X.org Making Fast Progress 778
prisonernumber7 writes "X.org is showing a lot of progress! The combination of the XFixes extension, Damage extension, Composite extension and XEvIE (X Event Interception Extension)
present in X11R6.8 present user interface designers with a wide range of
here-to-fore difficult to achieve possibilities. What does this mean for the enduser? That's window shadows and window shadows within windows as well as true translucency for the OSS community. Good samples of Gnome and KDE desktops with drop shadows, and so on can be found here, here,
here,
here,
here,
translucency here, here and here,
and its use on handhelds running Linux."
I hate to say it... (Score:5, Interesting)
...but I've been waiting for translucency under X forever. It hurts me to admit it because I always thought that I didn't care about "eye-candy" but this is really cool. Why did it take so long?
Oh well...off to look into downloading, though I suppose I'll have to wait for the next version of KDE to take advantage of the new features.
Re:I hate to say it... (Score:5, Informative)
Re:I hate to say it... (Score:5, Informative)
Meanwhile... (Score:3, Interesting)
There are WAY more fundamental issues that need to be addressed for widespread Linux desktop adoption, from APIs to core architecture changes. But hey, at least our cramped KDE menu has translucency now.
Re:Meanwhile... (Score:5, Insightful)
You make good points, but if there is one thing I've learned observing which software gets adopted and which does not, its that polish matters.
As the Linux desktop experience matures, acceptance will eventually hinge on its polish - OS X really has that aspect nailed down, and its not a bad goal to pursue. Sure, there are other issues that need to be addressed, and of course power-users will turn it off, but for folks like me that spend a lot of time in IDLE or Kate writing python code, I can spare the CPU cycles on my Athlon 64 3000+ for transparency. And hey, I like the look. =)
This is a good thing, and I don't think it materially adds "bloat" though I'm not knowledgable enough of the code to say that with certainty.
Re:Meanwhile... (Score:4, Insightful)
Exactly. Just like in the automobile world, polish matters, and can often make or break a deal. A lot of cars have "unneccesary bloat" (weight) which affects performance. But things like sound insulation, vibration reduction measures, and bells and whistles add to the overall "feel" of the car. Many manufacturers have admitted to paying particular attention to the sound made when closing the car door. They deliberately fine-tune the acoustics so that closing the door makes a nice, reassuring "thud", and deliberately fine-tune the springs and hinges on the door to make it feel "heavy" and "solid", when in fact, they have not actually changed the door structure itself. Look and feel plays an important role (consciously or subconsciously) in people's buying decisions.
Re:Meanwhile... (Score:3, Informative)
Re:Meanwhile... (Score:3, Interesting)
The shadows behind applications have actual utility. I noticed this when looking at the projection of one of my professor's OSX laptop desktops. It's much, much easier to discern the z-order of the applications, and which pixels be
What's with this polish thing anyway? (Score:3, Funny)
Could you please stop all this fuzz about polish. Call me old fashioned, but in my PC I want nothing else than good, old english.
Ah, those arrogant immigrant geeks.
Re:Meanwhile... (Score:3, Funny)
Re:Meanwhile... (Score:5, Insightful)
(Has Windows really had Translucencies and shadows since 1999?)
No, we're excited that there is a version of X-window that is progressing.
These latest enhancements aren't super exciting, but X.org has had a lot of enhancements added since it split off from XFree86 a short 9 months ago, and there are many more enhancements coming in the next few months.
I think that all major distros had adopted X.org over XFree86.
Re:Meanwhile... (Score:5, Interesting)
Yes, Windows 2000 had full translucencies support. How do I know? I wrote TransparentFX for Winamp [draconpern.com]
Re:Meanwhile... (Score:3, Insightful)
Re:Meanwhile... (Score:4, Insightful)
Right now, they're being used to do drop shadows and transparency - but the same extensions can also be used to reduce network bandwidth required for remote sessions or create a 3D desktop.
Composite, Damage, and the other new extensions are exactly the fundamental changes you claim are needed - the fact that the quickest way to show them off gives us some neat eye candy is just icing on the cake.
Also, Windows does not currently support this type of window compositing - it has basic alpha channel support, but there's a lot things these new extensions can do that Windows cannot, and won't until Avalon.
Re:I hate to say it... (Score:5, Interesting)
CB
Re:I hate to say it... (Score:5, Interesting)
Re:Never saw a point for it (Score:5, Informative)
Translucency is beneficial for notification, non-critical alerts can be shown without completely hiding the workspace under them.
