Re: Separate virtual desktops for different monitors



On Tue, 3 Mar 2015 09:13:21 +0100 Rhialto <rhialto falu nl> said:

On Tue 03 Mar 2015 at 08:03:52 +0300, Pavel Kretov wrote:
Most non-tiling window managers (if not all of them) combine all available
monitors into single virtual screen space which gets switched as the whole
thing when user changes virtual desktop, exactly as the specification says.
But there exist at least two tiling managers, namely i3 and xmonad, which
employ the different concept: they allow user to setup own set of virtual
desktops for each of monitor. (Mac OS X has a similar feature too.)

ctwm has a similar but even different possibility.

Say you have 9 workspaces (that maps to what the spec calls desktops).
Then if you have multiple monitors, you can show workspace #x on monitor
1 and workspace #y on monitor 2. You can make any combination, but not
show the same workspace twice (since windows have only a single parent).
Also windows that show up in multiple workspaces are shown only once
(for the same reason).

This works best if both monitors are the same size, of course.

This is also not very well matched by the spec. There are multiple root
windows at the same time, for instance, and multiple current workspace
numbers.

-Olaf.

enlightenment also assigned an array of NxM virtual desktops per screen. as
screens can (and often are) different resolutions, you don't assign a workspace
to a screen - it's bound to the screen by its nature, but yes - unlike fdo
spec, there can be a different set of virtual desktops per screen and thus e
supports none of the fdo specs when it comes to virtual desktops/pagers as the
spec is lacking ... and tbh e has it's own pager and panels/shelves all built
in and has no need for an external one. the pager in e is now bound closely to
the compositor as it keeps a "perfect up-to-date" miniature of all desktops and
any external pager would miss this plus a host of features, unless it went and
implemented a mini compositor of its own - way too much overhead there. and
given the move to wayland, indeed these specs become redundant as all of these
kinds of things (taskbars, pagers etc.) are intrinsically bound to their
compositor, and likely are implemented in-process or as a tightly-coupled
companion process.


-- 
------------- Codito, ergo sum - "I code, therefore I am" --------------
The Rasterman (Carsten Haitzler)    raster rasterman com



[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]