Beyond DirectX 10 - A glance at DirectX 10.1 236
Hanners1979 writes "Although we still appear to be some way away from the release of Windows Vista, and with it DirectX 10, specifications for the first point release of the 3D graphics API, DirectX 10.1, have already been finalised and largely made public. Elite Bastards looks at what's new and what will be changing in this release, set to become available not all that long after DirectX 10 — There's more to it than you might imagine."
Hopefully... (Score:2, Insightful)
DirectX does not seem good for the industry (Score:5, Insightful)
Why is directX still tied to windows? (Score:4, Insightful)
If I were a DX developer I'd be more interested in playing "ubiquitous developer" than "Windows Sock Puppet".
I may get modded down for this comment, but honestly, what is so special about windows that makes DX infeasible to implement for other platforms?
Tom
Re:Would it be that difficult... (Score:2, Insightful)
No company should want DX 10+ (Score:3, Insightful)
DirectX shuts out porting of games to LINUX/Mac (Score:5, Insightful)
Until somebody writes a game that does something on LINUX/MAC that can't be done on windows because of the underlying OS that is successful I doubt if we'll see any change.
Re:2 years for adoption (Score:5, Insightful)
Considering [...] DirectX10 is only available on Vista and that 50% of employers say they are not going to purchase [soon], it's a safe bet to say that we won't be seeing any games [...] for at least 2 years.
I think I found the flaw in your logic. Employers != Consumers.
The fact is, games will probably drive Vista adoption more than any other factor save factory pre-installs. We proabaly won't see much requiring DX10 for a year or more, but that is because most big games take 1-3 years to develop so that's about the earliest that we'll see stuff.
This may cause game manufacturers to change tactics since OpenGL is supported on ALL OS's.
That fact has always been true, and it hasn't made much of a difference so far, even back when OpenGL and DirectX were much closer in abilities (without needing extensions and such).
Tired of these articles. (Score:2, Insightful)
OpenGL vs. DirectX (Score:3, Insightful)
The best thing about DirectX 10.x... (Score:5, Insightful)
Re:Why is directX still tied to windows? (Score:2, Insightful)
Clearly a workable strategy. End users don't care one jot about the OS, but what they can do with it, and Microsoft have been very good at pouncing on those opportunities and communicating what you can do with Windows.
"what is so special about windows that makes DX infeasible to implement for other platforms?"
90% of gamers use Windows?
Microsoft owns both Windows and DirectX and wants gaming on PC to stay a going concern, and is the only company in a position to make a sizable difference. Microsoft has made gamers an active priority, while their competitors in the OS market haven't.
Re:DirectX does not seem good for the industry (Score:5, Insightful)
OpenGL is alive and well. It would be great if some of the Windows developers started using it though, since they are in the majority. Please, feel free to join the rest of the world.
Re:Would it be that difficult... (Score:3, Insightful)
Personally, I can't wait to see how well displacement mapping will make real-time terrain generation vastly simpler and adaptive to level of detail (doing this now is a fair amount of work.)
Not much of a business case for OpenGL ... (Score:3, Insightful)
If it were in a developer's best interest to use OpenGL they would. OpenGL has a history of having mediocre drivers if you are *not* doing things as Quake does them. In other words OpenGL was of such little interest to ATI and NVIDIA that about all the optimization attention it got was whetever Quake used. Now this was a few years ago and things are better now but developers remembers this and are a little gun shy due to "spotty" support and optimizations. They all know Direct3D will be at the forefront of ATI and NVIDIA's efforts. Now consider the arguments made by other posters where the new features and tools show up first, in Direct3D.
Again, what's in it for developers? Linux gamers? No they dual boot or emulate, they are already Win32/Direct3D customers. There is no new money to be made, a port would merely move a sale from Win32 to Linux, more work, no revenue. The Linux market is really only those who refuse to emulate or dual boot. Mac OS X gamers? Well at least they have a history of spending money going for them, at least when emulation and dual boot were not feasable since an emulator had to emulate the CPU not just a gaming API. However with the switch to Intel dual booting is now an option, and to make things more confusing there is Cider for emulation. Write for Win32/Direct3D and link in Cider to translate the Win32 calls to Mac OS X. I like OpenGL, I come from a scientific visualization background, but come on, there is not much of a business case from a developer's perspective "today". It had slightly better case "yesterday"
Please, feel free to join the rest of the world.
Uh, by "rest of the world" you mean the 5% running Mac OS X and Linux? Hey, if you are discussing soccer then phrases like "rest of the world" are meaningful, but in the context of computer gamers it is a joke.
