Hey, if you can't beat 'em, litigate 'em to death, I guess...
Yeah, because that lawsuit Microsoft file against Apple is utter crap. Oh, wait. Microsoft didn't file a lawsuit. I guess that's why people bitch and moan when you use the abbreviation M$.
And according to the article, the paper Microsoft's patent was based off of was published in the "2001-2002" timeframe. Given that the iPod was released in "late 2001" by your words, who was first cannot be determined with the readily available information.
LMMH is a typo which should have been LDDM.:) LDDM is short for Longhorn Display Driver Model. Basically, Vista needed new functionality not exposed in old drivers in order to make their new compositing system work.
The same "problem" happens (aero glass isn't enabled) when you don't use an LDDM display driver for your graphics hardware. In other words, if you use an XP DX driver on Vista the system doesn't enable aero glass either. That's why you see some many posts in forums for people asking about LDDM drivers.
The claim was that hardware-accelerated OpenGL in desktop windows was impossible unless Microsoft changes things.
Which is a bald faced lie. Not only can an OpenGL ICD be written which properly "shares" the hardware, XP era OpenGL ICDs still work. Legacy ICDs just force Vista to fallback to the XP method of drawing the display. If all you do with OpenGL is play full-screen games, you won't even notice or care.
Now, I won't pretend that writing such an ICD is "easy", because it probably isn't. But I don't see anything that makes it impossible.
As for the output of hardware being consistent, that doesn't require Microsoft to write all the memory management code themselves, thus hamstringing vendors [...] AFAIK today DirectX doesn't have any real advantage over OpenGL in the consistency department despite the common Microsoft code DirectX
No, it doesn't at all. In fact, I think that will be enforced more through the driver certification process more than anything else...
Do you have any other links that go into more specifics on that?
It's not a problem today when you run 2 or 10 or however many windowed OpenGL apps at once along with some DirectX ones too; somehow Windows seems to cope just fine
It explains the nuances in much better detail than I could.
But I fail to see why they would be making such a big stink in that case.
They aren't. A guy posting to an OpenGL forum is, telling everyone to complain about the situation to the vendors. And there is apparently a guy from 3dlabs posting essentially what I've been saying.
I'll bet these newer driver APIs give vendors much less control over the internals of various parts of the graphics system.
It depends on your perspective. The final output from various pieces of hardware should be more consistent, and I don't think that is an unreasonable goal. What process they use to generate that output is up to the vendor (that's why there is a driver layer to abstract the hardware from the software), so long as they implement the minimum featureset necessary for the checkbox they're aiming for. The biggest developments these days appear to be related to vertex and pixel shaders (and MS seems to be making a big deal about going crazy with them for aero), and that's probably where most of the biggest changes lie.
They're fearmongering and spreading FUD, and obviously don't understand how OpenGL is supported on Windows.
1) The DX wrapper does not add a 50% overhead, unless the calls you're making amount to Noops. Performance with virtualization suffers. The cost of "sharing" the hardware between multiple pieces of software is where the primary perf hit lies. OpenGL isn't the only entity which will suffer in this case. Basically they're complaining that they don't get the perf you'd get if only one app was allowed to use the hardware at a time. 2) Microsoft's implementation will be a vanilla version of OpenGL that conforms to v1.4 of the OpenGL standard. Nothing prevents third parties (including hardware vendors) from implementing their own OpenGL ICD. To put it into perspective, Microsoft's OpenGL implementation on WinXP is a software renderer that conforms to v1.1 of the spec. 3) See point 2.
Re: Last paragraph. That is 100% correct. Vendors are free to implement their own OpenGL ICD that supports the LDDM. However, if vendors do NOT implement an ICD that supports the longhorn device driver model (ie: you use the WinXP model, which does NOT support virtualization of the graphics hardware), the OS reverts to the WinXP model of rendering the display.
Spoken like someone who clearly doesn't know how anything actually works, and someone who clearly likes to point fingers and assign blame.
