Metisse (or métisse) means 'mixture' in French.

It's also the name of a cool Irish-French musical duo [metissemusic.com]

is how this is going to make me more productive. I can barely read the text when the windows are put into those weird angles.

tell me why I would want to look at my document while it's twisted sideways?

ask yourself, "What problem does this solve?"

Where are the 3D Apps? Why 3D desktop always have to be like 2D desktop, except for some 3D change..

Its not a evolution..

I'll stick with the command line.. ;)

Because I know that's the first thing I clicked on, and it was slow then. Here's the mirror [jhu.edu].

If you have spare CPU/GPU cycles...

Right now these folks would prefer a spare webserver and some bandwidth. :-)

I guess this is pretty technically cool, but as a user of a desktop system (aren't we all?) I'm not actually sure of how this would benefit me. Would I, for example, be more efficient in my job using this?

I think there are other assumptions that need to be challenged prior to this sort of thing being built. Namely, that "applications" are the best way to segment functionality within an OS. This sort of system really seems to address the problem of moving between windows to access and work with different information from different applications. I think the problem of having to move between applications ought to be addressed first.

Finally, is anybody aware of any studies of this type of interface that prove me dead wrong? That prove that people are fantastically more efficient using a mechanism like this?

has OpenGL support, transparency and all other whistles

I'm all for whistles, don't get me wrong [severinghaus.org], but without the bells, I'm just not convinced.

3D computing environments won't be quite useful until we get a 3d input. A mouse is meant to move around a 2d desktop, not a 3d environment.

We need a 3d input device, perhaps like the ones used in Minority report? That's how I see 3d displays becoming useful.

I'm running kde 3.2.2 on a p4 1.8 ghz with 512 mb of ram and it's sluggish compared to windows xp. I don't think time should be spent trying to make cool looking 3d wm's but trying to improve xfree (alright, now xorg) or kde.

I mean I think what they're doing is cool in terms of "hey let's try this" but I don't see this as where window managers are headed. People still want fast and colorful icons and a nice file browser that's well integrated with apps. If you're unsure what I meant, compare the "save as" dialogs in mozilla and kedit.

Sorry for the rant, I was just a little annoyed I had to reboot after having an uptime of only 8 days.

this 3d stuff is all lame without some video-integration and image registration to make is seemless check out some of my thesis screencaps http://roscohill.com/skool/index.html

Famous last words:

"Screenshots here."

1. scaled windows - it's one thing to resize your windows and tile them. That's very old news. Scaled windows are another beast. Scale your firefox window everything shrinks, you don't get a bunch of "A..." "B..." tabs. Instead you get "Apples" "Boxes" etc.. in what amounts to a smaller font. Not always better or worse than resizing, but a nice new tool.

2. Skewed windows - Yeap, I can't read em' either. What is the point? It _may_ be easier to browse multiple windows and forefront the one you want using skewed/rotated effects (instead of an alt-tab ring or taskbar).

3. Window peeling - this is kinda nifty. Instead of minimizing, resizing or moving your current window to see what is underneath you 'peel back' part of the parent.

Earthshattering breakthrough in UI? Nope. A reliable and consistant cut-n-paste would be of more immediate value. But as an experiment into improving the GUI it is fun stuff.

Keep in mind that 3d desktops have the ability to increase system performance because of rendering into pixmaps instead of rendering into the framebuffer...

Why does rendering into pixmaps possibly increase performance?

If you're rendering into a pixmap, having something occlude it onscreen (i.e. in the framebuffer) will -not- be a destructive operation, and you won't have to repaint..

In otherwords, sliding windows across the screen, animating some huge mouse cursor (larger than HW mouse accel would allow, for some strange and uncouth reason), or otherwise putting stuff up in front of windows would not cause them to redraw because their pixels would not be damaged by the operation.

This is good.

There is a negative-- You use more memory on your graphics card/AGP memory, but even this can be alleviated by switching what windows/buffers you render into offscreen pixmaps.

One could, for instance, render all but one window into the framebuffer, save the colorbuffer and depth buffer, then render things in the 'active' window into an offscreen pixmap, and render that into the scene. This would require less memory than a full-off every-window-gets-its-own-pixmap approach, and would still likely perform better than our window managers today (only one repaint of windows is needed when you switch contexts, as opposed to one every frame with the current method)

I'm really not trying to flame or anything, but it always seems to me that while open source geeks have great technical skill, they completely lack any sense of art.

This window system is cool. It's cool in the same way that Aero Glass will be cool and how the Java3D desktop is cool. But what really turns me off about those screen shots is that horrible window manager. It's like whoever designed it has absolutely no sense of aesthetics.

