Slashdot Mirror
New Cross Platform Alternative To DirectX
BlackVomit writes: "
There's a bunch of companies such as 3dfx, 3Dlabs, ATI, Compaq, Discreet, Evans & Sutherland, IBM, Intel, S3, and SGI that have formed a special interest group called Khronos to design a cross platform API for graphics, video, and audio. This is very cool, as it could be a huge leap for gaming on Linux as well as all platforms that choose to implement the API. Imagine games that work seamlessly on Winders as well as Linux/Unix, BeOs, Mac, etc. I am somewhat surprised that nVidia isn't in on this.
" Let's just hope they work with the other open-standards projects for these things. The promise of a "an industry wide, non-proprietary approach" just screams out for it ...
Just a kind of 'off the wall' rumour I heard. Do you think there is any truth to it?
For someone wanting to do this for "Simple DirectMedia Layer supports Linux, Win32, BeOS, along with ports in progress to MacOS, IRIX, Solaris, and FreeBSD." I've played around with it myself, and was actually planning on using it, but a project I was going to do it for fell through. But check it out, it supports OpenGL... But it would also be cool if they support OpenAL in the future... http://www.devolution.com/~slouken/SDL/
So microsoft is not in on this standard, so what makes people think they would implement it under windows?
Let's face it, economically the most important platform is windows, so if it is not supported, the standard is not really cross-platform.
There's GTK/GDK, for the Unix and Linux platforms. And DirectX for Windows.
What we need is a merger of both, using similiar function calls. Or maybe something that simulates DirectX for Linux.
However this is still great news, compared to the recent coverage of WAVE America and the first anniversary of Columbine fast approaching. This library will finally give us the respect we need.
Careful attention must be given to the process of making this library, or it could turn into a proprietary project run by one corporation, almost just like DirectX.
US businesses that currently accept chip and PIN/signature
Someone explain to me what "Winders" is?
I can't imagine why they WOULDN'T support OpenAL.
:)" endl;
LokiSoft wrote the SDL. LokiSoft is one of the main forces behind OpenAL. See the connection?
if (LokiSoft == SDL && LokiSoft == OpenAL) {
cout "Isn't that amazing?
}
Yes. OpenGL _IS_ cross platform, not _was_.
OpenGL runs on pretty much all desktop systems. (Can anyone think of a system that doesn't have OpenGL?)
At least Khronos is doing The Right Thing (tm) and using OpenGL as a base for OpenML. OpenGL is a nice clean, and othogonal API unlike that bastard child called Direct3D. (DX = 6 revisions in 4 years? Yeah, thats a good design! NOT.)
I'm kind of curious as to why OpenAL or SDL doesn't fit the bill here?
http://www.lokigames.com/development/
At least they have enough "big names" that this might actually take off.
<rant>
Unlike all the developers that SOLD OUT to M$ with Direct3D. As a 3D programmer I'm still upset that everyone let M$ shove down our game programmers throats wheter we wanted it or not.
</rant>
SDL, Simple Direct Media Layer, by Sam Lantinga (not Lokigames, but he his their head programmer) is supported in news groups and mailing lists, and is a very well documented API that is already in many shipping products from Loki games.
:)
I cannot remember all of the operating systems and platforms it compiles and runs under, but let's just say that it allows you to interface with TWO renderers under both Windows AND Linux, plus OpenGL and OpenAL support. Also, it is endian-portable, as it works on the Mac as well.
It also supports all of the other major facets of DirectX: DirectInput, DirectSound, oh wait, that's about all the other major facets.
About the only benefit that I can see from this new API is a unified driver programming interface. Right now, SDL interfaces with renderers elsewhere. This means many different types of drivers must be programmed in order to have your hardware work cross platform a'la SDL.
PS. SDL is lgpl'd.
SGI and NVIDIA may have partnered, some, and exchanged engineers and technology, but they, as far as I know, are still separate technologies and companies.
SGI does systems and solutions, where NVIDIA does components; they can coexist and cooperate, but they are not(yet) interchangeable.
-AS
-AS
*Pikachu*
I'm not the kind that goes around bitching 'bout things but to see heavy hitters such as 3Dfx and SGI among others involved in a neat project like this...and to have such a lame page (look at the graphic...they could have done something a little more attractive). .pdf, yuck...
And they have some of their docs in
Kudos to them for the neat project anyhow!
