OpenGL in HL2

[mrchimp]

I think it's worth mentioning that Linux is also used for scientific simulations and purposes and often uses OpenGL to visulize the results. If the companys that sponsor OpenGL droped it they would have problems. So there is a market for OpenGL on Linux even if it's not games, also alot of embedded devices use OpenGL or OpenML (i think it's called) because they can only run Linux on there hardware.

Unless OpenGL 2 comes out soon and is very good, I doubt many people will be useing OGL for games but for other applications there will be a market for a very long time.

 

[Mountain Man]

*nix was not designed with the prospective idea of playing games, it was based off of the idea of creating a free and open Unix, which was an operating system used primarily by engineers for servers and other mission critical computers.

It wasn't originally designed to be a competitive desktop OS, either. It's interesting how some things can transcend their initial design parameters.

It has so few games because to properly piece a game together you have to use many scattered libraries to do the various parts, ensure they're installed on the various machines, and all sorts of things. That's not to say it makes it incapable of playing games, however it places a considerable hurdle between it.

I can only assume your comments are not based on practical experience.

I'd rather people just try and not make it out to be something it's not. Linux is a free and open Unix, which is an operating system for engineers, not the average everyday joe, I doubt it ever will be, because to do that it wouldn't really be linux anymore.

I can only assume your comments are not based on practical experience.

Like my earlier comments about using graphics capabilities in OpenGL, it's not that it can't do it, it's that it's so much harder and requires more work to do it that way.

I can only assume your comments are not based on practical experience.

As far as I'm certain, the closest to this is the SDL, which while it does cover many aspects, lacks all the bells and whistles DX has. From the standpoint of a programmer, I can tell you DirectX looks infinitely more appealing then any other I've seen so far.

John Carmack apparently disagrees. Not to mention the fact that companies like Bioware and Epic among others have had little difficulty in using non-Microsoft API's to create cross platform games.

Cross platform support is insignifigant unless you have a real market you'll be able to make a profit off of on that platform.

Companies like id, Bioware, and Epic among others obviously recognize Linux as a viable market.

 

[nc17]

Right now DirectX is much more mature then mature OGL, OGL was the standard back when hl1 was released but it has not really advanced much since then

If OGL was updated now it would most likely be far superior to DX in all areas

 

[Fallout2man]

It wasn't originally designed to be a competitive desktop OS, either. It's interesting how some things can transcend their initial design parameters.

If it succeeds at being easy to use and competetive on all levels as a pure desktop OS, then it most likely will cease to be linux in that sense.

I can only assume your comments are not based on practical experience.


I can only assume your comments are not based on practical experience.


I can only assume your comments are not based on practical experience.

I can only assume your lack of a proper response is based on not having any real way of responding to those points.

Also, allow me to elaborate, The SDL does a lot, but it can't say, do everything OpenGL does, it can't offer everthing OpenAL does either. It doesn't offer Dolby Digital Surround sound or any number of things.

John Carmack apparently disagrees. Not to mention the fact that companies like Bioware and Epic among others have had little difficulty in using non-Microsoft API's to create cross platform games.

As I've said before, Carmack is a nut. He's a coding genius but he's got a few screws loose in his head. If you don't believe me read the PC Gamer interview on Doom III, this guy has no clue about anything outside of his beautiful art (making 3D engines). He's the guy who doesn't care if he has to program all his shaders in assembly, he'll do it, and it'll take ten times longer but the end result will probably be better whenever he reaches it. He seems to have an irrational hatred of microsoft, and considering he holds a good percentage of say over the small but powerful iD software, I think he can get away with any extra delays required in staying away from anything microsoft written. As for Doom III I honestly have to wonder if aside from the OpenGL, if they didn't code the entire rest of it from scratch without any library to assist them, though it's possible they may have used OpenAL as well I've never heard any info on that front.

