This is the next step: Delay DX10 until (end of) 2004

[CoolAsAMoose]

I'm really excited about these new upcoming 3D chips, together with the new possibillities given by DX9/OpenGL2.0. Too me this really feels like the biggest step in real-time 3D (at a reasonable cost) since the launch of Voodoo Graphics in 1995 - up until now it only seems as we have had speed increases. Allthough we have been "shown" next-gen architectures from 3DLabs and Matrox with some DX9 features, NVidia is likely to be the first with a "full" DX9-chip - followed by ATI. During 1H2003 we will hopefully see refreshes of Parhelia-512 and P10, adding the missing DX9-features, together with some optimizations (bandwidth saving techniques for Parhelia) and higher clock speeds thanks to the use of a 0.13 um process. This brings me to the point: isn't DX evolving too fast?

Allthough it's cool, especially for us 3D-geeks, with all these new features, the game developers need some time to catch up. If MS delayed DX10 until 2H 2004 (instead of 2H 2003 - if that is the plan, that is???), the game developers would have time to catch up and really start to use the new features of DX9 (as well as exploring the programmability). OpenGL2.0 would also be able to catch up, allthough this isn't an objective for MS.

In one sense it would be good for the hardware developers as well to be able to catch up. Instead of being forced to integrate a lot of new features once a year, they could refine their architectures, bring all DX9 features to the low-end, create PCI Express (formerly known as 3GIO) versions, etc.

Thoughts anyone? Maybe this is already MS' plan.....

Anyway, as I mentioned, I'm really excited about these upcoming 3D-chips. Now bring me some high-speed games (racing, flying, space combat) taking full advantage of them...........

 

[ushac]

I don't think delaying drivers will do any good. The game developers will design their engines so a lot of hardware out there is supported. If the drivers can do more they can just ignore it (or make the feature optional). Other applications less dependent on a broad costumer base can make use of that feature. On the other hand - if the drivers can't support some new hardware feature, no software with it will be made and then there's no reason for hardware developers to implement it. Or am I just rambling?

Regards / ushac