Written by testian the 19 Feb 13 at 08:38.

Thats it basically. The user is just like a package with variable dependencies, but here some hints on the implementation:

The user is allowed to execute apt-get install. The requested packages are installed and marked as requested-by or manually-installed-by the executing user. Just like the behaviour is today for root.

If the user executes apt-get remove his marks are removed.
If the package is no longer referenced it is removed (just like autoremove)

The base installation is marked manually installed by root.

If a user is removed his marks are removed (and then autoremove will remove his packages solely requested by him)

In case of conflicts the operation is aborted and the user is asked to resolve together with the conflicting users or root.

Thats it basically. The user is just like a package with variable dependencies, but here some hints on the implementation: The user is allowed to execute apt-get install. The requested packages are installed and marked as requested-by or manually-installed-by the executing user. Just like the behaviour is today for root. If the user executes apt-get remove his marks are removed. If the package is no longer referenced it is removed (just like autoremove) The base installation is marked manually installed by root. If a user is removed his marks are removed (and then autoremove will remove his packages solely requested by him) In case of conflicts the operation is aborted and the user is asked to resolve together with the conflicting users or root.

Written by testian the 6 Jun 12 at 13:26.

Any item being a problem should become crowdfundable per se.

Evidence should be collected to estimate Demand (For bugs, for example amount of affected people), and Supply Available resources for working on the problem, complexity.
However - maybe it's better to define a funding limit, kick-off and deadline moderated.

Money goes to the problem solvers, whoever that is. (Also needs moderation).

Any item being a problem should become crowdfundable per se. Evidence should be collected to estimate Demand (For bugs, for example amount of affected people), and Supply Available resources for working on the problem, complexity. However - maybe it's better to define a funding limit, kick-off and deadline moderated. Money goes to the problem solvers, whoever that is. (Also needs moderation).

Written by testian the 6 Feb 12 at 19:45.

This negative space is like the middle part of a paper fold in an N-shape.

The windows would then be the vertical parts of the N.

As you know windows with equal Z or depth location (this is normally not possible) can't overlap.

Now what happens to the N-shape if you try to make the two vertical parts onto the same layer/Z-Location? The N-shape is unfold and the negative space becomes positive (or 0 if you look at it as a unstretching rubber band). The 2 windows are now tiled.

In other words: Window tiling can be done by having one window trespassing the Z-Space of another window, forcing the windows to move out of their ways with a smooth animation.

To reduce the amount of locations in Z, windows which do not overlap have the same location in Z if there is no evidence that they can't be in the same Z-location (a window is between them and overlaps with both).

Note that in this solution there is no visible grid, but I think (but I'm not sure) mathematically any window arrangement situation as we have them in todays window managers can be described using a grid with irregular cell sizes IF negative spaces are allowed to do the overlapping.

This negative space is like the middle part of a paper fold in an N-shape. The windows would then be the vertical parts of the N. As you know windows with equal Z or depth location (this is normally not possible) can't overlap. Now what happens to the N-shape if you try to make the two vertical parts onto the same layer/Z-Location? The N-shape is unfold and the negative space becomes positive (or 0 if you look at it as a unstretching rubber band). The 2 windows are now tiled. In other words: Window tiling can be done by having one window trespassing the Z-Space of another window, forcing the windows to move out of their ways with a smooth animation. To reduce the amount of locations in Z, windows which do not overlap have the same location in Z if there is no evidence that they can't be in the same Z-location (a window is between them and overlaps with both). Note that in this solution there is no visible grid, but I think (but I'm not sure) mathematically any window arrangement situation as we have them in todays window managers can be described using a grid with irregular cell sizes IF negative spaces are allowed to do the overlapping.

Written by si14 the 1 Mar 12 at 07:38.

This page explains what's gTile: https://extensions.gnome.org/extension/28/gtile/
It's not a full-blown tiling WM, of course. But it's way better than anything presented in Ubuntu right now.
The only thing that is lacking in that extension is an ability to save workspace configuration or assign application to position.

This page explains what's gTile: https://extensions.gnome.org/extension/28/gtile/ It's not a full-blown tiling WM, of course. But it's way better than anything presented in Ubuntu right now. The only thing that is lacking in that extension is an ability to save workspace configuration or assign application to position.