If the OpenGL implementation you're using on a windows machine is hardware accelerated, you aren't using Microsoft's implemenatation of OpenGL; you are using your hardware vendor's implemenation.
Your reasoning for windows being "broken" is equally absurd. You might as well say all Fords are defective because they don't use Chevy engines. I bet you stay awake at night complaining about how Panasonic VCRs are broken because they don't use your beta tapes.
If you use an LDDM compatible OpenGL ICD in Vista, the system doesn't have to fall back into compatiblity mode and none of the features are turned off.
If you use a WinXP OpenGL ICD in Vista (ie: a driver that doesn't have the goo to support virtualization of the graphics hardware), the system falls back to "compatibility mode" and turns off the stuff that requires virtualization of the hardware.
Expecting it to work differently is like expecting an s3-virge card to play quake3 at 60fps.
Vista's graphics subsystem virtualizes the graphics hardware when running in "Aero" mode (ie: the mode with all of the eye candy); think of this as a multitasking scheme for your 3d card. This mode can't be used with a library which requires exclusive use of the graphics hardware, as you correctly surmised above.
However, when you run an application that runs an XP era OpenGL ICD, the system falls back to the XP-style compositing scheme (ie: eye candy gone). In other words, vender supplied OpenGL ICDs still work.
Furthermore, venders can create LMMH compatible OpenGL drivers, which would contain the right "goo" to share the graphics hardware with the OS.
There is nothing "lcoking" in the OpenGL implementation that can be used on a Windows machine at a pure 1.4 level. Nor is this some conspiracy to make OpenGL "slow" (the previous system supplied implementation was a SOFTWARE renderer only; the only way it could be slower would be if the video card in use has no 3d hardware at all -- in which case you were already screwed).
1) XP era vendor provided OpenGL ICD's force Windows to fall back into a compositing mode compatible with non LMMD drivers. THEY STILL WORK. 2) Vendors can produce LMMD OpenGL ICDs that work with the new composting model in Vista.
The big brooha here is that the "Aero experience" requires support for virtualizing the graphics hardware, which vendor supplied OpenGL implemenations don't support. This requires Vista to fall back to the XP-era style of compositing the display when running an application using an XP era ICD. If vendors implemented an OpenGL driver that supported virtualization we wouldn't even be having this discussion.
Of course, you won't find any rational discussion here pointing that out...just the normal/. MS bashing FUD.
XP era vendor supplied OpenGL implementations don't support virtualization of the graphics hardware. This isn't a problem when you don't have to share the 3d hardware, but it is a problem when you've got a multi-tasking operating system that uses the 3d to render content for every application running. As a result, WGF has to fall back to the XP era method of composting the display when using a vendor supplied ICD that "takes" the hardware away from the OS.
In order to enable the "Aero experience", vendors have to implement a LDDM compatible driver and a version their ICD that is compatible with the WGF virtualization model.
Just like the overhead of multitasking, virtualization of the graphics hardware is not free. The supposed 50% drop in performance is due in large part to the fact that the application doing the rendering no longer has exclusive access to the hardware in question.
Given that: 1) most desktop users that "care" about OpenGL performance play games which run at full screen 2) most desktop users that use OpenGL in a windowed application aren't doing anything very taxing (meaning the cost of the OpenGL=>D3d wrapper isn't important) 3) most professional users that "care" about OpenGL performance don't care about how "pretty" the UI is and/or don't run Windows 4) all video cards with support for DirectX but no support for OpenGL will see a performance increase 5) XP era vendor supplied ICDs still work 6) Vendors can supply ICDs that support virtualization
I don't think that anything unreasonable is occuring at all.
The biggest complaint you're going to see is graphics whores complaining that their framerates go down because they don't have exclusive access to the 3d hardware any longer, and conversely, that they don't get the benefits of virtualization when they do have exclusive access to the hardware.
There are only 2-3 snippets of "real" information in the article, and everything in it was revealed during E3. The rest is baseless speculation and theorizing with no applicable first hand knowledge. Nostradomas would probably have done a better job writing an article like this...