.oo00OO
games work all over
not just on bill's piece 'o crap
let's hope this happens.
Haven't we all seen this before? Remember Java? Sure, the backing that this project has received is great, but can we really be foolish enough to believe that this will lead to the "write once, run anywhere" model we all envision? I think that this is certainly a step in the right direction, and I hope that they are completely successful, but I take all of this with a HUGE grain of salt.
"The only source of knowledge is experience" -A. Einstein
Why isn't nVidia working with this? Simple. For the same reason 3dfx pushed glide for so long -
Because right now, nVidia is winning.
Thus far, there's no indication that any of this stuff will exist under Linux, BeOS, Mac, etc.
I mean, with the intent of the API, it could very well <em>exist</em> under every platform out there... but someone has to implement it. I expect to see this firstly on Windows machines, since all the companies seem to have big investments in the Windows world.
Essentially, this looks like OpenGL++, with extensions and support for video, sound, and more interaction and synchronization details. So Linux won't have it, unless there's a workalike(think MesaML or something)
I would wonder if the API is open source and available, but relies on the individuals involved to provide closed binary level compatibility? There's nothing saying that this is a great thing for consumers... what this is, is a great thing for developers and companies that sell hardware, to no longer be tied or controlled by Microsoft or any other organization.
Now, if all these corporations adopted Linux or BSD or something like that and produced a <em>reference</em> implementation, the way Sun produces a reference Java, or SGI did a reference OpenGL... But I fear, for market acceptance, that we will see a reference implementation under Windows.
Keep our fingers crossed!
-AS
-AS
*Pikachu*
Hmm, what does this mean for this guy, who seems to be trying to fill up Open*L with the rest of the alphabet? :)
b .html
http://people.we.mediaone.net/seanpalmer/openli
<tutorial level='novice'>If you preview your post, the entities get switched to the corresponding symbols in the new page's submission box. If you're going to use HTML entities and want to preview your post, hit your back button after the preview and make any needed changes there. Then submit your post. Or, if you like doing things the hard way, go to the submission textarea and correct all the altered entities after you preview.</tutorial>
Yes, it's annoying, but not that hard to deal with.
Cheers,
ZicoKnows@hotmail.com
Gaming on Linux is still something of a lackluster area
More realistically, gaming on any computing platform other than Windows is still something of a lackluster area. Mac may be proportionately better, but it still pretty much pales in comparison to the titles available to Windows.
- Scott
------
Scott Stevenson
Scott Stevenson
Tree House Ideas
Although I hope it doesn't go the way of the pet rock and so many of the other many-company initiatives we hear about all of the time. Here are a few items of advice, if they want to supplant the horrible beast tyrant that is DirectX:
1) Don't just go after video/audio! Many game developers use DirectX not for the wonderful (ahem, *sarcasm*) APIs in Direct3D, but for the DirectInput and DirectPlay APIs. That way they know that people will be able to use their new "Force Feedback Webcam Modem Ultra Joystick With Added Internet Buttons" with their three week old game that wasn't coded for it.
2) Update the release often to support new technology, but try to keep the APIs from getting beat to death. This is a fine line to tread of course, but it is one that Micros~1 has partially succeeded at. Most game players don't mind upgrading DirectX, since it can result in better performance in games they already have, and it's on their game CDs anyway. The chief problem with DirectX is that the Direct3D API is *very* ugly. Memo to Micros~1: it's bad enough to version your function calls with "Ex", but please don't append numbers to them! Memo to the media: you all seem to believe that games make kids kill other kids. Is it possible that these kids happened to come across the Direct3D API, read it, and flipped out? Investigate! Maybe we can use some filtering software to keep the kiddies from seeing the Direct3D API on the web.
3) In spite of #1, the only performance anyone cares about is the 3D. Go look for reviews on how many giga blams (gb) per second you can get out of a sound card in EAX vs. A3D. I bet you won't find much. While good APIs are invaluable to developers, they won't use your APIs unless you can at least provide comparable performance to the competitor. I know, the APIs don't theoretically dictate performance, but in reality, the design choices of the API can have a serious effect on the ability to incorporate future performance gains due to the latest hardware techniques.
Anyway, that's all I can think of for right now. Hope it becomes a reality! It'd be great for Linux, but what it will really be best for is the future OS that none of us have even dreamed up yet, that will be able to have support for games sooner due to some standard APIs.
err, I don't know about the Linux drivers etc.. But nVidia is a bastard at releasing specs for their cards.
