Slashdot Mirror
Longhorn's Windows Graphics Foundation Examined
Matt J writes "Dave Salvator at ExtremeTech goes over some of the graphics designs for Longhorn. 'David Blythe of the DirectX development team gave a very interesting talk about the upcoming 3D graphics architecture in Longhorn, the next major revision of Windows. Called Windows Graphics Foundation (WGF), this new architecture will usher in some major changes to how 3D graphics operations get handled by Longhorn. These changes extend well beyond Longhorn's Avalon technology, which will render the Windows Desktop using a GPU's 3D graphics processing power rather than the traditional 2D blitter. WGF will instead define the core 3D operations themselves.'"
Is this like Keith's Getting X Off The Hardware plans, where he suggests that having your xserver running on top of openGL instead of having to talk to all this messy hardware stuff will make it nicer and faster?
We have cairo.. same kind of thing, and people are modifying stuff to implement it everywhere.. Theres also many other technologies to make up everything that Microsofts new one will do (the difference is though that we are much closer to getting a stable version)
http://www.freedesktop.org/Cairo/Home
Just dont take all of Microsofts noise too seriously, just be aware that by 2006, linux will have completely equivilent technologies (in many cases we already do), and just cause we dont make much noise about it, dont think that they dont exist, or aren't planned for the near future.
Honestly, the stuff which I have seen for longhorn so far hasn't been mindblowingly amazing, and are really just things where they are trying to catch up to MAC OS X, or linux
Oh, moving some stuff to user mode? Well, um, better late and half-baked than never and not at all?
Seriously, putting stuff in the kernel that should have been in user space is one of the more serious architectural botches in Windows. It has caused massive stability problems. Now it seems that Microsoft is recognizing this, and is starting to undo it. (What they need is to completely undo it, but they have to start somewhere. What they don't get to will continue to bite them until they do.)
When was the word 'innovation' used? I can't find it. But since we're on the topic, it is interesting that despite MS being a monopoly, they're still doing major work on their upcoming OS. But... no, we'd rather talk about their OS taking a big step towards (possibly even past) what Apple has done.
"Derp de derp."
Because "Open Source Development Model" skips little things like QA and user feedback testing. That allows them to "release" something and claim to be first with a particular technology, when in fact it's half-baked by commercial user standards.
That's exactly what people said about Win95, and 2k, and XP.
'Standards' in computing only impress those who are impressed by things like 'standards'.
It is time for the "Z" Windowing system standard. A fully SVG compliant, and "X" compliant vector based (but bitmap friendly, via texture mapping) system. Who's down?
Looks good for your age..
One thing that's fairly consistent about Microsoft is that they wait for the hardware to catch up to the software, and then, when it's 'ready', they take maximum advantage. That's why they can be 'late' and still catch up and dominate. (eg. GUIs and 386 machines, or web browsers and 56K modems)
Anyway, when Windows 2000 was released, there was only one graphic driver that supported the menu effects in hardware (Matrox). Which is probably why they've been fairly conservative about effects, unlike Apple who forced them onto a bunch of Rage Pro owners for marketing reasons, and then had to wait for the HW to catch up.
Business. Numbers. Money. People. Computer World.
I read the article. And many other like it. I also read the comments. And ran Longhorn 4072 for a while.
Everyone's getting excited about the compositing. Which will not be in production for ages, and doesn't do anything we've not seen before.
If you read (and fully understood) the article, you would realize that a fully hardware accelerated windowing system is not all that Microsoft promises with this new stuff.
The other stuff I see as being BIG are the changes to DirectX such as removing a lot of the fixed function pipeline features. They are pushing the GPU to be more generalized which is a good thing.
Microsoft is really hyping up Longhorn and none of the meat of Avalon has made it into the technical previews. Judging by the Ctrl+Alt+Del animations, the smooth color fades in Explorer, the few existing vector graphics, the other random programmer art in the technical previews, Avalon is going to be IMPRESSIVE.
Whether you like MS or not (which you don't, this is slashdot), they have the programming and graphical resources to pull this off in a very big way.
http://brandonbloom.name
There's really two things going here:
A) The factual question of whether Linux have something like Longhorn in the same timeframe? This depends on not only the availability of the tech, but it's adoption into major frameworks and applications. Open Issue.
B) The slashdot rhetorical battle where MS Vaporware is countered by links to Open Source Vaporware. Nobody here really doubts MS's abililty to execute, so "So-n-So wrote a paper" or "Here's a sourceforge project" looks a little thin.