Less window manipulation (Score:5, Interesting)
Translucency means you can (for example) have an editor window open on top of a reference web site in Mozilla, and still read the reference information while working in the editor without having to repeatedly raise and lower the two windows. As a developer, anything that lets me focus more on what I'm actually doing and less on messing around with the interface is more than welcome.
Re:Less window manipulation (Score:5, Funny)
But now you can't convince your boss to spend this year's equipment budget on a bigger monitor for you -- say goodbye to that 30" 2500x1600 LCD monitor from Apple.
Re:Not allowed in my shop (Score:4, Insightful)
Eye-candy that has no effect on the OS (I'm not saying let the users go out and install Stardock or whatever utility of the month) has no effect on administration, and as a result should be allowed.
Case in point - desktop backgrounds. There is no reason not to let users set their own, and many reasons to let them do so, like the fact that if you let them do the things they want that don't affect things, they're more likely to listen to you when you say "You must use Firefox for security reasons."
Re:Enforcer (Score:5, Insightful)
Installing programs unchecked? Not usually a necessary freedom.
Modifying inherent eye candy properties? Not a necessary freedom, but one unlikely to affect the stability of a machine, and as such a freedom that should be allowed.
Re:Enforcer (Score:4, Insightful)
Given that many secretaries (in fact, *all* of ours - we have a security guard at our reception desk, and he doesn't get a computer) are not in the public view, and that of those who are in public view their machines are often not at all visible, why shouldn't they set hunk of the month as their background? What does it matter? If someone is offended by a bit of onscreen beefcake, too bad. As long as the company standards aren't being broken (which would prevent, e.g., nudie pics as backgrounds on engineer's machines in most cases), where's the harm? It's entirely possible that my wearing of a Star of David offends Muslims in my workplace, or that my visible facial piercing offends the strait-laced. However, neither of these interfere with *my* work and with the work of reasonably tolerant people. Similarly, beefcake on the desktop interferes with no ones' work except the overly sensitive.
Admins need to learn that users are probably more sensible in many areas (read: what is and is not an acceptable desktop background) than they are, and that the areas they should be locking are those where the admin actually does know more (read: lockdown installation privileges, lockdown inappropriate network use, enforce virus protection, etc.)
Re:Enforcer (Score:4, Insightful)
X.Org proof of Open Source Advantages (Score:5, Insightful)
However, because it is open source, the project could be forked under new, competent leadership. And also, because of its licensing terms, people could switch to that fork without any negative repercussions.
Look at the progress X.Org has made in such a short period of time! How can anyone say that Open Source software is not superior?
Re:X.Org proof of Open Source Advantages (Score:5, Insightful)
Code's reusability (Score:5, Insightful)
And their work would probably be lost, and any new project that wanted continue their type of work would have to write everything all over again just to reach the existing level of functionality - which is a waste of time and effort. Instead, the pre-existing project is forked. Open from closed source is an innovation in distribution equivalent to modular/OOP from procedural in development in allowing and encouraging reusability. Reusability then facilitates easier extension - like the sort of improvements we're seeing with X.org.
Re:Code's reusability (Score:5, Interesting)
This happened to the company that I used to work for. We went out of business, the employees were laid off, and the IP was sold.
Guess where the source code, the hardware design, and the existing prototype is now?
It's in a closet at the company that bought it. It was bought simply to keep competitors from using it.
Re:X.Org proof of Open Source Advantages (Score:5, Insightful)
Re:X.Org proof of Open Source Advantages (Score:4, Insightful)
because it is Open Source software another set of developers could step up to the plate, start there Windowing project and be only 5 years behind the competition a huge head start and a real money saver for the new developers.
This story is an example of the dramatic progress that has been made in the last year under this new leadership and shows what we could have been doing if the previous leadership had there act together. If Xfree86 had been a closed source project none of this new progress could be possible because, like the berlin project, we would still be waiting for the core to be finished.
This is what shows the superiority of Open Source Software
Re:X.Org proof of Open Source Advantages (Score:5, Informative)
Why are they so hard to implement, and why does WMP have to implement a horrendous hack to get into the taskbar.
Nitpick: WMP on the taskbar is implemented as a deskband, which has a fully exposed and public interface. This means that other people could write their own deskbands if they wanted, which I don't consider a hack. MSDN has an article with some details [microsoft.com].
Re:X.Org proof of Open Source Advantages (Score:5, Informative)
Start a print job. Depending on the print driver you're using, note the printed pages count update in the dock icon.