Here's the thing... if you want a minimalist system, then fvwm2 is great. It's not a really attractive look, but it's small and fast. But if you're going to require a lot of horsepower so that you can rotate windows in 3-space and all the other cool stuff, then it's not asking much to want a window manager with some textures and lighting and curves and some other stuff that looks halfway attractive.

</rant>

"I know this, this is UNIX!"

Well, some parts of this are nice. I have reservations about the actual 3D parts, but window scaling would come in handy.

Having another X server to mediate this stuff isn't very clean though; I understand that they went that way for early development, since there isn't really anything finished that would be better, and they apparently wanted to get to the effects stuff. On the long run, however, it seems that this stuff should be done by a Compositing Manager. Of course, this also requires that the X Composite Extension [freedesktop.org] be implemented in mainstream X servers (read: X.org [x.org]).

This kind of thing only facilitates using more than one window at a time, the only purpose of which is unauthorized copying. Let's shut down this piracy tool before our economy is destroyed.

Back in the early 80's, color CRT's started to become available. People's reaction was remarkably similar to current reaction to 3-D desktops. Some people thought is was pretty, and that was enough. Lots of people wondered what good it was, and whether expending more than one bit per pixel was really a good use of memory. Would X become bloated? Would bit-blit still work? Some programmers who liked black and white better because they found it easier to read.

Why throw a ton of people into quickly improving an emerging new technology when we can split all those people up into smaller teams to try & develop the same thing from several different angles.

Thus rather than continually improving upon potential "killer apps" (not that I'm saying Looking Glass is such an app), we can all slowly develop slightly different versions of the same thing, all the while ensuring that cross-compatability doesn't exist.

Oh! And don't forget the reunion party in 2 years when we all get drunk and lament the fact that products from the likes of Microsoft stole the fire that should have been ours. Even though our solution was technically superior to Microsofts.

Ok... Maybe I do sound a bit jaded, but it sure does seem that as soon as a killer new technology or application comes on the market, we suddenly have a ton of applications being produced trying to replicate the performance of that technology, rather than either building upon the strengths of it, or developing a totally different, non-copycat alternative. Wassup with that?

...End Rant...

as far as I'm concerned, what practical use does a 3D desktop environment give me? If anything, i'll get the windows so twisted that I won't be able to use them anymore.

first let me get a program that operates in complete 3D (like a 3D website that needs a browser that can display in 3D) then once windows appear as 3D boxes instead of planes, i'll use a 3D window manager.

How about compatibility with other wms like KDE and Gnome? Of course this thing is a nice piece of eye-candy but I'm sure there are many people set in their ways and comfortable with their familiar manager.

I would like to have some of these features, however I can't say I'm ready to just start using another wm cuz of some novelty.

Why not take some of the existing technology from this wm and incorporate it into other wms?

Just a thaught....

Dancing poodles are interesting, not because they dance well, but because they dance at all. These 3D desktops are interesting, but only as a sideshow. The desktop/document/filecabinet paradigm is tired not because it is inadequately modelled on out computers, but because no one wants to work that way anymore.

X, Aqua, Windows: they're all sufficiently 3D, with stacked windows. If anything, they could use less dimensionality, and simplify the inherent complexity: window groups, accessible process data/logic/presentation tiering with pluggable dataflow, pattern copy/paste, OS-level replication and triggers, omnimedia messaging by reference, flowchart programming.

Making a 3D "desktop" is worse than just bogging down in a 20th Century soulkilling paradigm. It's distracting us from using the 3D parallel processors to separate presentation from logic and data, to obtain all those other features which reflect the way modern people intercommunicate at work, play and everything in between. Hopefully the open source of these new arrivals will allow the best functions to be salvaged and dragged into a new paradigm created by a visionary person or group. Then these machines might start becoming less of their own problems to solve, and begin to disappear in the magic of a real solution.

I can't believe 100+ comments and nobody mentions this.

I think we all agree that VR interaction as presented in the early 90s was a complete crock. It is not more efficient to have to "walk" into a "room" to find stuff on your computer. And the necessary hardware at that stage was rudimentary, slow and bulky.

Forget all that other crap. Put on your VR goggles and run a system like Metisse. Run all your existing applications in windows as normal, only now you can put them anywhere in 3D space around your workstation. Have dozens of windows open at once, all easily accessible, without desktop switching.

Long webpages could be opened to full height, sticking up through the ceiling and down through the floor. Instead of scrolling, move the whole window so the area of interest is closest. Pick out interesting sections (images or whatever) in the distance before you've "scrolled" there.

I can think of endless ways a 3D window manager could be used in conjunction with VR technology, even without any specialised applications. If I could seriously set one up now, I would. I'd probably still use my CRT as a second display device (after all, I might need to show something to other people).