Business. Numbers. Money. People. Computer World.
Does anybody notice that most of the computing industry would be redefined according to Longhorn?
Now I know they need to build something really different, but are all these differences really worth the hassle?
Maybe it's just me tired of hearing about software that won't be in use for another 3-5 years as if it's the best thing since sliced bread...
Blogging because I can...
There have been widget libraries built on top of OpenGL for years. See GLOW, for example. It's straightforward to do, and works reasonably well. Works on any OS that will run OpenGL.
Aqua is also bitmap based. Despite what many have said, OS X icons are just bitmaps, as are the buttons and other controls. That means that they don't scale very well - just like the widgets in Windows XP.
Yes, Aqua is one mega-gigantic compositing engine. The power of that shouldn't be underestimated, but I'd expect Longhorn to be able to do that fine. However, Quartz 2D is also a complete vector rasterizing engine, implemented (I assume, it'd be stupid if not) in AltiVec. Why use a GPU when you have multiple vector processors on a G5? (With oodles of L2 and L3 cache to eat on). FYI, writing vector graphics code with AltiVec is very yummy. If you look at the Quartz 2D API, there are no direct compositing functions; it's all vector-graphics. You can take pixmaps and composite them together (using the 'over' operator). Although I guess when they added support for the PDF transparent imaging model (part of PDF 1.4/OS X 10.3), they added support for transfer modes of vector graphics/pixmaps; I haven't looked into that.
As for icons, it's a heck of a lot easier to 'paint' an icon with pixels than to define a drawing with shapes and gradients. Also, Tiger is going to support 256x256 icons (!). IRIX's window manager (forgot the name) had vector icons. No biggie :P
With Longhorn, everything is vectorized. You'll be able to adjust the DPI of your display and all of the controls will automatically update to match it.
Tiger supports a resolution-independent user interface. With Cocoa based on the PDF imaging model, where every coordinate is represented with floats (including mouse position, which kicks in when you have a graphics tablet), it's very easy to scale everything (and rotate! NSView supports arbitrary rotation of views, and all further drawing in the view will be rotated as well).
It doesn't seem that the Tiger release notes are online yet... perhaps I should shut up.
Moderators should have to take a reading comprehension test.
A number of benchmark tests have proven that, CPU cycle-for-cycle, a well configured UNIX or Linux server can outperform a Windows server in just about every server task there is.
Surely the base reason for this is that the UNIX pholosophy is not to waste valuable computing resource of GUIs and graphical processes when you don't need to. In other words, have your UNIX server running in a console mode, perhaps with a web server or X-Server running, and just do all your administration of that server either through the console or via a web interface or remote GUI session.
It's ridiculous the Microsoft still haven't released a non-GUI server variant of their OSes, especially when, in my experience, a large proportion of blue screens and crashes are as a result of something going wrong in the graphics sub-system somewhere.
The only logical conclusion I can come up with is that there is a conspiracy between Microsoft & hardware manufacturers for MS to constantly waste CPU cycles to ensure that everyone is forced to upgrade their hardware constantly.
Gentoo Linux - another day, another USE flag.
The point is that 3-D card compositing is actually much faster than 2-D compositing on today's cards. The hardware is no longer super-optimized for 2-D... nobody cares about 2-D hardware anymore. The way Windows moves windows is insanely slow.
If I drag the window this browser is in on my 2.4 GHz machine with a Radeon 9800, I get tearing and it jumps around a bit. I have "display windows contents while dragging" on.
2-D on Windows is a fifteen year old setup, more or less. It's time for a new model, with less programmer complication.
If I run Quake III with vSync on, I get no tearing and I'm running at well past my monitors refresh rate... objects have apparent physical reality instead of this flitty flit windows nonsense.
Today if you turn off 32-bit colour for 8-bit color, you don't make your machine faster. It's slower because Windows stores internally as RGB. Turn to 256 color graphics and you have to set the palette all the time through BIOS calls. Yuck.
The same is true here... Windows is moving to a newer, faster graphics model. It's faster to have each application draw to its own framebuffer and let the 3-D card composite it. It's faster to not deal with actual blitting loops within applications that tie up the whole processor.
No more clipping calculations for the various windows, no more trouble with more than one video application trying to use the god-damned overlay mixer, no more trouble with overlays not working on the second monitor of a dual-monitor setup....
Don't think of it as 3-D bloat... that's inaccurate. Think of it as enabling the 3-D coprocessor that almost every computer will have at that time. You already know your P2-266 won't run Longhorn. 3-D doesn't make that more true.