Launch Adium. Set the preferences to display buddy status in the Adium icon.
Drag and drop onto dock icons. Watch applications launch or files get saved, etc.
What exactly do you mean by not interactive?
A better question. Why are you even commenting on GUI features without even seeing what the new versions of OS X provide. No one would take you seriously if you criticized the Windows GUI using examples from Win 3.1 why should anyone take you seriously here?
Re:X.Org proof of Open Source Advantages (Score:3, Informative)
The doc icons are not swallowed apps, that is correct.
But they are another canvas that the app can write on.
rather than just displaying images the app can treat them just like a 128x128 pixel window and draw anything it wants, the OSX doc will then display it scaled to the correct viewl.
advantages over a swallowed app, you can reuse the icon spa
Re:X.Org proof of Open Source Advantages (Score:3, Informative)
2) Microsoft's anti-aliasing in Windows 2000 is also a hack. It's not based on a compositing model and doesn't provide windows with a complete alpha channel, just a single alpha attribute. The latter, in particular, makes it most
Re:Welcome to 1999, guys. (Score:5, Insightful)
X has been around longer than Windows. And it JUST NOW has a feature that has been around for years on Windows. Why is it that OSS weenies jump up and down when a tiny feature like this finally arrives for X, when non-OSS has had this for years, and act like OSS just pulled a rabbit out of its hat?
Idiot. XFree86 has been a weak link in the chain of wonderful OSS for years, and for years we (as in the community) have been trying to get XFree86 to pick up the pace, clean up their act, and get to work. But no, XFree86 decided to linger in political pissing wars instead of actually building. It has been stagnant for a long time, and as we've celebrated all the wonderful things OSS brings us, we've all been accepting and acknowledging what XFree86 represented: a complete and utter failure that was independent of the development model. A project both unmanaged and micromanaged to the point where nothing could be done with it but barely keep pace with video cards (and even then not always managing that).
The win here is that XFree86 finally made a decision that made it necessary for the very people distributing it to stop doing so, and for a rival project to fork and fix all the mistakes.
So, yeah, you could focus only on the fact that X.org has new features that supposedly have been had for years in other parts of the industry and talk about how OSS sucks because of that. In the process you will be ignoring all the other wonderful things OSS has that proprietary software doesn't, and never will.
And you'll be ignoring the fact that the very development at which you scoff represents one of the biggest strong points of the OSS movement, and one of the strongest arguments RMS ever makes about Free Software.
So you can be ignorant, and there's still plenty of room for you.
Re:Welcome to 1999, guys. (Score:3, Insightful)
You could make Linux a Windows killer to day WITH the current XFree86 just by duplicating the Mac OS X folder structure and usability.
I've never understood this massive desire to try and copy Windows on the Linux desktop. Windows is not a *Nix OS. You're trying to hammer a square peg into a round hole.
Copy Apple for crying out loud. They've
Re:Welcome to 1999, guys. (Score:4, Insightful)
OS X uses an openGL rendering engine, but you can reproduce 99% of the GUI functionality using XFree86 as it stands now and still be 3 years ahead of Windows.
Yeah, you could do that, if you wanted. The only problem is that then the desktop would be a third as responsive as it is now. There's a reason OS X uses an openGL rendering engine rather than dumping all that load on the CPU. "Arrogant" OSS developers are well familiar with that reason, are you?
Re:Welcome to 1999, guys. (Score:5, Insightful)
OSX most definitely has hardware accelerated compositing - it's a terrific example of the GUI backend done right. It's also not available for anything but Apple hardware.
Windows does not have hardware accelerated compositing. Even it's alpha-channel support is quite flakey - have you ever seen a program with a partially transparent window? With some applications, you can set an entire window transparent, but this quite often leads to corruption of parts of the window - there's a reason Windows doesn't have built in, supported transparency settings in the display manager.
When Avalon becomes part of Windows in 2006 or 2007, it will finally meet (and possibly exceed) the features of X.Org. But I also don't expect the X developers to just sit around waiting for that to happen.
The X server features this is demonstrating aren't "a tiny feature". While hardware accelerated composition is currently being used for transparency and drop shadows, it can also be used to accelerate a 3D desktop a la looking glass - it depends on whatever the composition manager can do. It's revolutionary because this is just the tip of the iceberg.
great advances in window managers (Score:4, Funny)
*cough* Sorry.
Really great work, guys. I'm pround to see progress. But aside from these uses, what good will it do?
And what's the implication of 'true' transparency? What kind of fake transparency have we all been using up until now?