Example, alot of developers who have been creating a driver for the GeForce card for BeOS ended up hitting a roadblock, how to access the framebuffer IIRC. Well many emails to nVidia later, nothing, It's as if they don't care about the people who use their cards. And all they are asking for is access to creating a 2d driver ONLY! I doubt there's any trade secrets in the 2d code, but apparently nVidia just doesn't give a damn about non-ms os's. So I say screw nVidia, for as much as I like there video cards, I hate the way the company treats it's users choices of os's.
Therefore the TNT card in my system will no longer be replaced with a GeForce card unless nVidia gets it's collective head out of it's ass and releases specs so users could support their favorite os, be it FreeBSD, Linux, BeOS, whatever os you're using.
Matrox here I come, just don't pull a nVidia on me after I get a G400.
I work for a game developer, I'm just the type of programmer who would have to use such an API. Maybe I lack the needed level of contempt for Microsoft to appreciate this initiative, but...
W iderInTheLeft). An API has to mature for a few years (like good wine), and I'm pretty sure they'll get it wrong the first time.
This is not going to work... if it appeared three, maybe four years ago, they did stand a chance of success, but now the DirectX team has handled practically all of the developer's complaints about it (well, maybe except for the function names like CreateDirectXSurfaceCoveredWithTinyDotsAndALittle
What's the developers' incentive to use OpenML instead of DirectX (except for "cuz windoze sux" and "cuz bill gates is satan")? Game publishers just stare blankly when you mention the possibility for non-Windows versions of a game. And if doesn't catch up in games, it will always stay a niche technology (as, say, OpenGL before Carmack).
One more thing: who's missing from the list?
NVIDIA, who are the driving force in graphics hardware for the last year and most probably for the NEXT few years;
Apple, makers of QuickTime, a very sensible media architecture;
Creative, the largest (and now that Aureal has quit the game, the only) manufacturer of enthusiast-level audio chips;
SEGA, Sony and Nintendo - makers of the most widely spread "graphics, audio and video media devices"
In the faq, they mention Windows Media and QuickTime as alternatives. I don't think this is like directx at all. I think this is for making and playing movies and such. (does anyone actually follow those links?)
Here is the link to get you started: http://www.khronos.org/SIGFAQ.htm
Do the names Khronos and OpenML have any significance?
Khronos is the Greek root word for time. As the specification tackles integrating different media types, a critical element is synchronization of the media. Open signifies a cross-platform, cross-OS, initiative. ML or Media Language, emphasizes the synergy with the existing OpenGL standard, and helps to amplify an industry wide, non-proprietary approach.
Well, that last part sounds like a ray of hope for getting an open source implementation going in Linux. But, it sounds like the license is still be hashed out.
Another interesting part of the FAQ:
Who will define and maintain the OpenML specification?
The Khronos Group will retain this responsibility as well as license and distribute the OpenML SDK and reference implementations.
Acutally, that is were it gets a little scary. It sounds like now they are a little confused about whether it will be close or not. Also in the FAQ it states that the project will incorparate QuickTime??? Maybe as a binary plugin or something. And if you really want to 'what if' you can start thinking/hoping/praying that Apple's going to open source that one again. But, don't anyone hold your breath.
Well, everyone cross your fingers that this will be open and take off.
This ignorance is really getting to me
/. this week. And that's really saying something.
Networking doesn't have an api other than sockets?
Uh DirectPlay offers an easy API for game networking, it's 'network transport' independent, works over TCP/IP, IPXSPX etc etc. It's an API so you can essentially use simple commands to send messages, and not have to worry about creating your own communications protocol over TCP/IP etc.
resources in supporting a standard that changes almost twice a year...
That has to be one of the most ignorant things I've read on
DirectX is entirely backwards compatible (ever heard of COM versioning?) all new additions of DirectX are done thru implementing a new interface IDirectDraw3, IDirectDraw5, etc. A DirectX 3.0 game will work fine in DirectX 8.0.
I guess I shouldn't expect more from people who don't know anything about Windows programming except that 'APIs change twice a year, Microsoft is evil, so they must'.
So lame.
starts out kinda crappy in version 1 (or, in this case, horribly crappy) and then with each successive version, gets better and better, until it takes over the world.
