from the embrace-extend-extinguish-again? dept.
After over a decade of fighting OpenGL, Microsoft has finally accepted that it might be useful for something after all and has joined the Khronos group to get involved with its development. This seems to be about WebGL more than anything else, so it doesn't look like Direct3D is going away just yet, but it's still unexpected. Others are as surprised by this as I am; one comment about the announcement, quoted in the article, claims "Hell just froze over."
From The Register:
Microsoft has quietly joined an industry party building 3D interactive graphics boosted by graphics chips, after 14 years spent in opposition.
The software giant has left jaws swinging with a decision to join the Khronos Group and knuckle down immediately as member of the WebGL working group.
There was no official announcement, just a simple tweet on Sunday from Khronos president and NVIDIA developer ecosystem vice president Neil Trevett.
(Score: 5, Insightful) by quitte on Tuesday August 12 2014, @06:19PM
embrace, extend ...
(Score: 2) by CoolHand on Tuesday August 12 2014, @06:25PM
exactly what I was thinking...
*grumble*
Anyone who is capable of getting themselves made President should on no account be allowed to do the job-Douglas Adams
(Score: 2) by kaszz on Tuesday August 12 2014, @06:43PM
Embrace step taken, 2/3 to go..
(Score: 3) by Vanderhoth on Tuesday August 12 2014, @06:19PM
Do I really need to say more than what's in the subject?
"Now we know", "And knowing is half the battle". -G.I. Joooooe
(Score: 2) by Hairyfeet on Tuesday August 12 2014, @11:22PM
Sigh. I really do get tired of guys acting like its 1998 and MSFT is still the big dog instead of the toothless old hound it actually is now but just FYI you have more to fear from Google in that dept [arstechnica.com] than you ever do from MSFT, which just a few minutes of thought would make perfectly cear...mobile outsells desktops by a pretty large margin and has for awhile now, MSFT has pretty much zipola when it comes to mobile presence while Google has become what MSFT was in the 90s, a single corp controlling the OS on (last I checked, it may be higher now) over 85% of the mobile devices being sold. For a successful EEE to work one has to have a way to tie it back to their market in order to profit and between PC gaming being almost entirely DirectX and MSFT having Blackberry numbers in mobile there is just no way for MSFT to pull any kind of EEE on web based tech, they just don't have the numbers.
BTW mark my words but in 3 years or less I'm gonna be screaming "TOLD YA SO!" and doing the dance of intellectual superiority all over the web when all those that bought the "Don't be evil" bullshit get ran over like a runaway train and sit there gobsmacked as Google becomes just as douchey and monopolistic as MSFT in the 90s. I don't think it can be helped really, its like these corps get to a certain size and then fear of losing what they have and of competition blindsiding them becomes more important than innovation or staying ahead of the curve and THAT is when they start getting really slimy, with the EEEs and backroom deals and whatnot.
ACs are never seen so don't bother. Always ready to show SJWs for the racists they are.
(Score: 0) by Anonymous Coward on Wednesday August 13 2014, @08:32AM
Yeah, let's just let MS to do it's shit, because google might be a bigger enemy. I don't know, hmm, who is collecting unverified patent licenses on android now? Google? No, that's MS. MS has not changed, get that through your head.
(Score: 2) by prospectacle on Wednesday August 13 2014, @03:12AM
The unwritten step between "Embrace" and "Extend" is "Dominate", and if they can't do that then they can't Extend very successfully, and therefore can't Extinguish the standard version.
A perfect example of this is the near-extinguishment of the (real) web in about 2001, with all the microsoft only features and ActiveX extensions. They could only do this because they were far and away the dominant web browser, and now that they're not, they're far more standards compliant and their ActiveX plugin system is little more than a joke for most website developers.
As they're coming late to this webgl party, I doubt they'll be able to get past "Embrace", a step which is good for everyone (maybe even microsoft).
If a plan isn't flexible it isn't realistic
(Score: 0) by Anonymous Coward on Wednesday August 13 2014, @09:03AM
I think the new triple-E weapon is DRM. We also see this on sites depending on Silverlight to show movies (thus effectively disallowing Linux). Also the push for DRM in HTML5 (of course in a way allowing incompatible proprietary implementations). I'd not be surprised if they pushed for WebGL to include DRM as well (you know, to protect the poor creators from evil pirates stealing their WebGL creations, it certainly has nothing to do with lock-in to the system providing the DRM).
