Qt Celestia

[ElChristou]

Just tested a Windows version, there is a slider to filter the number of hits; this feature seems not bad, so I wonder what was the reason to not use it under KDE...

It's just that when I "copied" it the slider wasn't there. Anyway, I'm not sure it's that usefull, 100 or 500 stars, we could just as well display all 500 by default.

Ok, then there is a limit in the number of hits...
Is such limit up to date? I mean we have now some good DSOs catalogue, soon perhaps pulsars and other exotic objects, soon or later some revised and extended asteroids catalogues, so...

...if the browser become "stronger" (no limit in the number of hits to not lose some results), the criterions of the query would limit naturally the hits, but for a simple search (1 criterion) this slider could be useful, no?

[chris]

Here's the result of some further experimentation with Qt4:



This is not intended to show a final look for the star browser; it's primarily an experiment. Some interesting things:

- The star browser is a dockable widget: you can attach and detach it from the main window, and place it either to the right or the left of the 3D view. You can also hide it completely. The state is saved between runs of Celestia, so you can get the tools arranged the way you like and not bother with them again.
- You can select multiple stars and ranges, then mark all selected objects with a specified symbol. Optionally, you can also label the markers (thanks to a recent change by Vincent.) The screen shot shows stars with planets all marked.
- You can right click on a star in the browser and get the same popup menu you do when clicking in the 3D window. Goto, Center, Follow, etc. are done with the context menu rather than buttons. The KDE browser has the same behavior. It's nice because it's clearer what object the action affects, especially when multiple items are selected.
- Unlike the current Windows version of Celestia, special characters (Greek letters, superscripts, etc.) are shown in the star names.

--Chris

[chris]

Another another screen shot, this time with custom color selection for markers:



--Chris

[MKruer]

Christophe, is it possible to combined binary stars listings into a single group instead of three separate items.
somethign like

Code: Select all

+Rigel Kentarus A-B

Expands into

-Rigel Kentarus A-B
  -Rigel Kentarus A
  -Rigel Kentarus A-B Barycenter
  -Rigel Kentarus B

Note: Barycenter name I think should always appears in between the to stars that are orbiting around it. Example for multi star systems it would be like

Code: Select all

Star A
Star A-BC Barycenter
Star B
Star B-C Barycenter
Star C
Star BC-D Barycenter
Star D

In a nut shell this says that Star A orbits around Star B and C, Star D orbits around Star B and C, But Stars A and D do not orbit around each other.
VS.

Code: Select all

Star A
Star A-B Barycenter
Star B
Star AB-CD Barycenter
Star C
Star C-D Barycenter
Star D

Says that Stars A and B orbit around each other, C and D orbit around each other and that Stars AB and CD orbit around a common barycenter other.

Somethings you may want to add is what constellation the star is in and if it is part of the constellation itself, also the Declination and Right Ascension I think should be listed somewhere. Additional stuff like temperature, radius can be included, but I don't think it is necessary nor currently accurate.

Either way both of these are great so far. Also i believe Chris was going to add more descriptors for all objects such as Pulsars, Neutron Stars, Black Holes, Super Massive Black Holes, Dwarf Planets, KBOs, etc...

[ElChristou]

Another another screen shot, this time with custom color selection for markers:...

Nice so far!

Somethings you may want to add is what constellation the star is in and if it is part of the constellation itself, also the Declination and Right Ascension I think should be listed somewhere. Additional stuff like temperature, radius can be included, but I don't think it is necessary nor currently accurate.

Perso the question I ask myself is: shall the browser be and aid to browse the OGL window or almost a final doc where the graphical representation become useless?
What I mean is that to me the browser should only help the user to detect an object in the OGL window; once you select the object of interest in the browser all other info is on screen (const, Ra/Dec, temp etc...).
The only data that should be visible in the columns, should be the data used to sort object... (if a sorting system is implemented)

[MKruer]

Somethings you may want to add is what constellation the star is in and if it is part of the constellation itself, also the Declination and Right Ascension I think should be listed somewhere. Additional stuff like temperature, radius can be included, but I don't think it is necessary nor currently accurate.

Perso the question I ask myself is: shall the browser be and aid to browse the OGL window or almost a final doc where the graphical representation become useless?
What I mean is that to me the browser should only help the user to detect an object in the OGL window; once you select the object of interest in the browser all other info is on screen (const, Ra/Dec, temp etc...).
The only data that should be visible in the columns, should be the data used to sort object... (if a sorting system is implemented)

Sorry I didn?€™t mean to infer to add the columns to sort by, but rather be added as an info box like in Christophe example. This would be information that would be good to have, but not necessary sort by. I agreed if you are pointed in that direction, you should know what?€™s in the general vicinity via the actual OGL. With that said the only thing I would add to be displayed in the browser, but not to sort. Would be a B or an M for binary/multi star systems, and P for Planetary companions. This way if you sort it via something other then name, you know upfront that the star has companion(s). Then again would you want to search for all stats that have planarity bodies or are part of multi star systems. The tree and the list both have advantages and disadvantages so I don?€™t think there is one correct answer. Having both and being able to switch between then might be the best option.

[Vincent]

Another another screen shot, this time with custom color selection for markers:

Chris,

That looks good ! That makes me feel like installing MinGW and entering the Qt4 play ground...

[t00fri]

I am almost there to compile Celestia-Qt4 also without MinGW, with my good old VS2003.Net and with batch (nmake).

F.

[cartrite]

I recently built qt-x11-opensource-src-4.3.3 and kind of hacked a qt4 build on linux. First I had to run make Makefile.cvs and then ./configure for kde. After that I ran qmake on celestia.pro and then make from the top directory. Probably a better way to do it but a pretty cool screen.



