Linux/Mac development Suite?

[cmaxwell]

If I were to start specing out the "road to success" for an OSX version of World Wind, what development software is the generally accepted standard? Visual Studio is what windows developers use, what about mac?

Same goes for Linux?

Thanks,

[Felix Oxley]

Originally posted by cmaxwell@Sep 24 2004, 03:58 PM
If I were to start specing out the "road to success" for an OSX version of World Wind, what development software is the generally accepted standard? Visual Studio is what windows developers use, what about mac?

Same goes for Linux?

Thanks,

Quoted post

I am not a Mac developer, however I believe that the standard used to be Code Warrior and is now shifting to Apple's own XCode development environment (free with OS X) and GCC.

I came here to post a question 'I am shocked that this only works on windows. Are there any plans for OS X / Unix versions?'. So I am glad to see some interest.

Is this idea just a twinkle in your eye or is it more definite?

[cmaxwell]

Originally posted by Felix Oxley@Sep 24 2004, 04:20 PM
I am not a Mac developer, however I believe that the standard used to be Code Warrior and is now shifting to Apple's own XCode development environment (free with OS X) and GCC.

I came here to post a question 'I am shocked that this only works on windows. Are there any plans for OS X / Unix versions?'. So I am glad to see some interest.

Is this idea just a twinkle in your eye or is it more definite?

Quoted post

it's much more than just a twinkle. Based upon responses today, the priority for porting has been elevated considerably. I just might need help on this though.

[Davey]

Originally posted by cmaxwell@Sep 24 2004, 04:42 PM
it's much more than just a twinkle. Based upon responses today, the priority for porting has been elevated considerably. I just might need help on this though.

Quoted post

I look forward to the OS X version!

Also, have you tried contacting Apple about help with porting it over? I'm sure they would be interested...

[cmaxwell]

Originally posted by Davey@Sep 24 2004, 05:44 PM
I look forward to the OS X version!

Also, have you tried contacting Apple about help with porting it over? I'm sure they would be interested...

Quoted post

I will bring this up with "the powers that be" B)

[Richard Johnson]

Originally posted by cmaxwell@Sep 24 2004, 04:42 PM
it's much more than just a twinkle. Based upon responses today, the priority for porting has been elevated considerably. I just might need help on this though.

Quoted post

Let me add yet another vote for a Mac version. Mac hardware is pretty well known for graphics capability and I bet this would really scream on a G5! (Unfortunately, I can't give you any pointers as to graphics packages for MacOS X, but I bet Apple would be willing to help. It'd make a great demo program for their hardware!

[adamhill]

Mono on OSX is like C in the olden days - editor and the command line. There is a editor called SharpDevelop but it is not anywhere close to VS.NET.

You will also face the same problems in the Mac port that you had in the Windows one -the OpenGL wrappers. And you have the added issues of what parts of Win Forms have been ported by the Mono team or rewriting the GUI in *cough* GTK#.

Most of the core seems to compile with Mono IRandom test of a few .cs files)

[mic]

I'd go for Eclipse, convert all code to Java and use OpenGL with JOGL. You then could use Java Web Start to distribute a single package for all supported configurations and you could use the same IDE (Eclipse) on all three platforms - Windows, Mac and GNU/Linux.

IMHO this is the easiest way to develop cross platform software. Of course all the C Sharp and DirectX things would need to be converted to Java and OpenGL, so it's quite a bit of work at first.

[Guest]

Porting from a total .NET environment and DirectX9 to Linux or OSX will be fairly time-consuming. Why not just put the code out as Open Source? Ports to other platforms and to other languages (Java for example) will happen fairly quickly. What you should standardize/publish is the API / Network protocols for retrieving the images and for any other server access required.

[cmaxwell]

Originally posted by mic@Sep 25 2004, 03:31 AM
I'd go for Eclipse, convert all code to Java and use OpenGL with JOGL. You then could use Java Web Start to distribute a single package for all supported configurations and you could use the same IDE (Eclipse) on all three platforms - Windows, Mac and GNU/Linux.