(Score: 4, Insightful) by Hairyfeet on Tuesday August 12 2014, @06:29PM
What a great idea, lets run unsecured code from unknown parties on a device with 1.- its own high speed processor that most folks don't know how to monitor, 2.- has its own RAM that no AV that I know of scans, 3.- and has its own storage (firmware) which will not only not be scanned by AVs but can also actually brick the hardware.../facepalm/
I thought ActiveX was seriously stupid but at least ActiveX couldn't put the actual hardware at risk. GPUs are designed to run much more "bare metal" because of how big a performance hit you get when you add layers between it and the gamecode and not only are the modern GPUs more powerful than the CPUs we were running not too long ago but malware writers have an incentive to target them thanks to AltCurrency being lucrative...man this is a BAD idea!
ACs are never seen so don't bother. Always ready to show SJWs for the racists they are.
(Score: 4, Informative) by opinionated_science on Tuesday August 12 2014, @06:33PM
go see my browser DRM rant...
(Score: 4, Interesting) by tibman on Tuesday August 12 2014, @08:06PM
This isn't direct access to the metal. WebGL is just a JavaScript API. Some things are passed through obviously, vertex arrays, shaders, and so on. Almost every modern browser implements WebGL natively (no plugin required). IE is the exception in that IE 10 and under require a plugin.
If you have JavaScript enabled and you visit a website, they could be using WebGL. http://caniuse.com/webgl [caniuse.com] -> shows what browsers support it
I have worked with the threejs library. It's pretty nice! Here is an example picked off their site (js required): http://ogreen.special-t.com/en/ [special-t.com]
SN won't survive on lurkers alone. Write comments.
(Score: 0) by Anonymous Coward on Tuesday August 12 2014, @08:32PM
http://lmgtfy.com/?q=web+gl+exploits [lmgtfy.com]
(Score: 3, Informative) by tibman on Tuesday August 12 2014, @08:59PM
You may have to abandon the visual internet if you want to use a 100% secure product, service, or application. You can pretty much google anything plus exploit and come up with a full page. Even images have been exploit vectors (scary!). I don't see people screaming from the rooftops that we need to disable images on all websites.
That being said, no security conscious person likes seldom used features like WebGL siting around to activate without user permission. But then those people don't exactly have that problem because they don't even allow javascript to run unless they trust the page.
SN won't survive on lurkers alone. Write comments.
(Score: 2) by Hairyfeet on Tuesday August 12 2014, @10:32PM
That is like saying "since we can't stop thieves from stealing money all banks shouldn't have guards and be on the honor system". Just because nearly anything, given enough time and effort, can be hacked does NOT mean that running code from just anywhere on a device with its own processor, memory, and storage which is NOT screened by AVs and which has lower access to the system than most thanks to graphics drivers taking a serious hit with added layers, is not a GIANT BAD IDEA. Again look at ActiveX, it was "just sitting there" and your argument would work just as well by simply replacing WebGL with ActiveX but I think we can all agree it WAS A BAD IDEA!
And the real bitch is its NOT the users pushing this shit, because for Joe and Sally normal the old way of simply downloading a game, installing a game, and running a game? Works just fine. Nope this is being pushed by the same ones that push "Cloud 3.0 Powered by Mountain Dew" because it will allow you to run games on devices more locked down than an iPad and give them another way to insure content is delivered as desired...be it to do datamining, use as another form of DRM, I can't see how its supposed to be in any way good for the users who I have no doubt will have to deal with crap like lag and overhead that they wouldn't have to if the game was downloaded and run, but I can see plenty of pluses for publishers....just another BAD IDEA being hoisted upon us by those with $$ in their eyes, bleech.
ACs are never seen so don't bother. Always ready to show SJWs for the racists they are.
(Score: 2) by tibman on Wednesday August 13 2014, @01:48PM
WebGL is a specification, not an implementation. There is zero DRM, it is a text document. http://www.khronos.org/registry/webgl/specs/latest/1.0/ [khronos.org] The WebGL implementation in Firefox is open source. It is not a closed piece of garbage like ActiveX. And yes, it is users pushing this. People do not want to install a POS flash plugin just to watch a video. They want an open specification to describe a resource and their browser to implement that spec. Bonus points for an open source browser.
You keep refering to WebGL as if it is an active thing, it is not. It does not "run" any code in the normal sense. The closest you can get to running code is applying a shader, which is like a very small C program. That little C program is compiled client-side from source provided by the server. It can calculate colors, move vertices, and small range of other things. It has a VERY limited control structure available. It cannot access outside memory. The spec does describe safe operation.
Would you feel safer downloading (probably closed source) executables to your system? Only if you trust it. Would you rather download the source code, compile it, and run it in a container that prevents any direct hardware access? That is exactly what JavaScript is.
SN won't survive on lurkers alone. Write comments.
