Slashdot Mirror
What Game Developers Think about DirectX 10
mikemuch writes "In the last of his series of interviews with the stakeholders in Microsofts upcoming DirectX graphics API, Jason Cross speaks with the developers of Hellgate, Crysis, Flight Simulator X, and Age of Conan. They seem pretty stoked about the new technology's ability to get visual detail to a much higher level of realism, and to offload physics and AI to the CPU." From the article: "Without hardware, it is hard to evaluate which features will really make the biggest performance impact. The geometry shader looks pretty full of potential. So does the fact that you can write to buffers from any shader and then read them into another shader. Texture arrays look like they will make a big dent our batch count, which should lead to much better frame rates. At this point I feel like I'm looking at a shiny new toy through a shop window: I can't wait to get my hands on it and play with it, but I don't really know what it can do."
I think that says it all. All we have now are lofty marketing claims and unfounded speculation. I am as excited as anyone to see what it can do (I admit to being a fan of flight sims) but this does not really help me understand any better. Since this is
Information wants a fueled airplane waiting at the hangar and no one gets hurt.
Woo hoo! I hope 'offloading' MS Office to the CPU is next!
Seriously, though, is this supposed to be a feature?
FUNK!
The articles responses sound more like what marketting thinks about it, rather than developers.
I'll tell you what developers think about it:
DirectX 10 is supposed to be Vista exclusive. Smart people know Vista is a pile. Of course Microsoft will shove Vista down people's throats eventually, but you're still being locked into a piece of software that will tank for the first 12-18 months.
Realistically there is very little the new API will offer in this generation of games, in some cases it detracts from it (Hellgate, which looked much better without the normal mapping), because Microsoft is buying out these folks in mid-development so they can say they've got support for this API and make marketting-articles like these.
Lastly, DirectX 10 is going on a platform that will rule out OpenGL. Not many people use OpenGL in the last few years, so, sadly, this is a minor point. But it's great to have choices, and to have your choices superficially limited always gets me in an uproar.
In a nutshell DirectX 10 is not bringing anything terribly new to the table, while removing very critical freedoms from developers in the first place. They will fill pages with positive bullet points on DX10, but when you plop down a DX10 title next to a DX9/whatever-else title in the end, they will not be noticably different.
Sorry for the ranty-rant.
Theres much bigger support for OpenGL 2.0 on pretty much all hardware and platforms. I never saw why people use DirectX when all it does is limit your application to MS platforms and is dictated by MS.
Since these developers clearly don't really know that much, the question is really asking how much they believe whatever marketing hype or hearsay they've encountered. I'd be more interested in finding out what they think about Microsoft potentially shafting OpenGL, or if they even care at all about vendor lock-in.
For the past few years I havent had nearly the urge to upgrade that I did in the past. Most games have become so gpu dependant that upgrading the video card has been sufficient enough to keep the latest games running on even a 2-3 year old machine. From what I read it appears that dx10 is more of an attempt to make the rest of the hardware insufficient than any real improvement in the gaming experience.
With all the talk of physics engines and vista exclusivity it has me concerned that what is already a small market is just going to get smaller. My last video card upgrade was $400 and it was an upper mid-range card (x850 xt platinum) not top of the line. My friends and family thought I was insane since many of their pc's cost less than that. I cant even count how many clients I have that get pissed off that their new dual core system cant run a two year old game because the video included was an onboard intel gpu.
IMHO thats whats wrong with the PC gaming world and what pushes users to "casual" gaming. Most folks I know dont even bother to look at the new games on the shelves because they assume they cant run it and refuse to put up the cash to upgrade what they see as a minor part of their pc. I understand Microsofts and their vendors desire to get people upgrading again but unless computer manufacturers either start uncluding decent video cards or intel increases their integrated performace by leaps and bounds the market is just going to dwindle even more.
Considering that Directx 10 is only on VISTA, any game developer launching games for Vista from the get go will probably be very upset when their sales numbers do not make up for development costs.
I think by now, everyone realizes that Microsofts product isn't good until the second or third version and people will wait. As such, games released with Directx will be directly affected by this.
This is my sig. There are many like it but this one is mine.
You forgot some of their most popular games:
Pac-man, Pole Position, Galaga, and Dig Dug.
