Wayland 3D Compositor with Oculus Rift and Razer Hydra [video]

I love the idea of coding on a giant virtual wraparound screen but I'm not sure even a DK2 is going to be up to it. Think about it: The text in your windows can't look "fuzzy" (antialiased or rendered) at all. It really has to be "pixel perfect". We go nuts around here over fonts that get this or that pixel in just the right place on a given serif.

If you doubt it, try setting your monitor to something other than its native resolution and coding for a while using the interpolated resolution. I'm afraid text in floating 3d windows on vr sets is going to look like that for a long time to come.

I'm still going to try, dammit.

(No, I'm not the author.)

Code: https://github.com/evil0sheep/qtwayland-motorcar-compositor

Thesis: https://github.com/evil0sheep/MastersThesis/raw/master/thesi...

Defense Slides: https://docs.google.com/presentation/d/1svgGMxxbfmcHy_KuS5Q9...

Heh. Reminds me of http://www.blendogames.com/qc/ http://en.wikipedia.org/wiki/Project_Looking_Glass http://www.nngroup.com/articles/2d-is-better-than-3d/

This is really cool. I ordered a DK2 to try out some similar things in a similar domain. The idea of a 120° workspace which can be expanded indefinitely seems like a reasonable alternative to getting bunches of physical screens on a desk.

The beauty of this approach is that it frees up window managers from managing desktop spaces and instead allows you to manage windows and the relation of the windows relative to other windows. Being freed of borders and boundaries presents amazing opportunities. Defining things in absolute space is very confining. Defining things in relativistic space, gives you the opportunity to move into multi-view/multi-screen modes with multiple participants much more easily. It's not longer about desktops, but about taskspaces.

I don't want to say "put window A on the lefthand side of the desktop and put window B on the righthand side of the desktop." Instead I want to be able to say "put A to the left of B in when mode X" or "put C on top of D when performing task Y."

It's simply much more natural and more intuitive to tell the computer where your focus should be (or if the focus should be split equally between 2 or more windows), and then defining all the peripheral windows that should appear around the focused window. Ideally these wouldn't even separate windows, but simply a rectangular stream of pixels from the buffers of your choice, so that the same application could be multiplexed to layout modes. This would basically be tmux for windows/apps in 3D space.

A prolog-like declarative language with relations and predicates for describing windows and how they are organized relative to one another under variable conditions would be awesome. Type one command to switch to the window you want to have focus in the center of your screen. Once that window is in place, bring into focus the last used "mode" for that window, with all the peripheral windows coming in from behind you or zooming in from the back (time machine style, but with your primary focused window held constant). Then you can activate switching between modes with another key stroke combination.

This is pretty cool as far as tech goes, but also seems rather ridiculous practically. Productivity with a HMD will likely amount to a lot more than the same 2D dynamic images we already have(screens) being placed arbitrarily in 3D space. What will it look like instead? one may ask, to which I have no definitive response.

What we have here seems like an extrapolation of what we have now, which may prove fruitful if we have nothing else to go off of (and is the nature of incremental progress). If I had many more monitors I'd place a class (or any other abstraction) on every screen, so maybe every class will have its own 2D plane of text and they'll be organized semantically & dynamically in 3D space. Also resolution here needs to be really really high. Idk, just interesting food for thought.

It'd be fantastic to have an Oculus-based desktop environment! Imagine not having to drag giant monitors around. Combine it with xmonad and you're golden. I hope this becomes reality soon, at least for development purposes.

Shut up and someone take my money already.

I feel like jumping straight into floating resizable screens is too great a leap.

I think we have to start off by simulating a desk and two fixed size monitors so the user understands the metaphor, nail that experience and then evolve from there.

Another 3D Wayland compositor example: https://www.youtube.com/watch?v=_FjuPn7MXMs

nice one. It reminds me The Lawnmower Man movie. Soon, you will be able to physically dive into the code.

The one effect I really really want is when you switch from 2D -> 3D mode with the Goggles I kind of want the 2D desktop to do the laser activation effect from Tron Legacy.

I hope the author ports this to something else than Oculus, since I bailed on it when it sold out.