OSNews

No, it will work in Windows 7 too.

What difference would it make? The gaming community for Linux and OSX is virtually nothing.. I mean it would be useless to waste man hours doing something like this.

But the reason the *nix and OSX communities are so small is because you can't play (most) games on those OS's.

That, my friend, is a catch 22

Edited 2008-09-04 17:04 UTC

The Linux based OS gaming community is alive and well with a large library of games to select from. These gamers also enjoy the win32 only games though so, being of higher technical skill level, usually have dual boot winXP partitions for DX crippled games.

I sort of think that well developed code could easily be recompiled against the libraries for various platforms with little issue and time for porting the differences over. Being interested in seeing the DX libraries ported to different platforms or even different versions within the windows family is a very valid interest; even if you mistake market share for any kind of remotely accurate measurment of platform usage.

Mind you, I'd be surpised and happy if it even worked with winXP along side whatever flavor of Windows the marketing department will be pushing by then.

Excuse me, Direct3D is a proprietary windows standard which is only supposed to work on windows. So the question of it running on linux doesn't arise.

Having said that, when Gallium3D will be operational, one can look forward to having full Direct3D 9 acceleration in linux (on cards that support D3D 9). This would be particularly useful for software running on wine. But I am not holding my breath for that one. If only game developers hang on to D3D 9 for a few more years....

Direct3D 9? What's this, the year 2002?

Game developers are continuously pushing the envelope of graphic cards and you're harping on about 6 year old technology?

I don't see why Gallium couldn't emulate DX10 as well, if the driver model below permits it. The framework is designed to support new API's like OpenGL 3.0, and DX10 should be quite similar in feature set.

Features like memory management that should (finally) go upstream in the following kernel (for Intel at least), and a bit later for Radeon will make it easier (i.e. provide WDDM 1.0 equivalent features).

The remaining problem will be optimisation of OSS drivers; I doubt it will be comparable to closed source equivalents that soon (AMD and Nvidia invest a lot of money to get each bit of GPU time used as efficiently as possible).