That also said Carmack is one of the lucky few to recieve Beta Drivers for OpenGL 2, which is something most people don't have the advantage of. So Carmack can program for this fabled new core OGL update that may, for all we know, solve every current issue/problem/hiccup/etc with the current 1.x's forray into mostly requireing vendor specific extensions programming for anything nice looking just about. As for Epic, they started with one engine back before 3D engines were that popular, back before even DirectX (before most anyone used it at least), Unreal was openGL first, they simply updated their existing code base as time passed, which isn't something that hard to do compared to writing a new engine from scratch. So they never had nearly as much of a time-penalty and because of that they could easily make a port to other systems.

Companies like id, Bioware, and Epic among others obviously recognize Linux as a viable market.

As for Bioware, I've honestly never heard of them releasing anything for linux, also while iD and Epic may make *nix ports, I've never heard of any boxed copies made. Which to me would make sense, due to the fact that if they have the capability to make a cross-platform game without the aformentioned costs due to the above mentioned situations, they'd do it because despite the small marketshare they'd still make a profit since it wouldn't cost them anything. If they (the actual companies, not third party companies doing it on their behalf) are making boxed copies then that's an entirely different story.

 

[Punzz]

You ppl know too much about this sort of stuff

 

[Fallout2man]

You ppl know too much about this sort of stuff

What can I say, I'm bored and an insomniac

 

[Mountain Man]

If it succeeds at being easy to use and competetive on all levels as a pure desktop OS, then it most likely will cease to be linux in that sense.

Do you even know what you're talking about? Linux is just a kernel. The interface that sits on top of it can be made as user friendly as you like. The folks working on the KDE project are doing an excellent job in that regard. You seem to be trapped in the Winblows world where the graphical user interface and the operating system are inseperable. Tell me, what exactly do you consider to be the "real" Linux?

As for Linux games:

Unreal Tournament 2003 shipped with a Linux client in the box, and UT2004 will even include Tux the Penguin on the box cover! :thumbs:

Bioware has ported Neverwinter Nights to Linux. You have to download an installer, however.

Savage by S2 games includes a Linux version in the box.

All of id's games have been available boxed for Linux.

And so on. I think you get my point. While OpenGL may not necessarily be as up to date as DirectX (a point that remains debatable), it seems premature to put it out to pasture.

At the same time, I don't understand why you're so hostile towards the concept of cross platform support. It doesn't take anything away from your precious Windows "experience" and it gives the world viable options that don't all involve filling Microsoft's coffers.

 

[Fallout2man]

Do you even know what you're talking about? Linux is just a kernel. The interface that sits on top of it can be made as user friendly as you like. The folks working on the KDE project are doing an excellent job in that regard. You seem to be trapped in the Winblows world where the graphical user interface and the operating system are inseperable. Tell me, what exactly do you consider to be the "real" Linux?

Physically it's just the kernel, the result of a project started in the early 90's by Linus Torvalds, but it has its whole stigma and community associated with it. The fact is that while it may still use the same kernel it won't be linux in the spirit of what linux originally was. Take lindows for instance, It may use the kernel, but it's really not linux as far as everthing else is concerned, it's something else, something new. That's my thought anyway.

As for Linux games:

Unreal Tournament 2003 shipped with a Linux client in the box, and UT2004 will even include Tux the Penguin on the box cover! :thumbs:

Providing they included it on the same disk, this probably cost them nothing extra, in which case it'd be silly for them not to, as it's pure profit.

Bioware has ported Neverwinter Nights to Linux. You have to download an installer, however.

I'm not familiar with the engine they used for neverwinter nights' graphical capabilities, so I'm not qualified to give a real response, however I'd guess it falls under the same circumstances as the two scenarios I mentioned earlier. Either someone there hates microsoft and has the power to get the extra time taken, or other circumstances naturally had them developing what could easily become cross platform, in which case it costs them zero and they'd be silly not to.

Savage by S2 games includes a Linux version in the box.

See above comment and the one about UT2K3

All of id's games have been available boxed for Linux.

As i said before, Carmack's an anti-microsoft nut, he has a lot of say at iD, so this makes perfect sense.

And so on. I think you get my point. While OpenGL may not necessarily be as up to date as DirectX (a point that remains debatable), it seems premature to put it out to pasture.