Something which requires you to execute a script on the computer is not a virus. Think if you execute a bash script in Linux and it goes on and put itself in all your bash scripts, would you call it a virus?
No, I'd call it a trojan that infects my system with a virus.
In addition to the wrongness noted by the other reply, Microsoft stock trading in the the $115 range at the beginning of 2000. If you compensate for splits and dividends, that would be roughly the equivelent of ~$40/share today.
The last year or two the stock has cycled between $24-$31/share. The fluctation is cyclic and fairly predictable; I wouldn't be terribly suprised to see a run up to around $28-29 over the next month or two (investor speculation about another large dividend). Unless they significantly outperform analyst expectations (not likely), I would expect the stock to fall to about $26 by the end of the year, then back down to $24ish by late spring.
I have trouble imagining the stock staying under $24/share for long periods of time so long as they continue to meet expectations. I also have trouble imagining the stock exceeding $30/share with their current performance. But, as it is I've produced about a 15% return the last few years trading them...
The current *VERSION* has been out for the last few months. The older versions have existed for quite a bit longer; the concept hasn't changed radically from the first iteration.
Err, yeah... Microsoft has gotten so good at ripping off other people's products, they are now capable of doing so years before their competition even thinks of creating one.
Slashdot Mirror
Hey, if you can't beat 'em, litigate 'em to death, I guess...
Yeah, because that lawsuit Microsoft file against Apple is utter crap. Oh, wait. Microsoft didn't file a lawsuit. I guess that's why people bitch and moan when you use the abbreviation M$.
And according to the article, the paper Microsoft's patent was based off of was published in the "2001-2002" timeframe. Given that the iPod was released in "late 2001" by your words, who was first cannot be determined with the readily available information.
Apparently the FUD belongs in your mouth.
If that were the case, the exploit would have been released hours after the patch, not days.
People still own VCRs?!?
LMMH is a typo which should have been LDDM. :) LDDM is short for Longhorn Display Driver Model. Basically, Vista needed new functionality not exposed in old drivers in order to make their new compositing system work.
The same "problem" happens (aero glass isn't enabled) when you don't use an LDDM display driver for your graphics hardware. In other words, if you use an XP DX driver on Vista the system doesn't enable aero glass either. That's why you see some many posts in forums for people asking about LDDM drivers.
The claim was that hardware-accelerated OpenGL in desktop windows was impossible unless Microsoft changes things.
Which is a bald faced lie. Not only can an OpenGL ICD be written which properly "shares" the hardware, XP era OpenGL ICDs still work. Legacy ICDs just force Vista to fallback to the XP method of drawing the display. If all you do with OpenGL is play full-screen games, you won't even notice or care.
Now, I won't pretend that writing such an ICD is "easy", because it probably isn't. But I don't see anything that makes it impossible.
Hmmm, that second link doesn't appear to be as useful as I had hoped. :(
As for the output of hardware being consistent, that doesn't require Microsoft to write all the memory management code themselves, thus hamstringing vendors [...] AFAIK today DirectX doesn't have any real advantage over OpenGL in the consistency department despite the common Microsoft code DirectX
...
... though I'm kind of curious so I tried to find a few:c ee2-0b64-41f2-893d-a6f2295b40c8/TW04079_WINHEC2004 .ppt#350,1,Windows Graphics Foundation (code snippet at the end)a milyID=23a22468-5807-4ff7-a363-ce6fe69b8f04&displa ylang=en (the WinFX SDK, which has docs -- all the specifics you could want :))
No, it doesn't at all. In fact, I think that will be enforced more through the driver certification process more than anything else
Do you have any other links that go into more specifics on that?
I didn't
* http://download.microsoft.com/download/1/8/f/18f8
* http://www.microsoft.com/downloads/details.aspx?F
It's not a problem today when you run 2 or 10 or however many windowed OpenGL apps at once along with some DirectX ones too; somehow Windows seems to cope just fine
3 ,00.asp
Read this: http://www.extremetech.com/article2/0,1558,179168
It explains the nuances in much better detail than I could.