Don't get me wrong, I love OpenGL, and I've never done a 3D project in anything else. However, D3D isn't the nightmare it once was... it's rather usable now, and there's plenty of docs and sample code available now. Unfortunately, it's not cross platform, which keeps most of my projects away from D3D, but it's perfect for games that don't feel the need to be cross-platform, and wish to take advantage of the best developed 3D drivers on the Win32 platform (the OpenGL drivers are usually a joke).
BTW, the numbering after the functions is standard COM protocol now. The idea is, if you want to update an API, but the changes you'll introduce are really kinda different from the existing API, it's best just to start anew with a new object, and leave the original alone (so nothing old gets broken). Hence the numbers, which refer to new objects. Yeah, there's some bloat, but it's better than breaking old stuff. The only solution is to design something really really good in the beginning, but that idea isn't really compatible with bringing something to market fast, fast, fast.
For all those who want to see more and download the files and help out, here is the homepage: http://www.devolution.com/~slouken/projects/SDL/in dex.html
Everytime you look at porn a devil gets their horns.
Everyone comparing this to SDL is missing something. From the list of companies involved this is going to be more about video editing and compositing packages and high end 3d animation packages. i.e. standards for streaming D1 or HDTV video streams to a disk array in realtime, or standards for streaming from disk to graphics with guaranteed framesrates.
This could be a good thing for support of those applications on Linux/BSD because if hardware vendors manufacture video cards, disk arrays etc to open published standard like this it will be much easier to write drivers. At the moment on the NT platform every maker of digital video cards has their own standard for accessing the video and streaming it to a disk array.
DirectX is NOT used in this market, it's a games API and apart from some sound editing applications it's not used in content creation packages.
One of the companies in the announcement is SGI and they are widely expected to come out with linux based high end 3d workstations sometime this year so I'd expect to see a Linux implementation of these standards although it may not be OpenSource.
It's a shame that DirectPlay simply Did Not Work until version 6.1a. To clarify this statement: all the bits of DirectPlay that weren't essentially wrappers over a socket-like interface, were too buggy to use. I believe 6.1a finally did mostly work.
The problem is that a good half of DirectPlay is abstracting over something that doesn't need to be abstracted over. It's not like each network card needs to be addressed in its own particular way from the application point of view. This is already handled by the operating system.
As an API, it's also very poorly designed, almost completetely lacking in orthoganality or layer distinction, and having some particularly odd choices of abstraction.
Of course, the most important failing of DirectPlay is that it is platform specific; there is no reason for a closed networking protocol - if you're talking to another machine it shouldn't matter what operating system it is running. DirectPlay essentially forces all participants to be Windows 9x/NT machines. For massively multiplayer games in particular, it's very possible that you don't want your servers running on NT. Further, it means one can't release a game for multiple platforms and have them talk to each other, or the same server.
I'd still recommend that a game developer avoid DirectPlay entirely, and roll their own protocol instead.
You all really don't get it, do you? the real platforms which do matter for the people who are going to DEVELOP for this OpenML platform, are THOSE platforms which are used by the MAJORITY of their potentional CONSUMERS, namely the people who are going to buy and play their games.
Ever wondered which platforms?
for starters: NOT Linux on PPC, NOT OpenBSD or MSX.
These platforms are the CONSOLES and consumer windows (win9x, winME). Write an API that works on THOSE platforms and you're set. Why? Because the gamedevelopers are developing games for the money, for selling products. They are only able to sell their products if the potentional consumers are able to use it and as a producer you only make any profits if you sell a lot.
The initiative is very ok in theory, but only if it's for the platforms where the most gamers are on. This mostly excludes Linux, no matter how much you probably hate that.
Futhermore, as some clever reasonable minds have pointed out already: an api has to mature before it's even USABLE. And if that time arrives, is it a BETTER alternative for what's available ALREADY.
If I assume this time will arive around 2002, I'm pretty sure the alternatives are unbeatable. It's sad but true. (O, and perhaps you think Linux is THE desktop OS... that's highly unlikely.)
--
Never underestimate the relief of true separation of Religion and State.
First things first, 3Dfx hardware does not suck (not as much as you imply, anyway). The problem with nVidia's method of porting the windows drivers over to their new (closed) linux API is that when (it's no longer an if IMHO) when nVidia stops developing drivers for linux, there will be two crufty, closed source drivers to reverse-engineer, not just one that plugs into the standard DRI interface.