Realistically speaking, this isn't practical without true see-through displays. (I want to be able to see the rest of my environment behind the windows -- such as the keyboard -- and the current displays of this type, to my knowledge, use camera passthroughs which are probably a bit laggy and nauseating to use.) But I want one, as soon as the tech catches up -- assuming it hasn't already...

And it's cool. A little slow (OK, so I ran it on top of KDE) but I think it has some cool features, the setup I really liked was: - surface like a sphere or cylinder - window auto scaling to 40% - auto flat (the window you focus on becomes flat instead of staying all distorted). That window-peeling stuff is very nice, although I don't know how useful it can be, unless it gives you enough time to bring the window in the back to the front before becoming flat again. This stuff can be really useful for having many windows on the same desktop (could be a little better than having several virtual desktops). I think this could at least be used as the foundation for an Exposé-like feature in other window managers. Of course someone is bound to say that OSS should innovate and not just imitate, but hey, Exposé is really cool, someone could add a similar feature (though not exactly the same) to KDE, GNOME or both...

Some of the features in Looking Glass seem kinda pointless, like that 3d cd jukebox idea. The organization of windows by tilting them out of the way looks like a good way to waste screen space.

Metisse seems to take a slightly different approach. The features outlined in the screenshots actually look like they make an effort to help users manage windows. I do agree somewhat with Steve Jobs' position that users shouldn't have to be janitors; a system that makes some useful decisions for the user seems worthwhile.

It looks like Metisee preserves window locations when using the 'shaped' screen and scales contents down, allowing users to utilize spatial relationships and visual cues to find data.

I also like the idea of folding over window corners to see the lower windows. Seems more useful then Alt-Tabbing. This feature is implemented from a research paper. I've seen several posts here that lament open source's propensity to copy rather than innovate. Here is a concrete counter-example of an open source project trying something new and leveraging academic research. Even if the project is a complete failure, it should be applauded for taking a risk and implementing something different.

As an aside, I remember some comments on the original Looking Glass article that critized flipping widgets over to change settings and view properties. The new Tiger Dashboard in Mac OS X is now doing the same thing. The reviews seem positive on that score. I can't really decide what I make of that. Property sheets sometimes obscure their apps, so I guess flipping is a wash, but I think I would prefer to see the application and its properties.

The second screen shot [insitu.lri.fr] about the "Auto Scale mode" looks like a feature that works better than Expose on OS X. The thing I don't like about Expose is having to work with function keys or desktop hot corners in order to activate it. The Auto Scale mode looks like a version of Expose that runs while you are working, without having to resort to function keys or hot spots.

It looks like a feature in which the shrunken windows are visible around the normal size active window, without any overlapping. I presume switching between windows requires you to simply click on a shrunken window, which would resize it to normal and shrink the previously acitve one, kind of similar to Expose.

This also eliminates the need for a Task Bar, and would also have the advantage of actively showing the windows contents, rather than just representing a window with a Task Bar button. OS X can display a minimised window's contents in the dock, but this Auto Scale mode can display minimised window contents in a larger fashion, depending upon the space available on the screen.

I also presumed that Expose was OS X's answer to the Task Bar, because their dock didn't allow you to switch windows as efficiently as the Task Bar. This Auto Scale feature looks like something that is a combination of Expose and the Task Bar, but works better than both. I think this is a really innovative concept for window managers.

The screenshots for Metisse suck and you can only use FVWM - what gives? It is a great demonstration of how productive Java is though.

Looking Glass: Looks awesome
Metisse: Looks like crap

Looking Glass: one guy in his spare time
Metisse: "a lot, see the source" (really one dude hacking other sources though)

Looking Glass: from scratch because of Java APIs
Metisse: hacked X server, hacked FVWM, hacked vnc.

Looking Glass: very secure
Metisse: insecure (it's in C and it's hacked up code)

Looking Glass: easy to write plug-ins, dynamically load
Metisse: hack fvwm in C, recompile
...

There was an article a while back saying that the language doesn't matter for security because it is bad programming that is responsible. Even without looking at the source I can guarentee there's no buffer overflows, double-free's, format string exploits, etc in Looking Glass. And I would bet my life savings there's at least several in Metisse.

There was an article recently about Java performance where most posters insisted it's still slow and jerky, but the movie of Looking Glass [sun.com] sure looks good to me. It's sad that people still use C/C++ to create lame hacks like this Metisse when there are such better alternatives. Can you imagine if the whole OS was written in a modern language? [slashdot.org]

Just because X is rendered with the help of OpenGL, it does not mean that the Linux GUI environment is up to MacOS standard. It is the what's inside windows that count, not how things are rendered on the screen.

OpenGL or not, the Linux GUI still looks like a hack...and the screenshots showing Emacs don't help.

The vastly different look and feel, from one app to the other, is one of the main problems, but it is never gonna go away...