IMHO this is the easiest way to develop cross platform software. Of course all the C Sharp and DirectX things would need to be converted to Java and OpenGL, so it's quite a bit of work at first.

Quoted post

How is the performance of JOGL? Compared to the more traditional C++/OGL approach?

[Henk Poley]

Maybe just a dumb suggestion but why not install Mono your Windows machine (in parallel to .NET) and try to first get it to to work on top of that.

Shift to using gtk# and OpenGL next to the already existing DirectX code. You should then be able to run it on Windows, Linux and Apple OS X.

Specifically for mono there is monodevelop, but it is not much more than an editor that knows assemblies, AFAIK.

[Henk Poley]

BTW I should add that I'm not a Mono nor .NET programmer. I would recommend you to look around what is out there on mono. It is not that much but most portable graphics libraries are there in more or less .NET-ified state.

From what I read Wxwindows and SDL should be rather easy to work with together with OpenGL, but GTK# is the most mature graphics library on mono. I don't know of there are any existing managed DirectX-alike libraries for mono.

[rgovostes]

I'm looking forward to an OS X version too...

The development environment of choice is Xcode, as mentioned before. It's what most of Apple's programs are written in and supports Java, C, C++, and Objective-C with Ant and GCC. Objective-C and Apple's Cocoa API are the "standard" for developing programs for OS X.

Java doesn't perform very well on Mac OS X. I'd recommend against this.

Instead, changing from DirectX to OpenGL would make the code more portable. Additionally, writing a generic wrapper class for sockets on Windows, Linux, and OS X, would allow image downloading to be fairly cross-platform.

Alternatively, you could use curl, which comes with every copy of Mac OS X (I think, maybe it's only if the BSD Subsystem is installed) to download images, if they're available by HTTP.

[LeighGus]

The usual high end development environment is Code Warrior. It is possible to use a single code base to compile for most operating systems.

[\]

My toughts would be to go for a traditional C++/OGL approach
if your aiming at support from the comunity that is.
The basic reasons being that in that case you'll hit the widest
possible contributer base.

Locking yourself in on 1 platform might cover the biggest
userbase but thats rarely where the most contributers/enthusiasts
are. It's not an OS religous thing but many people do contribute
not for the world but for themselves. They want a specfic thing
to run on their machines. Portability really matters at that point.

It also can extend the live of the project from a short period
to a lifetime just because of people being able to take over
where the original designers leave off. The C# and DirectX
choice are the right tools for most games since most of these
are just pretty graphics. The time to market is very short
but its general lifetime is as short too. This seems like the kind
of app which should be an evolutionary thing rather : new datasets,
improved graphics, extensions ...

About Java I would like to say that the GL is quiet up to speed. The
main problem stays there that is often a pita to have it properly working
all the time on all platforms. It often breaks down due to all the
different versions and user installations available. With the target
compiled code its meuch easier to avoid dependencies and if things
get broken because the layers above change the developer can solve
them iso the user. You might ask the people who worked on MAESTRO at NASA
about their experience and the current popularity of their project.
It might confirm or contradict my view.

So the choice is really : where do you want to go with this project.
How much external involvement would you like ?
If you intend to keep the project under thight control for quiet long time
and do 95% of the workload, then don't bother to do all this stuff.

In the meanwhile you could also just post the code and wait to see if there's enough
enthousiasts which care enough to start a port. If the port is really good,
than embrace and extend it if you still like to work on it.

I had some positive experience with the latter. Some time ago I did
such a port of a 3D Windows game (C++/DirectX/Ogl mix) towards Linux.
The original designer liked it very much. Decided it was the way to go
even it did not have everything the orignal version had at first.
Once that workload was done, BSD and MacOS people joined in and finneshed
the rest of the portabiltiy problem.
Now the code is SDL/OGL/Wx-windows and is fully cross platform.
Since then I haven't been active involved anymore but the number of contributers
have grown and the game is still evolving. (You can check out www.scorched3d.co.uk.)

Asside from all of this congratualtions on your current release and achievments. Whatever choice
you make : good luck.