(Score: 2) by cafebabe on Wednesday August 13 2014, @03:55AM
I was hoping that GPU mischief would be limited to rude words and pictures. However, I investigated DMA over PCI Express [infosecinstitute.com] and I found that the answer is complicated. All PCI Express devices can potentially initiate DMA. And, due to PCI bridging, devices with 32 bit drivers may access memory outside of the 4GB limit. So, the limiting factor is whether the graphics card memory space is paged (as viewed from the CPU), whether the graphics card can initiate DMA and, in particular, whether the SIMD section of the graphics card can initiate DMA. The drivers may not support it but I cannot guarantee that all past, current and future GPU designs are safe.
1702845791×2
(Score: 2) by kaganar on Wednesday August 13 2014, @02:31PM
What a great idea, lets run unsecured code from unknown parties on a device...
Same as running any program I download -- even if it's signed by a software company, should I really trust them? (EA games? Microsoft? SmallUnknownBusiness? WildTangent? Free Software Foundation? Individuals?) The key difference is that I must decide to download, install, and execute such software. With web technologies I do not -- I click the wrong link and it starts up right away -- and that's the real flaw.
WebGL and other cross-platform easy-delivery instant-content enabling technologies aren't going to go away -- web developers like them too much, so end users like them too much. (All my friends thought Flash was great, the steaming pile of crap it was in the 90's, because people made cool content with it.)
People creating new features for web browsers have a chicken-and-egg issue -- if the feature isn't enabled, people won't use it and find existing insecurities. But if it has existing insecurities, it shouldn't be enabled and people won't use it.
Right now new features get pushed out off a cliff -- implement it, test it, beta test it, goes live with many edge cases unexplored. At best the feature can be toggled somewhere obscure in an options screen before it's rolled out for real, but most end users aren't going to be savvy enough to really use that.
Browsers ought to have a classification for its feature sets. When visiting a webpage that use features classified as likely more insecure than a configurable threshold I ought to receive a confirmation about if I'd like to visit this site and which features I'd like to enable if I do. (Maybe there's a browser or extension that does this already?)
(Score: 2) by tibman on Wednesday August 13 2014, @05:15PM
I agree, having everything on by default is a bad idea. At least in FF you can go to about:config and disable WebGL. Not sure if there is something similar in Chrome and IE.
SN won't survive on lurkers alone. Write comments.
(Score: 2) by kaszz on Tuesday August 12 2014, @06:50PM
An "open standards" group headed by one of the most notorious suboptimal graphics manufacturers? Well optimal for NVidia, suboptimal for everyone else.
What could ge wrong except everything when this is combined with the Embrace-extend-extinguish Inc....
As for Active-X and other crap. Let's just abandon it cold turkey.
(Score: 1) by Wootery on Wednesday August 13 2014, @12:59PM
If that's the situation, surely it's good that there's another multi-billion-dollar party involved with an interest in opposing nVidia's mischief, no?
(Score: 2) by kaszz on Wednesday August 13 2014, @01:16PM
I wouldn't rely on that.
(Score: 1) by Wootery on Wednesday August 13 2014, @03:54PM
I'm not too worried. Neither nVidia nor AMD are fully 'loyal' to the 'cross-vendor' APIs of OpenGL and Direct3D. See: nvapi, Mantle. Despite AMD and nVidia being in direct competition, OpenGL seems to progress. Apple has its own 'Metal' API now, too, but their Khronos membership doesn't seem to have caused issues.
Intel, nVidia (even with their own CUDA), and AMD are all interested in OpenCL (also a Khronos API), but again, good stuff has happened there.
Of course, Direct3D is a much bigger competitor to OpenGL than nvapi/Mantle/Metal, so I admit Microsoft is unique here.
(Score: 2) by kaszz on Wednesday August 13 2014, @05:44PM
Ah, the usual scenario where every company tries to twist the standard to hinder customers from migrating ever and to be "uniquely incompatible" ;) Guess we don't have to worry it's all balkanized anyway. What is killer software for a 3D API ..? (except games optimized to insanity)
Thinking, make killer software on API, that would force vendors to comply or stand aside.
And how is the availability of OpenGL on Microsoft-Windows platforms?
(Score: 1) by Wootery on Thursday August 14 2014, @09:50AM
nVidia seem [nvidia.com] to have beta support for OpenGL 4.5 on GeForce 400 series (released 2010) and above. Unclear whether older OpenGL 4.x series APIs are unsupported, or already stable.
AMD aren't terribly clear [amd.com], but OpenGL 4.1 appears to be supported for Radeon HD 6800 (released 2010) and newer [amd.com].
(Score: 2) by kaszz on Thursday August 14 2014, @02:58PM
Perhaps standards compliance is where the dirty laundry is ;)