Slashdot Mirror
What is Happening with OpenGL?
Trapped In Windows Hell asks: "I was just at the local game store looking for a new game, and I noticed the absolute lack of ANYTHING other than DirectX games. Where has OpenGL gone, and what does this mean for games on GNU/Linux? If DirectX is so hard to program in, so clunky to use, and limits the game to being sold on only one OS, WHY do so many programmers use it? It seems logical (to me, at least) that programming as portably as possible, as simply as possible, and using standards where possible, leaves a lot more sales options open for the future... and DirectX seems to close all options *but one*." OpenGL use in Windows gaming has decreased dramatically in favor of the use of DirectX which is improving with each release. Will OpenGL continue to mature on the Windows platform (which arguably is the platform that drives most of the mainstream demand for graphics) or will it continue to stagnate as game and driver developers concentrate on the offerings from Microsoft?
If DirectX is so hard to program in, so clunky to use, and limits the game to being sold on only one OS, WHY do so many programmers use it?
By the sheer amount of DX games out there, isn't it obvious that the game industry doesn't find DX clunky and hard to program in. And remember DX offers MORE than just graphics. It also does sound, input and networking.
As we've seen this year, game companies don't care whether DX limits them to windows or not because:
1- Almost everyone uses windows to play games on.
2- If they ever did want it ported to other platforms, there are companies (you know who they are) that will do the porting for you (yes, that includes porting DX games).
OpenGL only offers graphics, dx offers: graphics, joystick, sound, forcefeedback, mousecapturing.
There are libraries giving you all these features using GL graphics, but they are not made full time and/or they are made by hobby programmers. The keyword here is time. MS have employed programmers to create, maintain DX. This means better time, and resources making a better product overall. Not just the graphics part. The graphics part of DX is not better than OpenGL, it's just everything around it.
Look a monkey!
Btw, you're comparing apples to oranges here; DirectX is a gaming architecture, while OpenGL a graphics one. You should compare Direct3D with OpenGL. :-)
:-)
Simple answer:
Direct3D is now simpler to use, better supported by video card manufacturers, and more OO than OpenGL.
Long answer:
Up until recently, OpenGL was considered more advanced because it was more high-level. You didn't have to write as much OpenGL code to get stuff to work as you did in Direct3D. (Immediate mode, we're talking about here; only a complete moron would write Direct3D retained mode code, what with the overhead it causes.)
However, with the release of DirectX 8.0, this has changed. Direct3D is now a lot cleaner and easier to use. No more does every card have all these annoying mode bits for every single little thing; the market has consolidated in such a way that the major players all support simultaneous use of the most important stuff. (In the olden days, you had to check mode bits to see if you could use alpha blending and bilinear filtering at the same time, for example. Some cards did it, some didn't. You had to have a work-around in case each of these things failed. This is no longer a problem.)
Additionally, Direct3D 8.0 takes OpenGL's way of looking at things, adopts, extends and surpasses. Ever heard of vertex shaders? Pixel shaders? In OpenGL, these are just extensions that are implemented differently by the video card manufacturers. (If they're available at all..) In Direct3D, they're an integral part of the API. And they'll become even more important in years to come. Here's why:
Anyone remember back before video cards could do stuff on their own, when you did everything on the main processor and pumped it to the video card raw? Well, we're in those days right now with video card design. These processors have very specific things they do; if you want to make a dynamic texture, you'll have to ship it back to the main CPU to process it there. Weighted vertices that use in excess of a certain number of matrices (I'm talking about skinned animation systems here) have to be done on the CPU. The message here is that what the video card can do is very limited, and is "hard coded" in the hardware. There isn't much flexibility here.
Enter "shaders." These are little applications (more like scripts, though, since there are no branches in execution) that you can run inside the video card. Instead of going through all the vertices on the CPU and transforming 'em or whatever on the CPU, you can write a little program called a vertex shader which'll do it for you on the video card. And pixel shaders will do the same for your textures; now you can have zoom effects, warp effects.. heck, you could probably implement Photoshop as a set of pixel shaders. (In fact, I don't doubt they're looking into it as I type.)
You see, the graphics API's no longer just a one-processor API with DirectX 8; Direct3D has become an operating system unto itself! Vertex shaders and pixel shaders are like specialized mini-drivers that you load to access the additional functionality. It's really quite neat! I can't wait to see what the demo scene does with these things; the possibilities are endless!
As for the OO thing, have you ever considered just how many games are OO? Considering OpenGL's very C-ish nature, C++ programmers are easily going to gravitate towards Direct3D, simply because of it's C++ OO design. (And lets not even get started on what happens whenever you throw more than one monitor into the mix. I mean, with DirectX, all you do is use another pointer; what do you do in OpenGL..?) Add in the fact that you'll need DirectX for input and audio either way, and..
Wrapping things up, until OpenGL catches up to Direct3D in terms of it's integration of vertex and pixel shaders directly into the API, (programmers are lazy; they don't want to have to go searching for a frickin' function pointer to access an extension..) it'll be playing second fiddle to Direct3D in the minds of developers. Especially with the X-Box coming out, and everybody and their cow wanting to port to it.
Anyways, carry on..
James
I write 3d game engines for a living, and have been fighting this issue for a decent part of my career.
The OpenGL ARB really doesn't give a crap about games. Sure, there are a number of vocal game advocates, but the majority of the membership is far more interested in maintaining backwards compatability to older SGI and Evans and Sutherland hardware than keeping up with accelerator progress.
If the ARB did care about games, there would be a concerted effort to standardize on vertex and pixel shader instructions between card vendors, and a move to get these into the standard AS FAST AS POSSIBLE, and a push to actively participate in ongoing features. Instead, it took them years to drag in a few interesting extensions, and Microsoft has assumed the unifying role in the gaping vacuum.
As a game developer who has spent too many man-years fighting abysmal M$ API bugs and design limitations since Win 3.0, even I will admit that Direct3D has completely exceeded OpenGL as a 3d game development platform. Why should I invest six+ months tuning seperate nVidia and ATI shader support engine features under their respective OpenGL extensions, knowning that this GL code is barely reusable and is tied to a VERY limited set of cards?
Add to that M$'s role at the ARB, and the influence they throw around with their money to keep other members in line (remember Farenheight?)
Unless the ARB makes tremendous changes in its policy of staying 3 years behind the hardware, I strongly feel OpenGL is relagated to the niche BASIC fell into. Sure, you can get it on all platforms, but its so slow and feature poor, why bother?
I wouldn't hold my breath...
> I think it's only been established that Id didn't do well with the Linux gaming market
All linux games sales EVER don't add up to one medium selling windows title. We are one of the creditors that aren't likely to see money that Loki owes us, so we have some idea just how grim it is.
That isn't saying that it can't change in the future, or that doing linux ports isn't a Good Thing, but it isn't an economic motivator at the present time.
John Carmack