This is GOOD for Linux because it means that bug fixes in the Windows drivers will apply to Linux as well. It won't introduce any sort of incompatibility because it is all contained in a standard XFree4 module.
This is also BAD for Linux because bug-fixes that change the interface fill break the nVidia driver. Besides, x86 Linux is not the only non-windows OS out there. What if the *BSD, Be, x86 solaris, etc. users want 3D acceleration? Right now, they're SOL.
WTF?! Since when do we expect companies to release open source drivers?
Since 3Dfx, Matrox, and ATI have open-source drivers (or at least published specs) WTF shouldn't we expect nVidia to play along with the rest of the industry?
0 1 - just my two bits
Probably a "revenge moderator." You know, someone out to get someone they disagreed with in an earlier post. Irritating, but I'm not sure what can be done about it.
All the creatures will die, And all the things will be broken. That's the law of samurai. (Jubai, 1605)
I would just like to point out that this was out a week ago on OpenGL.org. What took so long? Aside from that bitch, I think this is a revolutionary API. Some people are saying that this isn't really a DirectX competitor and that you should use SDL or something for games. Wrong. First, DirectX is not just a games API, it is applicable to all media apps. (I think even MS Office uses DirectDraw.) Second, it the state of OpenML is similar to OpenGL. It was originally intended for pro apps, but then people realized that it was great for any 3D app. OpenML will probably be the same way, although the companies behind it now have some hindsight. (Remember, the OGL ARB is on the board.) Third, this IS going to be cross platform DirectX, whether or not its creators intented it to. There is nothing like DirectX on any other platform, and DirectX is a big reason why people don't start developing for alternative OSs. Not just because of the API, but what DirectX means. Direct access to hardware without hardware dependance and without conflicting with other apps. BeOS has some of these APIs in a fairly unified manner, but they aren't as featurefull as DirectX. Linux has nothing remotely close to DirectX. SDL is nice, but is inherently slower since it is layered on top of native APIs. Also, it does not take the world's best advantage of hardware acceleration. Face it, most hardware today are DirectX accelerators. If a large board like the Khronos board can pressure developers to include OpenML acceleration, a fast, powerful cross platform API that can compete with DirectX can emerge. Finally, SDL has nowhere near the features of DirectX. DirectX is addictive. It is the only thing that makes me keep Windows on my harddrive, even though I like programming BeOS much better. It is the only API I've ever seen that can blit 21,000 icon sized sprites in the space of one second. Even OpenGL can't touch it for flexibility. The HAL/HEL is a work of genius, combining the power of hardware acceleration with the niceness of device independance. Lovely. However, it is limited to Windows, which is its one major flaw. Thats why I think OpenML will be so revolutionary. If it can match DirectX, (And I'm sure it will, after all, the OpenGL ARB is onboard) then I can give up windows and game developers can get all the features and power they're used to with DirecX with the freedom to port easily to different platforms. On a larger scale, OpenML will essentially fill in a large gap in POSIX. POSIX has become a very well accepted cross platform API, but is limited in usefullness because of its lack of media/graphics/etc features. This is a major thorn to game developers, because aside from code unique to the application, most of code is graphics or sound API code. By making this API cross platform, we can now see cross platform apps that aren't cut down to the least common denemenator of the OSs that it supports.
A deep unwavering belief is a sure sign you're missing something...
Yeah, Linux is still a niche market, but Linux games are making money. Loki isn't going out of business any time soon.
There is only one thing Linux needs to start to get out of the niche gaming market and into the main stream, and that is a killer game app that every gamer wants but that doesn't have a Windows port. Heck, the game could be packaged with a Linux distribution and partitioning tools, I bet any gamer would love to try it. Gamers are not cowardly little computer users, they have no problem messing with their systems to get that little bit of extra performance out of them. I doubt a userfriendly Linux distribution would intimidate them. People who use their computer for fun are more likely to mess around with it than people who just use it to do work and surf the Web. Of course, you may be an arrogant Micros~1 employee who believes that no one knows about computers and nobody wants to. If that true, why are their even PC gamers at all why doesn't everone just get a Playstation, the epitome of the non-technical users game machine? Even with Playstations, how many people have modified them to play games from other regions (or had one of their tech savvy friends do it? The same guy they'll have set up their Linux partition?). Or maybe you think that the people who pay $60+ for a PC game are just "casual gamers?"
Here's hoping you end up in one of the unprofitable Baby Bills when the government breaks up your company.
