3Dwm Updates

Robert Karlsson writes "3Dwm, the Three-dimensional window manager, an open source project at Chalmers Medialab, has just released a new, extensive release of 3Dwm, release 0.2.2 - VNC support, 3D scene graph, big texture splitting, client connection, framework, 3D materials support, testsuite added." miles away from a real desktop, but a great testbed for those ideas that are way ahead of their time.

Comments

Guys.... (and gals)

[BluedemonX]

The critique of the idea of a 3D windowmanager doesn't take into account it was built for a computing environment that ISN'T a desktop (the Cube) and secondly, there may very well be some applications of this we aren't yet aware of.

I agree that burying Linux, typically the darling OS of the "why do I need a GUI when I can just grep -v -x | sed -qvf | tar -qt32pir2 : smeek -frfk > kibble.rpm ?????" set, in a 3D window manager seems weird: a total waste of CPU cycles, a criminal waste of memory, and an exponentially bad version of Microsoft Bob (eye candy that actually got in the way).

But who knows? Maybe someone will come up with a way of visualising and representing data that'll only make sense this way.

Time mechanics data? Physics data? Who knows?

Just because I can't think of a reason for this yet doesn't mean one exists.

Sweet Merciful Crap...

[Isotrope]

Don't get me wrong, I'd rather this existed than not, but wtf, how is this "way ahead of its time"!?!?!?

If the only thing we can vary in our user interfaces is how many dimensions they display, its going to be a long time before we make any inroads into dealing with large amounts of data using arbitrarily generated paradigms that minimize the energy input required to acquire a particular piece of information.

DATA--that is, what computers deal with--DOES NOT require a user interface mapped to the tangible nature of our reality.

People seem obsessed with the question, "Where is it??? How is it arranged?? How is it layed out???" while the question "WHAT is it??" seems to elude them.

Sorry if this post seems angry, but this isn't ahead of its time at all. Our interfaces and computers are BEHIND the times. (assuming objective time... yeah, we're both wrong I guess)

Just my $.02, and yes, I am working on my own interface, but I am only 1 person, so it's going to be awhile.

--
Disgruntled AC

Why not use the Quake engine?

[Mr. Flibble]

John Carmack thoughtfully GPL'ed the Quake engine, why can't this be used for a WM?

Quake is a lean 3d graphics engine, and now with programmers working on it at sourceforge it can only improve. Heck, since you can run it from the command line you can completly bypass X! (Although too much software depends on X now...)

Re:Oh, but if only...

[nellardo]

I seem to remember some people having gotten a Nintendo Power Glove connected and working with a PC as an input device

The first example I remember of that was Randy Pausch's group, then at UVa, in a paper called "Virtual Reality on $10 a day," in the UIST conference. It was a cute (and effective) demonstration of how researchers were pursuing expensive toys and not actually addressing the issues. Rather than a half a mil in SGI and one-off things, Randy put together a complete, head-tracked, head-mounted, glove-input (using the Nintendo glove), 3D virtual environment, for less than $10000, in 1994. The most expensive component was the magnetic tracker for about $6000. It was monoscopic, black-and-white, and wireframe, but it did 30 HHz, using SPHIGS (the simplified educational version of PHIGS used by Foley, van Dam, Feiner, Hughes).

Unfortunately, glove interfaces (even with a more expensive and more accurate DataGlove) have two problems: fatigue and lack of haptics.

Re fatigue, how long can you mouse around on the desk for? (hint - how long can you play quake at one sitting?) Now try swinging your arm, unsupported, in the air for that amount of time - that's called calisthenics.

Re haptics, a great deal of control comes from kinesthetic feedback (why you can touch your finger tips on opposite hands together with your eyes closed). When using something like a mouse, the inability to move in certain directions gives you a great deal of information. A mouse is relatively crude at this, but a pressure-sensitive tablet is better and an actual paintbrush is better still. Now pick up a virtual paintbrush (hint - you can't feel it in your fingers) and try to paint with it (hint - you can't tell when it touches the canvas). Researchers have tried substituting sound (a noise when you run into something), and "ghosting" to show where the virtual object actually is (because of physical constraints) and where you put it (i.e., straight through the canvas).

This has, in part, I think, led to the recent research popularity of augmented reality and "virtual desktops".

The killer app

[Docrates]

IF you ask me, this could well be the killer app that would put linux and/or other opensource/free software OS' on every desktop. (you do realize we need a killer app for that right?)

Surely a lot of people will just say no! waste of cycles, waste of memory, waste of time!, but i couldn't disagree more. If i could have a 3d view of 6 different programming sessions with a real time representation of runtime results i would. it doesn't matter how much hardware i throw at my current system, i just can't do this today. same for designers, engineers, etc. this would mean a real productivity boost.

which brings me to my point... with the current pointer/icon/double-click/window interface, there isn't much to gain on the productivity side even if you throw in 5 years of moore's law. we're stuck. it's the same thing that happened when the GUI first came along, everyone said it was a waste of resources, but now that we have the hardware we're glad they worked on it nonetheless.

of course we need a good navigation system an efficient way to handle objects, but both exist today: i'm pretty efficient in moving around in quake and homeworld, and i can move objects in 3d pretty easily using truspace or Max...

actually i think we're late in doing this. the hardware to support a totally 3d desktop is here (GTS Ultra, 1GHz CPU, 512MB Ram would probably do the job). it's just a few years before similar hardware is standard configuration for most. If by then linux (or similar) are the only platform that will give you an intuitive 3d environment, lots of people would get it.

Clearing up some points

[NickElm]

I'm one of the core developers of 3Dwm, and I've watched with horrid fascination as the webserver was nearly toppled by the tremendous /.-onslaught just recently (have a look at the logs). Now, browsing the comments, I thought I should post and clear up some points.

First of all, yes, 3Dwm is misnamed. 3Dwm is NOT an X11 window manager, it is a user environment (the beginnings of the 3D-equivalent of X11). However, the name has stuck with us since our first appearance on Slashdot, so we don't want to change it.

Secondly, the main platform for 3Dwm is not normal desktop computers (though it does run on desktop systems), but Virtual Reality devices (like this one). In Virtual Reality, you have some amazing 3D interaction possibilities that few existing applications exploit.

As for VNC support, 3Dwm has VNC client (not server) functionality, just as one observant slashdotter pointed out. This allows us, in a network-transparent fashion that is in keeping with the distributed nature of the rest of 3Dwm, to display graphical desktops of any major windowing system (including Windows, X11, and MacOS) in 3D.

There's always skeptics who wonder what you would use a system like this for when 2D is perfectly fine. To that I can only answer that there are, in fact, areas where 3D could help a great deal, mainly in the fields of design, modelling, and information visualization. Why, take a look at this (and this and this) screenshot for a prototype 3D web browser.

Btw, today marks the one-year anniversary of our last slashdotting (I wrote up a short summary of the comments we got last time). Cool, eh? :)