Oh I never said it's done for, it can do everything DX can, it's just more time consuming because of how you have to do it, the core itself needs an update, and providing OGL 2 comes out, this might give it the boost it needs to come back out on top like it was before. Until then however DX has a very big advantage as far as a developer is concerned. I was saying however that if they don't change their ways and get more timely with updates and doing things to make it easier to use these capabilities, it may damn itself. It can still come back on top, but that depends entirely on what the OGL board decide on.

At the same time, I don't understand why you're so hostile towards the concept of cross platform support. It doesn't take anything away from your precious Windows "experience" and it gives the world viable options that don't all involve filling Microsoft's coffers.

First of all, I'm speaking as a programmer, and for practical purposes cross-platform support, like many other such things, needs a good cost/benefit justifacation to consider. A lot of things go into making a game and deciding what to use, unforseen circumstances can occur and you have to be careful or you may find yourself up shit creek without a paddle. I'm not against cross-platform support, but I am not going to endorse what's a generic statement not founded in the business-reality of software.

The fact is you develop for different platforms when the market can justify the costs, you have to compare tools, audience, market share, and development times carefully before you can decide if it's worth it. I currently saw what amounted to a lot of people saying "OGL is free and good and runs on linux and macs, let's all use only that." and I just wanted to ground the thing more in terms of reality. The idealistic approach of free is better is great for a hobby or if you're working on developing an infra-structure you know will make you more money long-term, but for most business simply won't work that way, they need the best and most cost-effective solution, sometimes free just isn't worth it. Well, that and I love a good debate

 

[Fallout2man]

Since the time limit on editing's expired (why did the admins enable it? honestly what good does it serve?) I'll add this, which I probably should've explained earlier. By grounding in terms of reality I mean I want people to think twice before demanding everything become ported to X OS (replace X with whatever OS you want) just because they use it or don't use the OS it was designed for. I see it so often on slashdot it just gets to me a bit.

you don't see me asking for windows pors of Final Cut Pro, iMovie, and Pro Tools.

 

[rebb]

LoL, i didnt expect this thread to still be going.
Impressive .

 

[Brian Damage]

Yes, it's a fine example of post-disection.

 

[Mountain Man]

You said:

First of all, I'm speaking as a programmer, and for practical purposes cross-platform support, like many other such things, needs a good cost/benefit justifacation to consider.

But in response to the fact that Epic includes a Linux version in the same box as the Windows version you said:

Providing they included it on the same disk, this probably cost them nothing extra, in which case it'd be silly for them not to, as it's pure profit.

So what you're reallying saying, there is profit in cross-platform support!

I want people to think twice before demanding everything become ported to X OS (replace X with whatever OS you want) just because they use it or don't use the OS it was designed for.

This is ignoring the fact that there are legitimate reasons to promote alternatives to Microsoft Windows.

 

[mrchimp]

http://nehe.gamedev.net/data/contests/2004_creative/demo/florian_rudolf.zip

hhhmmm this demo was made in glslang and it's one of the most realistic looking uses of vertex shaders I have ever seen.

Since I last posted I'v been learning a bit of ARBfp to try and make my own smartshaders or at least modify the ones made by ATi and I must say, considering what it does ARBfp is not as tedius as basic assembler language that you would use for running instructions through your CPU. makeing shaders in ARBfp is probably quite a bit harder than useing a high level language, but it isn't a major difficulty for an experienced programmer.

If I was going to make anything that used 3Dgraphics I would definately go for OpenGL, even if it meant slightly more work. Also downloading the DX9 SDK is a pain in the arse just to try it out.