Re:great advances in window managers (Score:5, Informative)
Ever notice how if you have a transparent menu open with a xterm/etc under it and, say, you're compiling something, that the text in the transparent menu doesn't update? `Fake' transparency refers to what we've been using now, which is basically taking a screenshot of the app and then pretending that your window is transparent by using the screenshot in the background of the window/menu. This real transparency means that it's not handled by the application, it's handled by X, and since the contents under your window weren't overdrawn, it can just.. you know... render it properly.
That's what I got from it, anyway.
Re:great advances in window managers (Score:5, Informative)
True transparency means that it's truly based on alpha values and computed as the window is drawn; current ways of "faking" transparency - in Konsole or XChat, for example - essentially take a capture of the background wallpaper and use that as the background of the window. If you move the window, you can see that it takes a moment for the background to adjust; with apps that are aware of and use the new X server features, this would be done as the window moved, and would also show windows and icons behind the currently focused window.
As long as the composition manager has good hardware acceleration (something which is already the case with NVidia, but not so much with ATI), this combined with a double or triple buffered desktop could well provide a Linux/Unix desktop on par with OSX, at least technically.
Of course, it's up to the window manager to really take advantage of these added features; metacity can already support window borders with alpha values, for true transparency.
Re:great advances in window managers (Score:5, Informative)
the kinda where you don't see what's right under the transparent window, rather you just see transparency against the background image.
this is supposed to be true support for it, without any goofy hacking from within the program, i think.
Re:great advances in window managers (Score:3, Insightful)
For instance: say you are monitoring commands you are running on four different machines in four terminal windows that will take a while, and as a sanity check, you also have xosview running behind the terminal windows corresponding to the machine the terminal is on, and showing through enough to be useful but not enough to totally obscure your commands window. Now you get to use mor
Groovey (Score:5, Funny)
Re:Groovey (Score:3, Insightful)
Apple computers are so much more than just drop shadows and translucency. To get the designers, you would also need hardware as well designed as Apple hardware (read: designer hardware), you need to completely remove the need to go to the command line, you need to simplify the interface.
But most importantly, X needs proper colour matching support. Designers need to work with ICC profiles
Re:Groovey (Score:3, Informative)
Re:Groovey (Score:4, Informative)
Right click your desktop, -> Configure Desktop
Behavior -> Menu Bar at Top of Screen -> Current application's menu bar (Mac OS-style)
For more fun, add a small Kicker child panel at the top of the screen, then right-click it and pick Add -> Applet -> Menu.
Then you can have your clock, systray, or whatever else in the menubar. At the top of the screen, but managed by the application.
Full size pngs and jpgs galore! (Score:4, Funny)
Comment removed (Score:4, Funny)
Performance (Score:4, Informative)
Why GPUs Matter (Score:5, Informative)
Chances are you were running your X server with unaccelerated drivers - which offloads all the hard work to the CPU. In Panther, Quartz Extreme allows the transform and lighting engine of your GPU do all the hard work, leaving the CPU for things that a CPU should be doing it.
Properly implemented and accelerated, eye candy does not have the take away from CPU power and can greatly enhance usability - as it does with OSX.
Re:Why GPUs Matter (Score:5, Informative)
It's partly that. But Quartz is also fast on a plain unaccelerated 2D framebuffer. To prove this, simply run Panther inside Mac-On-Linux on the Linux PowerBook. Transparent windows and drop shadows are noticeably faster inside MOL than on the Linux desktop.
The issue is apparently the interaction between XAA (XFree86 Acceleration Architecture) and the XFree86 driver model. It isn't designed to handle Composite and Render properly. There is a hack in the 6.8 release so drivers will work, but suboptimally. There is considerable work going into a new driver architecture called Keith's Driver (kdrive) and XAA which will give near-Panther performance. But the powers that be have decided to leave those improvements until X.org 6.9. They want the extensions out there now, even if they're slow, so GNOME/KDE/others can start designing applications that use them.
Good, but... (Score:5, Insightful)
I really wish the default font situation would be better in the world of X and nix/bsd distros. Switching back and forth between Macs and PC's (windows), it's amazing how much better the mac fonts look and feel than windows. Likewise, Windows looks as much again better than the typical gnome/kde setup I have seen.
Re:Good, but... (Score:4, Insightful)
Re:Good, but... (Score:5, Informative)
This is a mere configuration setting that most of the distributions choose to set. Fontconfig is perfectly capable of turning off AA for a given range of sizes.