But I fail to see why they would be making such a big stink in that case.
They aren't. A guy posting to an OpenGL forum is, telling everyone to complain about the situation to the vendors. And there is apparently a guy from 3dlabs posting essentially what I've been saying.
I'll bet these newer driver APIs give vendors much less control over the internals of various parts of the graphics system.
It depends on your perspective. The final output from various pieces of hardware should be more consistent, and I don't think that is an unreasonable goal. What process they use to generate that output is up to the vendor (that's why there is a driver layer to abstract the hardware from the software), so long as they implement the minimum featureset necessary for the checkbox they're aiming for. The biggest developments these days appear to be related to vertex and pixel shaders (and MS seems to be making a big deal about going crazy with them for aero), and that's probably where most of the biggest changes lie.
They're fearmongering and spreading FUD, and obviously don't understand how OpenGL is supported on Windows.
1) The DX wrapper does not add a 50% overhead, unless the calls you're making amount to Noops. Performance with virtualization suffers. The cost of "sharing" the hardware between multiple pieces of software is where the primary perf hit lies. OpenGL isn't the only entity which will suffer in this case. Basically they're complaining that they don't get the perf you'd get if only one app was allowed to use the hardware at a time.
2) Microsoft's implementation will be a vanilla version of OpenGL that conforms to v1.4 of the OpenGL standard. Nothing prevents third parties (including hardware vendors) from implementing their own OpenGL ICD. To put it into perspective, Microsoft's OpenGL implementation on WinXP is a software renderer that conforms to v1.1 of the spec.
3) See point 2.
Re: Last paragraph. That is 100% correct. Vendors are free to implement their own OpenGL ICD that supports the LDDM. However, if vendors do NOT implement an ICD that supports the longhorn device driver model (ie: you use the WinXP model, which does NOT support virtualization of the graphics hardware), the OS reverts to the WinXP model of rendering the display.
Spoken like someone who clearly doesn't know how anything actually works, and someone who clearly likes to point fingers and assign blame.
If the OpenGL implementation you're using on a windows machine is hardware accelerated, you aren't using Microsoft's implemenatation of OpenGL; you are using your hardware vendor's implemenation.
Your reasoning for windows being "broken" is equally absurd. You might as well say all Fords are defective because they don't use Chevy engines. I bet you stay awake at night complaining about how Panasonic VCRs are broken because they don't use your beta tapes.
LDDM = Longhorn Display Driver Model (http://www.microsoft.com/whdc/device/display/PCIe _graphics.mspx)
Wrong.
If you use an LDDM compatible OpenGL ICD in Vista, the system doesn't have to fall back into compatiblity mode and none of the features are turned off.
If you use a WinXP OpenGL ICD in Vista (ie: a driver that doesn't have the goo to support virtualization of the graphics hardware), the system falls back to "compatibility mode" and turns off the stuff that requires virtualization of the hardware.
Expecting it to work differently is like expecting an s3-virge card to play quake3 at 60fps.
That is partially correct.
Vista's graphics subsystem virtualizes the graphics hardware when running in "Aero" mode (ie: the mode with all of the eye candy); think of this as a multitasking scheme for your 3d card. This mode can't be used with a library which requires exclusive use of the graphics hardware, as you correctly surmised above.
However, when you run an application that runs an XP era OpenGL ICD, the system falls back to the XP-style compositing scheme (ie: eye candy gone). In other words, vender supplied OpenGL ICDs still work.
Furthermore, venders can create LMMH compatible OpenGL drivers, which would contain the right "goo" to share the graphics hardware with the OS.
There is nothing "lcoking" in the OpenGL implementation that can be used on a Windows machine at a pure 1.4 level. Nor is this some conspiracy to make OpenGL "slow" (the previous system supplied implementation was a SOFTWARE renderer only; the only way it could be slower would be if the video card in use has no 3d hardware at all -- in which case you were already screwed).
OpenGL on windows is not broken.