Re:Not much of a business case for OpenGL ... (Score:2, Insightful)
Not so simple (Score:4, Insightful)
Re:DirectX does not seem good for the industry (Score:4, Insightful)
Well, you can run games using older versions of DirectX in Vista, so how exactly is this a problem. DirectX 10 isn't out yet and you're complaining about the lack of DX10 cards? ATI and nVidia are on development cycles and they would obviously not include DX10 support on current graphics cards, simply because DX10 wasn't mature when today's architecture was under development. I am very sure both nVidia and ATI are dying to get DX10 cards out as soon as Vista is out, since selling the first fully Vista compatible cards is a huge thing. You can't expect today's tech to support something that isn't even released yet. That doesn't make sense.
Even then it makes me wonder what the point is. When games like WoW get all the headlines using technology that barely require DX8. Maybe if you are into FPS games this will matter but for everyone else the featureset you are calling dated far outpaced the software that runs on it years ago.
Excuse me, but are you even aware of how many of today's games rely heavily on DX9 technology? You are also saying that games like World of WarCraft barely use any DX9 tech. Do you have any documentation on that? Also, what is "games like WoW"? In fact, many "games like WoW" today use pixel shaders and therefore require DX9. Yes, FPS games are clearly taking advantage of the latest technology, but the simple reason is that realism is far more important in such games than, say, strategy games. But on the other hand, there are so many types of FPS titles. We have racing games, action games, MMORPG and even adventure games. World of WarCraft may not be the most graphically advanced game, but this is for three main reasons: it's huge and would simply require too many gigs of space if it was more complex than it already is. Blizzard is also targeting a larger mass who isn't constantly upgrading its hardware as much as some other groups do. Last but not least, have you ever thought of how much more time and money Blizzard would have to invest to include the latest technology with the highest level of detail?
I'm pretty skeptical of how OpenGL will survive in games now that Carmack has sold out. It's a bleak future for those of us gamers who want to someday drop Windows and use a real OS.
You know, Carmack has all the money in the world already. I doubt he would kiss Microsoft's ass if OpenGL was more promising than DirectX. After all, as a DirectX developer, you also get to release your games for the Xbox without too much hazzle, which is only one of many reasons why DirectX is more successful. OpenGL is in no way a poor API, but most developers currently support DirectX because it looks more promising than OpenGL. Also, Carmack was one of the main donators to the OpenGL foundation. Why would he fund a project like this and then kill it? I'm sure it was painful but as Carmack is always set on developing the latest in technology, he is also looking at the two API:s individually before deciding what to go for.
Re:Not much of a business case for OpenGL ... (Score:4, Insightful)
Also, since when is the entirety of Linux considered stable? Half the time the drivers are up in the air, and there's enough competition and conflict between various builds and organisations as to achieve bugger all on a standards level. If game companies started porting to a stable Linux core, they'd be porting to Wii - THAT has a Linux OS with Opera installed.
DirectX is not just graphics (Score:2, Insightful)
If there was a good overall package that leveraged OpenGL for graphics, then you'd see OpenGL being used more often. At the moment there's really only SDL, and to be frank, while SDL is great for some things, it's just not on the same footing as DirectX having come late to the party and not had the level of funding and development.
Re:No company should want DX 10+ (Score:3, Insightful)
Why would any company want to lose out on the win98,2000,XP crowd when they market their game? Only Microsoft has any interest in selling stuff that uses DX10+. To me DX10+ is dumb, stupid, and inane.
People said the same stuff about DirectX 9, DirectX 8, DirectX 7.... you get the idea.
Corporations realize $$$ when they can market the newest, fastest, shiniest whatever. For PC games, this is especially true - how realistic a games graphics are drive sales, and often make a game more fun.
More importantly, programmers will want to use DirectX 10. IMHO, the biggest improvement so far seems to be the elimination of "capability bits" - flags a programmer can query to see what features a GPU supports. The implications of "optional" features that video cards may or may not support means two "DirectX 9" cards can render things very, very differently and make life difficult for the programmer. Features not supported by the video hardware are automatically emulated in software by DirectX, but that is much slower and bugs the crap out of people who dropped $500 for speed. The elimination of capability bits in favor of "dot standards" lets developers code for a specific flavor of DirectX, knowing all cards made for that flavor will behave the same, resulting in better code.
Virtualization of the GPU is also interesting. It applies the same time-slicing multitasking operating systems use to run multiple programs (semi)simultaneously to the graphics hardware. This means that if the GPU chokes on an instruction for whatever reason (i.e., a page fault, the needed texture is compressed, etc.) other threads and processes can continue drawing. Currently, in multitask-less DX9, a page fault chokes the CPU until the needed page can be loaded, whereas DX10 would allow other taskts the GPU was working on to continue.
Better yet is "predicated rendering", which is French for "putting an 'if' in front of a drawing command." Predicated rendering allows the hardware to ignore a command to draw an object if it's not visible - i.e., a very sophisticated hardware clipping.
DirectX 10 has amazing new features and performance enhancements, and (so far) looks like programming with the new API will be much easier and faster. That means cheaper development and happier coders. Doesn't sound so inane to me.