Re:Good, but... (Score:3, Insightful)
Re:Good, but... (Score:3, Insightful)
I've found exactly the opposite to be true. The Mac OS X rasterizer seems to "over anti-alias" the fonts - it makes everything too soft and makes any font below 9 points hard to read. With ClearType on, Windows fonts look sharp but not aliased.
So, for larger sizes, Mac OS X wins, but for the smaller sized fonts, Windows XP wins.
Note that the default font size on Mac OS X is la
Re:Good, but... (Score:5, Informative)
Where the Linux desktop really shines, however, is when it comes to customization. I prefer to operate in a very Windows-like manner, with maximized windows and taskbar. KDE allows me to do that (and gives me a nice launcher command bar with autocompletion - I haven't used the "start" menu in ages). Some want a nice file manger - KDE gives you Konqueror, GNOME gives you Nautilus. Others prefer doing everything in the shell, where you can use Midnight Commander and feel like you're back in the old DOS days.
Some want virtual desktops or virtual screens (larger than the physical screen size). Any decent window manager provides that. Some want a very efficient, slim system - they use something like Windowmaker [windowmaker.org] or XFCE [xfce.org]. Others want all the bells and whistles and install KDE or GNOME with lots of applets. Some like to experiment with innovative new UIs and try out window managers like ion [cs.tut.fi]. Others are happy just using a cloned Windows or Mac interface.
If you're willing to experiment, no system offers you as many possibilities as Linux. If you just want a clean, working desktop, all the major distro makers provide that by now.
There's room to improvement, and the devil is in the details: clipboard interoperability is still buggy and incomplete, performance in some areas can be improved (try resizing your window very fast with content visible), the driver situation is unsatisfactory etc. But none of the problems before us is unsolvable. It's just a matter of time.
Re:Good, but... (Score:4, Informative)
I'm going to have to call your bluff. Consider FreeBSD as one example. Same desktop. Same graphics subsystem, including DRI. And if you're into proprietary graphics drivers, NVidia's is even available.
This service announcement brought to you by the Pedantry Police...
Re:Good, but... (Score:3, Informative)
If you have a license to use Apple's patents on anti-aliased fonts, then you can rebuild FreeType's Font Engine [freetype.org] to support pretty fonts. They switched it off due to legal uncertanities [sourceforge.net], but it is still included with the source code. You have to edit an header file, determine where your Linux desktop distribution stores FreeType's libraries, remake, and install it (as root) to enable it. That sounds complicated, but it is actually really easy to do. Just follow the instructions [freetype.org]; you don't have to be a h
Dock vs. Taskbar (Score:5, Interesting)
OK, so that rant above is coming from someone who has mostly converted to using Mac OS X on the desktop. I still use Linux on servers. Anyway, I remember about a year ago when I made the transition, that the dock seemed rather confusing. However, after a couple weeks of usage, I was cursing every taskbar system I ever had to interact with (Windows and Linux). The dock is just so much more condusive to having many windows open. Add in Expose to the mix, and you are in desktop heaven.
So, my question is then, especially to the GNOME developers (GNOME is my preferred Linux DE), what are the plans with regard to application launchers such as start menus, taskbars, docks, etc. The progress has been remarkable, but, to me at least, this is the area most sorely needing standards and consistency.
It keeps getting better (Score:5, Informative)
Just think: A day in which support for the tuner on ATI cards is simply in the X server, rather than taking a great deal of pain and suffering to get working!
(Of course, this only applies to cards supported by GATOS, the older cards. But perhaps, just perhaps, if enough people bring pressure to bear upon ATI, then ATI will use the GATOS code to support the newer cards as well.)
Re:It keeps getting better (Score:5, Informative)
Right now the stumbling block is to upload DSP microcode using VIP bus FIFO. For some reason how to do this was obvious to ATI folks (as docs don't mention much of it) but very hard to accomplish in practice.
losing contrast (Score:5, Insightful)
Re:losing contrast (Score:3, Insightful)
"distance" when used to refer to window depth is, in reality, non-existant. Even if you were to code such a system you would need a way to actually place a given in active window at a fictitious depth "in" your monitor. To facilitate that false sense of
Re:losing contrast (Score:3, Interesting)
Re: (Score:3, Interesting)
Users Experiences (Score:3, Interesting)
Using the nvidia drivers with RenderAccel enabled with xcompmgr makes the desktop fly! Its amazing the psychological difference the redraw elimination makes. It certainly feels much lighter, and gives my iBook a run for its money. The transparency effects have no noticible speed hit whatsover. I've had multiple transparent videos playing, moving around, etc and its all smooth the way it should be.