There is a qt3to4 program and I was wondering if the kde programs and gui's could be ported to qt4 with that program.
cartrite

[t00fri]

The building under Linux I did already some time ago. Just take a fresh CVS copy. Type in the Celestia root directory

> touch config.h

then
> cd src/celestia/qt
> mkdir build
> cd build
> qmake ../../../celestia.pro
> make
> cd ../../../..
>ln -s src/celestia/qt/build/celestia-qt4 celestia
> celestia

Aaaaah!

Bye Fridger

NB: You dont need to run ./configure!

[cartrite]

The building under Linux I did already some time ago. Just take a fresh CVS copy. Type in the Celestia root directory

> touch config.h

I tried to just run the makefile generated by qmake and it would fail after building 2 objects because it was looking for config.h. I didn't realize it only needed an empty file. I only ran ./configure to generate that file. Thanks Fridger.
cartrite

[chris]

I am almost there to compile Celestia-Qt4 also without MinGW, with my good old VS2003.Net and with batch (nmake).

F.

Impressive! MinGW seems to be a fine compiler, but it would be great to have the option to use MSVC.

--Chris

[chris]

Another another screen shot, this time with custom color selection for markers:

Chris,

That looks good ! That makes me feel like installing MinGW and entering the Qt4 play ground...

Even the this project is still in the early stages, help would be appreciated. The Qt4 UI is missing lots of features from the old UI. There's not even a set time dialog box yet, though I'm not even certain that it should be a dialog box. A toolbar for controlling time would have the advantage of being more readily available than a menu-accessed dialog box.

I do want to be careful not to create a cluttered UI for Celestia. Dockable windows have the advantage that they can be hidden or shown as the need requires. For an application like Celestia with several different usage modes, this is important, e.g. if you're only interested in exploring within the solar system, the star browser can be hidden; but, if stellar astronomy is your primary interest, it's quite nice to have the star browser panel open and docked.

I suppose it's time to open a Celestia UI suggestions thread . . .

--Chris

[ElChristou]

...I suppose it's time to open a Celestia UI suggestions thread . . .

Exactly, before thinking in porting the existing UI to Qt, it would be better to think carefully every aspect of the UI vis-? -vis the future use we can predict.

(Example, for a professional point of view (and with the increasing power of calculation) we could think in multi screen scenarios where one screen is dedicated to settings with why not a fully customizable window where one could dock the element he need...)

[cartrite]

I am almost there to compile Celestia-Qt4 also without MinGW, with my good old VS2003.Net and with batch (nmake).

F.

I tried to compile qt4 on WindowsXP and failed. It looks for at least 2 files. msvcprt.lib and msvcrt.lib Not part of my sdk or 2003 toolkit. There was a warning that official qt4 will only compile on MinGW.
http://psi-im.org/wiki/Compiling_Qt4_on_Windows
There is a patch that can be applied that is supposed to get it to compile with msvc but that didn't work either. If you can do this, God bless ya. I give up. Windows (EDIT out choice words). Works on Linux though.
cartrite

[t00fri]

cartrite,

first of all, I downloaded this archive

http://www.trolltech.com/download?targe ... -4.3.3.zip

Then I built it (which takes quite a while):

> configure
> nmake

After this is done I try batch compiling

> nmake -f Makefile

+++++++++++++++
Do you have the commercial MSVC (VC++ 2003.Net) installed??
+++++++++++++++

F.

[cartrite]

Do you have the commercial MSVC (VC++ 2003.Net) installed??

No, I got the free version of Microsoft Visual C++ Toolkit 2003. Thats probably why. I got my qt4 version from a different server but the file name is the same. I can run configure but nmake fails when it can't find msvcprt.lib. This file is not part of the Microsoft Visual C++ Toolkit 2003. It is in Microsoft-SDK Lib\AMD64 and Lib\IA64 folders but not the Lib folder. I have a copy of the files in a Microsoft Visual Studio 8\VC\lib folder that is from some program I downloaded and they get the build moving again only to fail further down the line. I knew they wouldn't work but I thought what the hell. Windows always makes me crazy.
Anyhow, I'm glad it works with Linux.
cartrite

[chris]

I am almost there to compile Celestia-Qt4 also without MinGW, with my good old VS2003.Net and with batch (nmake).

F.

I tried to compile qt4 on WindowsXP and failed. It looks for at least 2 files. msvcprt.lib and msvcrt.lib Not part of my sdk or 2003 toolkit. There was a warning that official qt4 will only compile on MinGW.
http://psi-im.org/wiki/Compiling_Qt4_on_Windows
There is a patch that can be applied that is supposed to get it to compile with msvc but that didn't work either. If you can do this, God bless ya. I give up. Windows (EDIT out choice words). Works on Linux though.
cartrite

For the moment, I recommend downloading the MinGW compiler--it works great, without any special modifications. I'm encouraged by Fridger's progress with MSVC. I believe that the missing files are the non-multithreaded C libraries; I had the same problem when building with the free MSVC Express Edition 2005 and fixed it by rebuilding all of Celestia's supporting libraries against the multithreaded C++ library, libcmt.lib. We'll get this all sorted out soon.

--Chris

[t00fri]

I'm encouraged by Fridger's progress with MSVC. I believe that the missing files are the non-multithreaded C libraries;

--Chris

That precisely what I think, too. I am working on it. I have some previous experience with this sort of thing...We'll see.

F.

[cartrite]

Yep,
I am downloading the MinGW version of qt4 as I write this.

cartrite