Slashdot Mirror
The People Behind DirectX 10
ThinSkin writes "In the first of a three-part series covering the people behind the new DirectX 10, ExtremeTech interviews Microsoft's David Blythe and Chris Donahue to discuss the development, decisions, and future of the new API. They answer several questions such as how different it will be than DX9, why it will only be for Vista (and not for XP), and when we might be able to see it."
Call me anti-Microsoft, fanboy, zealot if you will but, I find it seriously hard to believe that they were having trouble getting DirectX 10 to work at a decent speed on Xp, an operationg system that hogs probably half as many resources as Vista will. Theres no reason given in the article why it should run faster on Vista it just says there is one.
What I find really amusing is how many MS fans will say 'Oh, obviously; DX10 depends on LDDM, and XP doesn't have Longhorn's nifty new driver model!'.... Yet 98SE supported both WDM and it's predecessor, at the same time.
What's keeping MS from backporting some of the new Longhorn kernel/driver niftiness to XP? Oh, right. Money. There's no money in adding new things to an already-sold product. It's all about selling the new hotness.
The preceding comment is my own, and in no way construes an opinon of the Emperor of Mankind.
I must be the only person who misread it as "DirectX 1.0", like we were going to get some sort of historical retrospective from The Man Himself, Alex St John, on the travesty that was 1.0 :)
OK, I'll bite on this one.
First, compatability. Let's say you're a desktop user with a 2.4 kernel. Why would you not switch to a 2.6 kernel. Hmmm, well maybe you have some legacy hardware that only works on 2.4. But hold on a second, if your machine is using such legacy hardware, then what are the chances that it's going to have/need a new skookum video card. Hmmm.... not very
So in most cases, your newer kernel is going to work with the hardware. Cue in Vista... well it might work with your hardware, but chances are pretty good that not all older software, let alone hardware, are going to work with the new OS... hmmm, not good, but then we run into...
Secondly, performance. Well, chances are that (again, if you can fit a newer video card) your processor won't need any upgrades to run 2.6. In fact, in many cases you'll see better performance using 2.6 with the same resources, provided you don't bloat your kernel.
Now let's check with Vista: Oooops, nope it looks like the hardware requirements are a little steeper for that one, aren't they? Moving on
Third, modularity+environment. While some parts of the kernel definately affect userland, the two are different. The desktop environment, such as KDE, doesn't really care if you're running 2.6 or 2.4.
Modularity of windows.... buzzzz. Nope, sorry. Chances are that with that shiney new system core you will also be seeing lots of UI and interface changes. You can't get a "kernel upgrade" in the windows world in quite the same way.
And lastly, the big deciding factor for many: cost. Windows XP Pro has it's "end of life" listed in (I believe) 2009. However, directX 10 basically forces a few issues here:
- Hardware vendors don't get newer capabilities from DX9. Therefore, they develop for DX10. Consequently, your games are rather suckish or possibly playable in DX9. Microsoft might still be patching XP, but they'll happily blame the vendors for not releasing compatible products when Vista rolls around. So for all those that want your stuff to still work... it's time for an upgrade. In the case of businesses (not everyone needs directX10, but I'll be betting that some graphics software might), that can be a hefty chunk of change.
Somehow, it seems one has a lot less alterior motives when saying "we won't suppose this, but if you update for free we will" versus "we won't support this, but if you hand us another wad of cash we'll try"
Why argue with an external body dominated by guys from a different side of the industry (cad/cam) just to add some capabilities to your platform?
Many on Slashdot would say its to tighten M$'es dominance - true, but they are probably fed up with standards body bloat too - why let an idea slip due to bureaucracy at some working group when you can implement it now yourself and do the standards thing later when your customers FORCE you to.
Interestingly DirectX adoption seemed to accelerate after NVIDIA tightened its market share in the early part of this decade as 'custom' OpenGL render paths for 3dfx and others disappeared (as the hardware vendors did) in games and were replaced with DirectX 6/7 paths.
If your user-visible API dictates the structure of your drivers to the extent that you can't back-port it to another driver model, then you're doing something deeply wrong. Or you're using technical buzzwords to confuse people into thinking that a management decision is a technical one.
I am TheRaven on Soylent News
Too late:3 7/137972.html
http://www.ticalc.org/archives/news/articles/13/1
The printing on game boxes that read "Requires Windows 95" "Requires Windows 98" and "Requires Windows XP" will soon have a brother. Big shock guys, there is going to be a "Requires Windows Vista"
Windows has more viruses because linux has more virus coders.
That's the wrong way to look at it. The DX guys are, in essence, acting as the negotiators between the game developers and the hardware manufacturers. The game dev folks tell the DX people what they'd like, in terms of feature set. The DX people then work with the hardware manufacturers to implement those features. It really does make a lot of sense. The game developers get what they need, in terms of feature set and APIs. The DX folks then get to work with the hardware manufacturers to ensure that the required capabilities are available and relatively consistent across hardware vendors.