This project really is an example of how re-opening the project from the XFree86 'cathedral' has increased development activity in leaps and bounds. Congratulations to all the X.org and freedesktop.org developers on a great job.
-theoddbot
Caveats on the new X facilities (Score:5, Informative)
The original poster meant well, but did not include the explanatory text with the screen shots...
Drop Shadows - choose a light source now guys (Score:4, Insightful)
Now is the time to get the esthetics worked out before things get entrenched. The screenshots of the drop shadows show a shadow around all of the edges of the windows. This looks really funny since this implies that the light source is directly over the center of each window. Why is there a shadow on the top? If we're going to have a rendered-style look we should choose a decent place for the light so we can have some consistency.
I vote for the light source to be at offscreen at the top left.
Re:Drop Shadows - choose a light source now guys (Score:5, Interesting)
Re:Drop Shadows - choose a light source now guys (Score:3, Funny)
Wouldn't it be nicer to link the lightsource to the systemclock and have it move like the sun over the desktop? So you could tell what time it is by the length and direction of the shadows. It would have to move backwards at nighttime of course or you would have a black desktop at night...
Re:Drop Shadows - choose a light source now guys (Score:3, Funny)
Good for those late-night hacks
Some smart aleck will tie it to the lunar cycle, so they can use the excuse, "Nah, I can't code that tonight, it's a new moon"
Re:Drop Shadows - choose a light source now guys (Score:4, Insightful)
Drop shadows, effects, etc are nice but... (Score:3, Insightful)
Why can't it detect when I'm docked and switch to dual head (LCD + CRT)?
The effects are cool, but alot of us would like to see these usability features too. I like using a graphical login, but I can't because I'm forced to have two X configurations. One for my LCD and one for my CRT.
I don't give a flying fuck about fancy-stuff (Score:3, Insightful)
Never understood gripes about eye-candy (Score:5, Insightful)
First, as long as you take the approach that Apple took with Aqua and Quartz in offloading the graphic work to the graphics card, then who cares? It barely affects CPU load and you get a better looking interface. It's just putting unused potential to work.
Second, why would you want to look at an ugly interface? Car makers put a lot of work into what you see when you're sitting in the driver's seat, right? Steering wheel, seats, dashboard... they've all been carefully designed for looks just like the outside of the car? Those of you complaining about UI eye-candy: do you also look for totally stripped down cars too? There is something to be said for aesthetics. Unless you're a robot, it affects you.
Third, some "eye-candy" can actually serve a purpose. For example: the "slurping" effect in OS X that so many people complain about actually acts as a visual cue, almost like a moving arrow, to show you exactly where your window is minimizing to. I never lose track of minimized windows in OS X, but I do it all the time on Windows. (Of course, it helps that OS X also has the added "eye-candy" of showing a minimized version of the window itself in the dock.)
Just a few things to consider. I don't think eye-candy is the Great Satan it's often made out to be and it's good to see X keeping pace.
Is this good or just a shitty hack? (Score:3, Interesting)
In particular I am concerned that things like transparency seem to be accomplished at the application level rather than the rendering level. In other words, at least on a quick read, it seemed that transparency was handled by the application wishing to display a transparent window asking that window to be rendered off screen, having that composited window returned and then rendering this to the X screen. It would seem a more robust solution would be to allow simple rendering of windows with an alpha component.
I know this might provoke a war over the sufficency of X but I'm hoping to get a few serious responses with technical knowledge about how reasonable it is to do these things without re-enginering X.
Re:Is this good or just a shitty hack? (Score:3, Informative)
Re:Another Step (Score:5, Insightful)
Re:Windows Winplosion (Score:4, Informative)
Re:Yay! (Score:5, Informative)
I know you're trolling, but it's rather important to recognize that X is a protocol, nothing more, and nothing less. How fast or slow it is depends on the implementation; some are clearly better than others.
Comparing X to direct video access is rather like comparing ssh to the linux console. The latter will always be faster because there is no network bottleneck, but you can't use it to connect to your machine remotely.
Furthermore, there are extensions which practically do give an application direct access to the hardware, or at least, as direct as one can safely manage. But what's the point, if all you're doing is drawing widgets? I could understand it for games or playing DVDs, where speed is important and you'd be stupid to play it over the network anyway, but for a window manager?