Here's an ATi smartshader

shader edgePixelShader = 
"!!ARBfp1.0
# Uses Sobel Edge detection algorithm

PARAM  texCoord00 = program.local[0];
PARAM  texCoord01 = program.local[1];
PARAM  texCoord02 = program.local[2];
PARAM  texCoord10 = program.local[3];
PARAM  texCoord12 = program.local[4];
PARAM  texCoord20 = program.local[5];
PARAM  texCoord21 = program.local[6];
PARAM  texCoord22 = program.local[7];
PARAM  negOne = {-1.0, -1.0, -1.0, 0.0};
PARAM  one = {1.0, 1.0, 1.0, 0.0};
PARAM  negTwo = {-2.0, -2.0, -2.0, 0.0};
PARAM  two = {2.0, 2.0, 2.0, 0.0};
PARAM  grayScale = {0.30, 0.59, 0.11, 0.0};

TEMP   filter1;
TEMP   filter2;
TEMP   coord00;
TEMP   coord01;
TEMP   coord02;
TEMP   coord10;
TEMP   coord12;
TEMP   coord20;
TEMP   coord21;
TEMP   coord22;

OUTPUT oColor = result.color;

ADD coord00, texCoord00, fragment.texcoord[0];
ADD coord01, texCoord01, fragment.texcoord[0];
ADD coord02, texCoord02, fragment.texcoord[0];
ADD coord10, texCoord10, fragment.texcoord[0];
ADD coord12, texCoord12, fragment.texcoord[0];
ADD coord20, texCoord20, fragment.texcoord[0];
ADD coord21, texCoord21, fragment.texcoord[0];
ADD coord22, texCoord22, fragment.texcoord[0];

TEX coord00, coord00, texture[0], 2D;
TEX coord01, coord01, texture[0], 2D;
TEX coord02, coord02, texture[0], 2D;
TEX coord10, coord10, texture[0], 2D;
TEX coord12, coord12, texture[0], 2D;
TEX coord20, coord20, texture[0], 2D;
TEX coord21, coord21, texture[0], 2D;
TEX coord22, coord22, texture[0], 2D;

MUL filter1, coord00, negOne;
MAD filter1, coord10, negTwo, filter1;
MAD filter1, coord20, negOne, filter1;
MAD filter1, coord02, one, filter1;
MAD filter1, coord12, two, filter1;
MAD filter1, coord22, one, filter1;
ABS filter1, filter1;

MUL filter2, coord00, negOne;
MAD filter2, coord01, negTwo, filter2;
MAD filter2, coord02, negOne, filter2;
MAD filter2, coord20, one, filter2;
MAD filter2, coord21, two, filter2;
MAD filter2, coord22, one, filter2;
ABS filter2, filter2;

ADD oColor, filter1, filter2;
END";


shader copyPixelShader = 
"!!ARBfp1.0
OUTPUT oColor = result.color;
TEMP pixel;
TEX pixel, fragment.texcoord[0], texture[0], 2D;
MOV oColor, pixel;
END";

surface temp = allocsurf(width, height);

edgePixelShader.constant[0] = {-ds_dx, -dt_dy, 0, 0};
edgePixelShader.constant[1] = {0,      -dt_dy, 0, 0};
edgePixelShader.constant[2] = {ds_dx,  -dt_dy, 0, 0};
edgePixelShader.constant[3] = {-ds_dx, 0,      0, 0};
edgePixelShader.constant[4] = {ds_dx,  0,      0, 0};
edgePixelShader.constant[5] = {-ds_dx, dt_dy,  0, 0};
edgePixelShader.constant[6] = {0,      dt_dy,  0, 0};
edgePixelShader.constant[7] = {ds_dx,  dt_dy,  0, 0};

texture[0].source = backbuffer;
destination temp;
apply edgePixelShader;


texture[0].source = temp;
destination backbuffer;
apply copyPixelShader;

This shader is called Sketch and looks like one of the photoshop filters.

 

[Mountain Man]

If I was going to make anything that used 3Dgraphics I would definately go for OpenGL, even if it meant slightly more work. Also downloading the DX9 SDK is a pain in the arse just to try it out.

Ineed. There has to be a reason why a genius like John Carmack would prefer OpenGL other than him being an anti-Microsoft nut job. I mean, he still writes his programs for Microsoft platforms, so there has to be more to it than that.

 

[jonbob]

I mean, he still writes his programs for Microsoft platforms, so there has to be more to it than that.

With respect, he'd be mad not to write software for the people who use the platform that constitutes 9/10 parts of his target audience. Why doesn't someone e-mail him and ask why he chose OpenGL over DirectX for Doom3? But then again, if he's the "anti-Microsoft nut job" you say he is, I don't expect he'd have an unbiased view to give us.