1) XP era vendor provided OpenGL ICD's force Windows to fall back into a compositing mode compatible with non LMMD drivers. THEY STILL WORK.
2) Vendors can produce LMMD OpenGL ICDs that work with the new composting model in Vista.
Stop trolling.
OSX doesn't virtualize the 3d hardware. End of story.
You're not mistaken.
/. MS bashing FUD.
The big brooha here is that the "Aero experience" requires support for virtualizing the graphics hardware, which vendor supplied OpenGL implemenations don't support. This requires Vista to fall back to the XP-era style of compositing the display when running an application using an XP era ICD. If vendors implemented an OpenGL driver that supported virtualization we wouldn't even be having this discussion.
Of course, you won't find any rational discussion here pointing that out...just the normal
XP era vendor supplied OpenGL implementations don't support virtualization of the graphics hardware. This isn't a problem when you don't have to share the 3d hardware, but it is a problem when you've got a multi-tasking operating system that uses the 3d to render content for every application running. As a result, WGF has to fall back to the XP era method of composting the display when using a vendor supplied ICD that "takes" the hardware away from the OS.
f e47-dfc3-4e74-92a3-088782200fe7/TWPR05007_WinHEC05 .ppt#273,16,OpenGL
In order to enable the "Aero experience", vendors have to implement a LDDM compatible driver and a version their ICD that is compatible with the WGF virtualization model.
http://download.microsoft.com/download/9/8/f/98f3
Just like the overhead of multitasking, virtualization of the graphics hardware is not free. The supposed 50% drop in performance is due in large part to the fact that the application doing the rendering no longer has exclusive access to the hardware in question.
Given that:
1) most desktop users that "care" about OpenGL performance play games which run at full screen
2) most desktop users that use OpenGL in a windowed application aren't doing anything very taxing (meaning the cost of the OpenGL=>D3d wrapper isn't important)
3) most professional users that "care" about OpenGL performance don't care about how "pretty" the UI is and/or don't run Windows
4) all video cards with support for DirectX but no support for OpenGL will see a performance increase
5) XP era vendor supplied ICDs still work
6) Vendors can supply ICDs that support virtualization
I don't think that anything unreasonable is occuring at all.
The biggest complaint you're going to see is graphics whores complaining that their framerates go down because they don't have exclusive access to the 3d hardware any longer, and conversely, that they don't get the benefits of virtualization when they do have exclusive access to the hardware.
There are only 2-3 snippets of "real" information in the article, and everything in it was revealed during E3. The rest is baseless speculation and theorizing with no applicable first hand knowledge. Nostradomas would probably have done a better job writing an article like this ...
Only against fruit.
Its simple if you use Sony fanboy math. :p
Something which requires you to execute a script on the computer is not a virus. Think if you execute a bash script in Linux and it goes on and put itself in all your bash scripts, would you call it a virus?
No, I'd call it a trojan that infects my system with a virus.
In addition to the wrongness noted by the other reply, Microsoft stock trading in the the $115 range at the beginning of 2000. If you compensate for splits and dividends, that would be roughly the equivelent of ~$40/share today.
...
The last year or two the stock has cycled between $24-$31/share. The fluctation is cyclic and fairly predictable; I wouldn't be terribly suprised to see a run up to around $28-29 over the next month or two (investor speculation about another large dividend). Unless they significantly outperform analyst expectations (not likely), I would expect the stock to fall to about $26 by the end of the year, then back down to $24ish by late spring.
I have trouble imagining the stock staying under $24/share for long periods of time so long as they continue to meet expectations. I also have trouble imagining the stock exceeding $30/share with their current performance. But, as it is I've produced about a 15% return the last few years trading them
And yes, I did say years when I meant months. Start.com has been going for roughly 6 months if I recall correctly.
The current *VERSION* has been out for the last few months. The older versions have existed for quite a bit longer; the concept hasn't changed radically from the first iteration.
Err, yeah ... Microsoft has gotten so good at ripping off other people's products, they are now capable of doing so years before their competition even thinks of creating one.
(fyi...this page has existed for several years)