You could write that wrapper in pure Java and have a nice simple binding to DirectX though…
Cas 
Well, think about it: There are already lots of PCs without decent OpenGL drivers. The OpenGL drivers on most new PC’s with onboard Intel gfx are rubbish already, and those are a very large chunk. The number one priority for most graphics boards vendors on Windows at this moment is stable DX drivers. I have the feeling that the number of OpenGL games is dropping (I don’t know the numbers, but how many AAA OpenGL titles have there been compared to DX titles the last year?). iD publicly stated that they’re starting to embrace DX so who knows their next engine will be DX too?..
Sure, OpenGL will have its place in the foreseeable future and top of the line gfx board manufacturers will probably continue to deliver decent OpenGL drivers for a while, but OpenGL’s position on Windows PCs is already declining. With Vista having some sort of D3D wrapper for OpenGL, I would suspect graphics board manufacturers like Intel don’t have a lot of incentive developing real OpenGL drivers anymore (chances are Microsofts OpenGL wrapper works better than their current OpenGL drivers anyway).
Sounds a bit like FUD to me Erik… can’t say that I’ve noticed any difference whatsoever and the Mac market is actually growing in size rather than shrinking.
Cas
I’m not trying to spread FUD here, I’m sorry if it’s perceived that way.
I just have the feeling that Vista’s OpenGL wrapper might be a reason for some gfx card manufacturers to stop delivering real OpenGL drivers. But yes, I’m just guessing here.
Anyway, the fact still stands that DX drivers are generally more stable on Windows, DX is available on all MS OS-es, and most windows developers seem to have a preference of the D3D API above OpenGL nowadays.
I think that should be enough reasons for a D3D binding.
Xbox … sure
360 … maybe not.
AIUI, the 360 has a more rigorous approach to memory protection, such that it is currently impossible to run proprietary VM’s on the machine, because it won’t let you execute stuff pre-ordained to be data, and without MS’s co-operation, all your VM-run program code will be marked as data.
Haven’t tried it myself, and (fortunately) I’m not under NDA with them so YMMV, this is just the rumour I heard.
Excellent!
So … know any good graduates who want jobs in games development?
Well there’s me! Although my plan is to go back to uni and do a Masters this September, and I’ll be hitting all the major teams throughout the year and also creating my job-worthy demo’s!!!
would anyone be writing this directx binding, if they we’re stimulated with some small amount of cash?
Like a bounty!
Probably depends on how small an amount you were talking about?
Kev
I think one intresting question is would anywone here be intrested in buying such a product ?
It’s not really a “buying” kind of thing, it is, after all, just a veneer over someone else’s tech.
I’d donate $100 to it. In fact I think it’s high time LWJGL supported a D3D binding. Maybe we should do a donation drive to get someone to do it. I mean, I don’t even want to use it, I just see a pressing need for it to become available, to enable a whole load of dependent projects.
I think ChrisM should donate some not-very-small-sum of money to the project. Like, $10k, with Sun’s blessing…
Cas
A lwjgl supported dx project should be possible. I am sure elias will hate it (since it’s windows only) - but we do have a chunk available for donation…
I started to do the DirectX binding. I didn’t want to say anything until I had something to show for it. I am having a small problem that is hard to debug. There is a post on LWJGL forums about it. http://lwjgl.org/forum/index.php/topic,2363.0.html. Once that problem is sorted out I should have something visual to show. I developing it the same way as LWJGL. With static accessors to the DirectX devices. I have tentatively title it Windows Game Library for Java (WGLJ). So, some questions.
- Do you want to host it under java-game-lib on sourceforge or start a new project?
- If anyone doesn’t like the title, I am open for suggestions.
- Once its hosted, does anyone want to help out?
- Do you want it as a totaly separted library, or a section of LWJGL? (If part of LWJGL then the title doesn’t matter)
What would be ideal is that it’s just part of LWJGL and integrated into the existing “API agnostic” bits, eg. Display and input etc, and just distributed in a lwjgl_d3d.jar, and we’d move the OGL stuff into lwjgl_ogl.jar.
Cas
[quote]I think ChrisM should donate some not-very-small-sum of money to the project. Like, $10k, with Sun’s blessing…
[/quote]
I think that’s a great idea - I believe with just little money as support for some open source developers could accomplish a lot! Port to XBox and PS anyone?
I personally would have just volunteered to take the bounty and pretended I hadn’t started working on it. Now we all know you’re almost finished I guess there’s no need to “donat-ivate” you!
Damn, hadn’t thought of that. 
Seriously, I am just happy to do it for the experience. Plus if we get Java a higher profile, all the better.
Ok. I’ll get LWJGL out of SVN and start working on it. I’ll get the LWJGL group’s opinion before I do my first commit. Send me an email or PM me if there’s anything you think I should do. (captainjester at hotmail dot com)
Wow, thanks CaptainJester for your self-less efforts. Your ‘actions speak louder than words’ approach is impressive.
In regards to the Direct3D C++ JNI calls, maybe the Java3D source will have your answer since I hear that it is now fully open source and it uses D3D.