[email protected] for whoever wants to know, at least that was the address that Google (in all its wisdom) provided.

 

[Fallout2man]

You said:


But in response to the fact that Epic includes a Linux version in the same box as the Windows version you said:


So what you're reallying saying, there is profit in cross-platform support!

I'm saying there CAN be, but it's not always there, I'm trying to destroy the myth it's always good and should be done 100% of the time to everything. Epic's a case scenario where it costs them zero extra to make a port so they can do it, your average developer will not be able to make that same zero-cost port when building an engine from scratch. Things have to be considered a lot before making these sorts of decisions, there can be profit but it's not always there.

This is ignoring the fact that there are legitimate reasons to promote alternatives to Microsoft Windows.

Legitimate reasons and legitimate reasons in the eyes of a business are very different things. The fact is that they only should be doing a port if they can make more then it'd cost, in a lot of cases usually some third party asks to port the game, and sell it, in which case most companies that aren't too paranoid about their source code, usually accept that because they get a hefty licensing fee in one form or another from the company doing the port. Whether or not it succeeds or fails they still get that licensing money, so they could care less about the rest.


The fact is that "promoting windows alternatives" is just not always in sync with the bottom line. A business is about profit, not the ethics of bringing down a monopoly. Honestly The only point I'd ever see to switching away from windows entirely is if palladium couldn't be cracked (fat chance, everything can be cracked, and the hardware vendors are all backing out of the hardware part of the solution). If i did such a thing though, you wouldn't see me asking for Half-life 3 or X software to get ported over, I'd be capable of admitting I'm using an OS that's just not designed for games. So unless there's some magical worst-case scenario, I imagine I and many others will be using and programming for windows for a long time. That's not to say things can change, but they don't seem very likely to change at this point, only time will tell.


As for figuring out DX, I found it a bit steep to learn but once I got the idea of what was what and got more accustomed to the OO orientation of it, I was able to do things just fine, and as for Gslang, it may be a powerful opponent to HLSL, however not just yet, the Gslang compiler included in ATI's drivers just isn't quite good enough yet for it to be used for practical purposes, as well there's no similar compiler in nVidia's drivers yet, which there'd really need to be for it to become on the same footing as HLSL.


If you want to promote windows alternatives, then you're more then welcome to make your own (non-windows only or whatever else you want) game and release it under the GPL or whatever license you choose, just please don't make unrealistic demands of other developers.

 

[Mountain Man]

Point is, you have not presented a single legitimate reason why supporting OpenGL and multiple platforms is a bad idea. The very existence of such games proves that all your arguments are in some way flawed. I'd just like to see more developers take the lead of companies like id, Epic, S2 Games, and Bioware.

 

[GRIMEY]

You're right.

He's right.

 

[Fallout2man]

Point is, you have not presented a single legitimate reason why supporting OpenGL and multiple platforms is a bad idea. The very existence of such games proves that all your arguments are in some way flawed. I'd just like to see more developers take the lead of companies like id, Epic, S2 Games, and Bioware.

my points are scattered throughought my posts, so I'll make a quick summary.

OpenGL: It takes too long to work with, when 2.0 comes out things may change, but until then it's second fiddle to D3D, which comes as part of a nice integrated windows solution.

Multiple-platforms: Can be good and bad, depends on what you're making, it falls into the trap of "can we make more then it'd cost?" In terms of games, if you allready have an engine that could be multi-platform at little to no additional cost, likelyhood is it'll make a profit. Epic as stated had allready had an existing OpenGL engine and they simply added DX later, because of their original forging and gradual updating of an engine before DX they were able to do so at little to no cost. Generally the very poor market share isn't worth the extra cost as far as games are concerned, as they'd never recoup development costs. The fact is OpenGL itself doesn't mean cross-platform, you have to use cross-platform sound, music, input, and net code. That's a big task usually when you're faced with using a bunch of different libraries versus one complete solution like DX.