If you want to use a single user, non-network aware system, then by all means, use Windows*. In the meantime, we should keep working on improving our X implementation in areas like optimizing for local connections. There are lots of shortcuts that haven't been implemented and I'm sure there are lots that haven't even been thought of yet.
Anyone that has had to administer machines remotely appreciates the network transparency of X. It's the only window system that has this feature. Let's not throw it out because some trolls don't understand its value.
*Yes, I know Windows is now multi-user. But it wasn't designed that way, and it shows.
Re:Yay! (Score:4, Insightful)
Actually, that's not a good comparison, because when X runs locally it does not use the network, but instead uses shared memory. This is really fast, so the assertion that network transparency slows down X is a total myth.
What really makes X on linux slow is that there is almost no hardware acceleration (even with accelerated drivers). The RENDER framework, used for a lot of the gee-whiz graphical effects, is almost entirely non-accelerated. This is due to incompatibility between the X driver design and the RENDER framework which makes it incredibly difficult to write an accelerated implementation of it. This will get fixed when X.Org moves onto the kdrive driver framework.
What also hasn't helped historically is the fact that X runs in a separate process, and so you have to wait until the kernel wakes up the process before you see drawing occur. Older kernels were poor at recognizing when X needed to draw stuff, and so there was a noticeable delay between user action and the corresponding on-screen result. Ofcourse, if X ran in-kernel, any X crash would take down your entire system. I personally would rather have a small speed hit than have an unstable system.
Re:Yet again, zero innovation (Score:5, Insightful)
Sure. From the User perspective, looking at screenshots, it probably looks that way.
You're giving Apple too much credit. The news here is not 'Oooh.. now we can look more like OS X', the news here is 'Now we have proper support for the things OS X supports'.
There is a difference here, because what 'the things OS X supports' are, by which I mean the 2D rendering API, is not a thing developed by Apple alone. Firstly, Apple's Quartz uses the PDF rendering model, which was created by Adobe, and PDF was in turn based on PostScript.
That this is a good way to do 2D graphics is a no-brainer. Postscript was invented in the early 80's. The Mac later supported it's own kind of device-independent images (QuickDraw, and PICT files). Windows had Metafiles, and GEM (if anyone here used the DOS or Atari version) had it too.
Given the success of Postscript and PDF, it's pretty natural to support the things they do. But Adobe (creators of PS and PDF), shouldn't get all credit either. They just implemented stuff developed by others, like Porter/Duff compositing.
(Another early 80's innovation)
So basically, none of this stuff is actually new. It has simply come of age. Apple has been in the forefront, and that is tribute to them. But if you think that this is all Apple's ideas.. You are wrong.
Re:Yet again, zero innovation (Score:5, Insightful)
Re:Yet again, zero innovation (Score:5, Insightful)
Ok, accuser. List every innovative thing you've done and released under an open source license.
It's either put up or shut up, and real quick. Your criticism is hardly constructive, and while I'm willing to grant any random user the right to constructive criticism, I'm not willing to grant it to flaming criticism. So now you have to prove your credentials or shut the fuck up.
So let's have it. What have you done?
Oh brother.... (Score:5, Insightful)
You're bitching at the wrong project. It is the window manager and desktop environment devs who most directly determine the look and feel of what you see on the screen. X.Org writes the bits that expose the functionality of video hardware to application developers and various layers of the OS.
All those screenshots are meant to do is advertise the availability of certain effects and capabilities that up till now could only be achieved with dubious hacks.
As what desktops look like, they can look like anything. Out of the box, they can look like Windows, OS X, or other things entirely. Everybody has the basic elements of windows, widgets, icons, and some sort of pager to work with. As it happens, my desktop doesn't particularly resemble either MacOS or Windows. Get off it already.
Given statements like "10 years behind" coupled with general ignorance and I have to come to the conclusion that your troll-fu is extremely lacking. The low userid only makes it worse. It implies you've been around long enough to know better.
Re:Yet again, zero innovation (Score:3, Insightful)
You're looking at Gnome, a desktop environment that runs on top of an X server.
This news just means that x.org now has more capabilities that desktops can take advantage of.
As far as your comment about Linux being behind the times, I'd agree that Linux is playing 'catch up,' but I don't think that's a bad thing, just a necessary step.
Re:Yet again, zero innovation (Score:5, Interesting)
As far as your comment about Linux being behind the times, I'd agree that Linux is playing 'catch up,' but I don't think that's a bad thing, just a necessary step.
Actually, we're not ten years behind. Microsoft just got a ten year head start, and UNIX even longer.