And I'm pretty sure those mainly ran on their own hardware (with coin slots).
I'm not with some big studio, mind you. If anything, I'm more along the line of the currently so popular (gee, I didn't notice it in sales, to be honest) "little lunch break games".
I've taken a look at DX10 (well, at the parts you get to see if you're not one of the huge studios, at least). Yes, a few things look promising. But considering the downside, I'll think twice before switching over to DXX.
First of all, it's a new beast. The transitions between DX7 and DX8, and to a lesser degree to DX9 have been considerable. Yes, the changes were minimal, but if you wanted to go with the new flow, you had to rework your code. For rather little gain, if you ask me, since "small developers" certainly don't use the wonderful new features offered. Yes, a programmable shader is nice, HLSL was a big, big help for creating a better shader and it was a big speed boost in code creation. But, frankly, it's not something I'd wet my pants over. Whether the easier creation of some parts and more flexibility in others was worth the time input to revamp your code to the new calling conventions is debatable.
Second, and more important, Vista-only. Now, as I've said, my "target group", so to speak, are people who want to play a li'l game now and then, not hardcore gamers that spend their vacation money rather on new hardware than on a trip to the Bahamas. In other words, my clients will have second line computers, not bleeding-edge hardware. And likewise they will most likely not jump onto Vista the moment it is released, so why should I? I would lose a considerable part of my market.
It might be useful for studios that really work on state-of-the-art games, that HAVE to go to the top and grab the most advanced features available in drivers. It's likely that soon you'll only get top performance out of hardware with Vista, because driver development for older systems will be slacking, just like it was for the Win95/98 line at the wake of 2k/XP. Because test pages will always go for the most current system, so drivers for those will invariably get perference when it comes to tweaking and performance tuning.
Likewise, MS will soon start to abandon fixing anything but the most glaring bugs in older DX versions and newer features of hardware will only be supported in DX10, just like it's been with older versions of DX.
So yes, game development will shift to DX10 and Vista at some point. The question is when their target audience shifts. If consumers don't buy into Vista, studios will have to continue making games for DX9. Or MS will have to port DX10 to XP, appearantly it is somehow possible (though it's quite possible that they only ported the SDK and the runtime won't be working with XP).
Whether or not games will require DX10 is finally up to a few questions:
1. How many people will go to Vista (and thus DX10), so how big is the market?
2. What "killer" features will DX10 offer that DX9 won't?
3. Is it easier or harder to use DX10 compared to 9?
4. How much more horsepower will DX10 need to run games comparably fast with DX9?
We used to have a Bill of Rights. Now, with the rights gone, all we have left is the bill.
In Vista, GPU resources are virtualised by the operating system; it's similar to the transition between real and protected mode operating systems on x86. It's kind of like asking for DirectX 9 on Windows 3.1. You could argue that it's just a different driver model, like 98->XP, but I would imagine that adding virtualisation would make the transition much more difficult than that.
I'm reading claims that this is some conspiracy between Microsoft and hardware companies to force people to upgrade.
I think what's really going on is that Microsoft is addressing the demands of the gaming industry with DirectX 10. The game industry is obsessed with realism. There's this overwhelming desire to make games look and feel real, with gameplay taking the backseat.
I also can't help but think that developers are getting every more sloppy because they're depending on consumers having sufficiently powerful machines. It almost feels like it's done intentionally. Release a game that struggles to run well even on some of the fastest machines out there and suddenly the game is used for performance benchmarks. It's free marketing. The game is mentioned in computer magazines everywhere. Now the consumer has a game that they absolutely need to get in order to test the limits of their machine.
DirectX 10 seems to offer a few neat features, specifically those dealing with physics. Unfortunately, DirectX 10 is simply perpetuating the poor habits of the gaming industry. It's ensuring we're going to be seeing contrived FPS games for years to come.
However, it's not going to stop anyone who wants to be more creative. Microsoft is simply responding to market demand.
The /. title is one "do" and a punctuation mark short of being a question.
Thats the whole point of DirectX and an API! So the game developer doesn't have to worry about the underlying hardware. All that's required is porting DirectX itself to the Xbox OS and hardware, so most of the work is done by MS themselves.