, while there are many other things that Celestia does, the planet itself can at least be disabled when you switch to terrain.