All the creatures will die, And all the things will be broken. That's the law of samurai. (Jubai, 1605)
Gamers could care less about the Operating System, if a game company produces great games for Linux that everyone wants, gamers will get Linux and stick it on a parttion of their Windows machine. It's not hard, since Windows and Linux are not exclusive, (they can easily exist on the same machine,) there's no reason not to do it.
The real truth is that Windows existence depends on three things in the consumer market, that it has the best Web browser (and I mean significantly better than the Linux Web browser), the best hardware support and that the best games are still written for it.
In the PC gamer market, however, the only thing Windows has going for it is that the majority of good games are still written for it. If a PC gamer sees he can get all the same games he can on Windows, and a few really good, exclusive Linux games on Linux, he/she will switch to Linux (though probably keep Windows on a partition if he still has games he wants for that.) Remember a PC gamer builds his machine specifically for gaming. This means picking out modems and graphics cards with good support for his OS of choice. The only reason Windows is the OS of choice for gaming systems is that it has more games. (In fact it isn't even the Windows OS, it is the DirectX API that means games will run on the system.)
Gamers are not crawling PC sheep (those are a seperate group of users), they go where ever the best games are. Micros~1 had better pray and fight tooth-and-nail to keep Linux from becoming a viable gaming platform, because once it is Windows days are numbered in the consumer market.
All the creatures will die, And all the things will be broken. That's the law of samurai. (Jubai, 1605)
This reminds me of one of johnc's plan updates, only he put it better, I think. Paraphrased and from distant memory: "I'm confident that Direct3D (IM) will continue to suck less with each subsequent release, but why go through that painful process?"
--
Fuck the system? Nah, you might catch something.
I agree. In this particular case, though, there is a reason to be pissed off at nVidia. The reason is that they lead us to believe they were Linux-friendly and that drivers were just around the corner.
--
Fuck the system? Nah, you might catch something.
Of course they're going to revise the DirectX API.
Four years ago, the gamut of hardware graphics acceleration features was quite limited. Texture mapping. Displacement mapping. Environment mapping. MIP mapping. Alpha mapping. Specularity mapping. (The only people who had heard the word "specularity" in 1990 were writing raytracing dissertations on SGIs and Crays for SIGgraph.) Of course the API has had to grow in that time.
Microsoft's OS sells a lot of games. Sue them. (Oh, you are? Fine.) In 1983, there were not 150 million computers out there with twitch-happy game playing buyers. Microsoft phased out MS DOS because it was way too limiting; it needed to add SOME games-oriented APIs to Windows. Maybe not the best, but it beat writing a 700Kb real-mode app trying to page 128Kb of video ram at a time.
Hardware manufacturers are optimizing and revising their whole approach to many features, and the API tries to both abstractify and follow these implementations. That leads to a very ugly API but at least the programmer has access to the hardware's features.
The DirectX API keeps generation-numbers on the API so that old games aren't obsoleted. When a new generation of API comes out, the whole universe of available hardware features has changed. Bugs fixed. Concepts abstracted and simplified.
The DirectX API has tried to incorporate software emulation of each compelling feature that a hundred hardware vendors add to their products. This (1) enables more games to sell, since games aren't dependent on the ONE current card that supports wizzy-feature-foo, and (2) enables more hardware to sell, since players want to upgrade from emulated-wizzy-feature-foo to accelerated-wizzy-feature-foo.
Hardware comes out every couple weeks, and people don't buy hardware if apps can't use it. Hardware producers don't like funding the software development process, but need to get their hardware sold.
Games live on a nine-month life cycle. (18 month gestation, nine months on the shelf, not too good, eh?) The game industry needs to reinvent itself every season, or game companies die. Diablo is antiquated before Diablo II can be released.
(Anecdote: one reason Win95 went to unidrivers instead of hardware-supplied monolithic drivers was ATI's horrible track record for writing software. ATI released to all customers over fifteen versions for the same rev of the same card within a month or so. Each driver rev broke some OTHER feature of Windows 3.1 GDI, regressed back to bugs they'd fixed earlier, and rarely fixed the bugs they were trying to fix.)
Six revisions may seem big to you, but it's not unusual for that industry. I wish they did fewer revs, too. Then again, think back: Wolfenstein 3D, DOOM, DOOM II, Quake, Quake II, Quake III.
(Six revisions in four years =anagram>I vary, or in seriousness, fix.)
[