Other types of software do have the type of market they need, for instance adobe makes photoshop cross-platform beause the business market for their graphic design program on other platforms is more then enough to justify the costs. Games just aren't there yet, and I don't think they will be for a good while. A real competitor to ALL of directX, not just one aspect, needs to be made before cross-platform support will start becoming easy enough to justify the costs. Right now there are a few libraries that are poor or okay competition with only one DX aspect, such as DirectSound or Direct3D, not all aspects, it's more then that.

 

[Mountain Man]

See, the thing you keep forgetting is that there are game developers who are making cross-platform products using OpenGL (among other APIs) and are seeing a profit from it. That fact alone is sufficient to nullify your arguments. My only hope is to see more developers follow this lead.

 

[Fallout2man]

See, the thing you keep forgetting is that there are game developers who are making cross-platform products using OpenGL (among other APIs) and are seeing a profit from it. That fact alone is sufficient to nullify your arguments. My only hope is to see more developers follow this lead.

And see, you keep forgetting that there's a thing called circumstances. The fact of tha matter is it can be profitable but it is NOT always profitable. I explained why both Epic and iD, which I know at least a little about, have chosen to do what they've done, I never knew much about savage of any of the released Bioware products and such, so I can't comment on that. However what you seem to be saying is that if one person jumps off a cliff and lives, we all should just run and jump off that same cliff right after him, without actually seeing why that one person lived and if that'd apply to anyone else, because that's what it amounts to.

You can't do this always you have to anylize and make the best decision and the fact is it just isn't always the approach you keep trumpetting. Yes there are special case scenarios that do allow for it, however they're just that, special case scenarios.

 

[Mountain Man]

However what you seem to be saying is that if one person jumps off a cliff and lives, we all should just run and jump off that same cliff right after him, without actually seeing why that one person lived and if that'd apply to anyone else, because that's what it amounts to.

See, kids, this is what we call a straw man argument. What you do is rather than actually address your opponent's argument, you create a different but weaker argument, defeat your made-up argument, then pretend you defeated your opponent's argument!

You can't do this always you have to anylize and make the best decision and the fact is it just isn't always the approach you keep trumpetting. Yes there are special case scenarios that do allow for it, however they're just that, special case scenarios.

Wouldn't it be neat if native Linux games were no longer considered "special case scenarios"?

 

[crabcakes66]

See, kids, this is what we call a straw man argument. What you do is rather than actually address your opponent's argument, you create a different but weaker argument, defeat your made-up argument, then pretend you defeated your opponent's argument!

that pretty much sums up all you have done in this entire thread. :dozey:

 

[Fallout2man]

See, kids, this is what we call a straw man argument. What you do is rather than actually address your opponent's argument, you create a different but weaker argument, defeat your made-up argument, then pretend you defeated your opponent's argument!

*shrugs* I was bored

Wouldn't it be neat if native Linux games were no longer considered "special case scenarios"?

Neat, yes, realistic, no.

I also think you misunderstand what I mean by special case scenarios. I'm implying that specific circumstances came together to negate the downsides of doing native ports for those games, so when there is no downside and only profit obviously you do it, however those circumstances only exist for those companies, not all of them. Obviously if there was never any downside required and only profit, every corporation would do native ports for everything.

Yes MS does have a habit of atrociously anti-cometitive behavior, however from the application/game developer's stance, providing you're doing it for money, is to maximise profit while still making the game you want. Microsoft's abhorent behavior has created a situation where over an estimated 90% of all desktops use Windows in one iteration or another, and obviously the application/game developers develop for that first in many cases because it maximises their profits. For any extra development for additional platforms it has to be decided if the benefits justify the costs of making a port. The benefits do not always justify the costs, and that's the truth of the situation. Yes This situation is the cause of microsoft, however you can't expect to hold all application/game developers responsible for what MS did, they just need to be able to make a good profit for what they do, and if that means they can't afford to develop for every platform then that's just how things go.

So yes, in an ideal world every piece of software would run on every OS, and every piece of hardware would run every OS, however in the world we live in that's just not the case right now. If you want to change things then by all means go ahead and set the sterling example, but please don't demand everyone follow suit.