Let's not forget how young today's open source operating system is compared to these others. Sure, the BSDs can claim direct lineage to the original UNIX, but all the fragmenting of UNIX that happened in the 80s and the lawsuits of the 90s made sure BSD would fall behind. And Linux came to the table pretty late in the game, and before Linux there was no open source X implementation. XFree86 was given to us (iirc) when Linux finally had enough POSIX implemented to run it.
So, yeah, you could say we're ten years behind if you really wanted to. That means we've closed the gap considerably, when you get right down to it. ;)
(Side note: I don't think we're ten years behind anymore, I think we've closed the gap completely. In some areas we have some truly innovative stuff, while in others we lag behind, so the aggregate of all the OSS stuff you get with an average distribution puts us on par with Windows XP and Mac OS X. The thing is, nobody will notice until we actually surpass them, and then it'll be too late for them to catch up. Also, people focus on individual features as being behind, rather than looking at the whole forest to see how thick it has grown and how much true innovation is in it, and that's a great disservice to the OSS world as a whole)
Re:That's nice, but... (Score:5, Informative)
It does this by diverting the X apps to draw to an offscreen buffer, and then compositing it to the screen in one feel swoop. This makes windows really solid. There is no need to wait for the X client to redraw the window contents when you move another window over it for example.
Re:Great Progress... (Score:4, Informative)
The shadows/transparency are only going to be rendered when changes occur to the windows they show anyway. Unfortunately there's often a lot of other things going on at the same time (expose events etc.). With the window manager I use I set the moves and resizes to only show outlines until the window is place where it is going to go.
Since it's part of the X server the effects are all going to be done locally anyway. If you only have a monochrome display on the remote box you can still run most X applications correctly anyway, this will be no different, and you'll probably never see the effects. However, if you have a program that specificly requires an extension which your local X server doesn't support (eg. like we currently have with OpenGL), then it won't work.If you have some VNC type thing that takes bitmaps then you'll probably want to turn the effects off in the X server that feeds it.
It makes a difference, switching between the VESA driver and the Nvidia driver for a card makes a large difference. X has had hardware acceleration for a very long time (pre-linux I suspect).Re:list of questions (Score:4, Informative)
- The eventual goal is to implement the X server on top of OpenGL, so transparency and 2D operations can be hardware accelerated.
Re:list of questions (Score:4, Informative)
Of course, the key piece here is Cairo/Glitz [freedesktop.org], which is already quite usable.
Re:Drop shadows are wrong (Score:5, Interesting)
If it doesn't look right, it's only because the WM/toolkit programmers didn't think about it enough. But the actual X.org extensions are very, very flexible. Don't worry; this is just a showcase of what can be done, but it's not all that can be done.
Remember, X.org is producing the X server, which is the lowest level of the X window system -- all it is is an implementation of the X11 protocol. Everything you actually see is drawn by other processes, like the Window manager, individual apps, etc, etc. The X server by itself isn't usable and provides no UI whatsoever.
Re:Drop shadows are wrong (Score:3, Insightful)
The current xcompmgr program is just a demo --- who cares what the drop-shadows look like? It can and will be replaced as window managers
Re:X.org or freedesktop.org? (Score:3, Informative)
Long answer: Kdrive and much of the Xorg work are hosted on freedesktop.org. Kdrive is on the back burner which Xorg kicks up. Much of the work in Kdrive is informing the implementation in Xorg (like the Composite extension, and the driver extensions for hardware Render acceleration)
Re:Eye candy is nice... (Score:3, Interesting)
There are quite a few legtimate reasons to do transparency:
1) It looks much nicer. In particular, non-square window edges can be nicely alpha-blended with the background, instead of standing out as they do now. Compositing allows for fully flicker-free redraw and resize, so you never have to bear ugly partially-drawn windows.
2) Transparent windows have their uses to allow i
Re:Wow. (Score:3, Funny)
You're talking to the guy who wrote an article called "Drowning in Aqua".
I think my dislike of user interface CPU-wasters is pretty generic.
Re:Wow. (Score:3, Interesting)
And modern CPU's have insane amounts of CPU-cycles at their disposal. Modern CPU's are so powerful that it's kinda ridiculous. How many CPU-cycles do you need to do your work? What do you do for living? I assume it requires massive amounts of CPU-cycles? 3D-rendering?
Most people don't do stuff like that. They run word-processors, browsers, email. you don't need uber-machine for that. Modern computer are ridiculously
Re:XFixes (Score:3, Informative)
Re:it's too bad that the fonts still look ugly (Score:3, Insightful)