 

[bobvodka]

meh, it was only 2 days ago so i'll correct something, MS pulled out of the ARB sometime last year (or maybe even the year before) since they have stopped blocking everything things have progressed.

The Core of OpenGL has been updated numerous times, thus OGL1.5, its only the lib and dll which MS should provide which hasnt been (wanna guess why children?)

technicaly speaking the OpenGL API is more mature than the DX API, its been stable since its conception which was a LONG time ago, DX on the other hand has had various changes to the API

OpenGL 2 isnt important (well, it is but it isnt, meh, cba to explain go read the OGL2.0 specs as to why it wont be magical), the fact is we have the important bits already really (fast vertex submission, HLSL for PS2.0... other... stuff... i'm tired, sue me), infact I think with OGL1.6 we might get something D3D doesnt have... uber-buffers (read/write to any buffer, multi-color buffers, all kinds of fun things)

As a side note I really wish i could find the document i read a while ago which showed that OpenGL was twice as fast on vertex submission as D3D, and this wasnt some fanboy thing, it was an ATI/Nvidia presentation

for other comments see the other Opengl vs d3d thread ( http://www.halflife2.net/forums/showthread.php?t=19374 )

 

[Fallout2man]

I've heard of this OpenGL 1.5, yet never seen it, figured it was just some development version or something. As for having what it needs, does it have a unified (Works on all hardware) high level shading language? CG is Nvidia only and GSlang is ATI only (as well as extremely inefficient, ATI really needs to optimize their GSlang compiler).

As for more Mature, yes that's very true, however mature doesn't necessarily mean better, just that it's been around longer.

The buffers however do interest me, one of the things I've been a bit miffed about in DX are all of the restrictions in what you can/can't copy/read/write/etc between.

 

[bobvodka]

My ATI drivers are currently showing an internal version of 1.4, however based on the OpenGL Extension Viewer ( http://www.realtech-vr.com/glview/ ) they are full 1.5 core complient (the number tends to lag a bit).
The spec tends to get updated about once a year, its then upto the IHV to write the driver portion of the code, from a user point of view not much fuss seems to be made because you average gamer doesnt care, and devs tend to keep their ears and eyes on the right places to notice these things, so they slip out year after year.

True mature doesnt always mean better, however i'd argue that stable means better, still i'm led to belive that very little changed between DX8.1 and DX9.0 API wise so maybe MS have sorted it out now

The buffers thing i also found intresting, it might make it into the next version (heres hoping, a year is a long long time)

Couple URLs which might be of intrest;
http://www.opengl.org/documentation/opengl_current_version.html
ARB Meeting notes from Sept where they talk about the future of OpenGL on the windows platform : http://www.opengl.org/about/arb/notes/meeting_note_2003-09-10.html
ARB meeting notes from Dec where next version is talked about (Also some minor 'super buffers' stuff) : http://www.opengl.org/about/arb/notes/meeting_note_2003-12-09.html

Next meeting is this tuesday/wednesday... wonder if these notes will get out quicker..

 

[Fallout2man]

In the card drivers? I thought you had to install OpenGL seperately? That's the last thing I remember having to do dealing with it, and having a hell of a time finding what I needed too.

I'd say plenty changed, the only problem is it wasn't all for the better, my main gripe is the D3DPOOL restrictions in the UpdateSurface and UpdateTexture member functions (of IDirect3DDevice9). In DX8 you could copy from and to any texture or surface you wanted, however now the destination and target must be created with specific, different, D3DPOOL parameters (D3DPOOL determines how DX allocates/manages memory for the object). This isn't so bad except when you're trying to do something like I'm doing, which is creating a hierarchy of textures that eventually are going to all end up on a "Master Texture" before being drawn to the screen.

The only really major changes I've noticed, were mostly just related to pixel/vertex shaders, such as the introduction of MS' High Level shading Language.

 

[Arno]

In the card drivers? I thought you had to install OpenGL seperately? That's the last thing I remember having to do dealing with it, and having a hell of a time finding what I needed too.

That was 4 years ago maybe. Today the OpenGL drivers are included with Nvidia's ForceWare and ATI's Catalyst drivers.