> Multiply this by 4 monitors and 8 Linux workspaces and users have
an awful lot of screen real estate to get lost in. The main synoptic
typically gets buried in the background, so the users open another one,
then you have multiple copies of everything and can never find what
you're looking for.
* 30 workspaces, dual screen per OPI x 5 (only 2.5 are heavily used though)
* Great for commissioning, not so for normal operations
* Our stretch goal is to have 4 summary work-spaces (i.e to be used for normal operations)
* Currently, we have motif, tk, matlab, labview, CSS, Qt... in a GTk environment.
Well, we are moving everything to Qt.
> Now the simplistic approach when converting to CSS BOY would be to translate every EDM screen into a BOY screen, launched in its own window. But then we aren't really solving any problems, we're just making BOY look like EDM, and it takes a lot of work to persuade eclipse not to put toolbars everywhere...
As Emma mentioned, if you want to have a look at dock areas and dock-able tabs, simply start qt-designer.
Regardless of the framework, with this approach they are limitations.
* You also cannot dock 2 tabs outside the dock area (i.e outside main window or on central widget)
* If you close the main window (the one with the menu and the central widget), you also close the detached tabs.
That behavior may or may not be the desired behavior
If you want 2 independent windows (you can close one without closing the other), you need 2 main windows.
(In Qt, a widget with no parent is automatically contained in an menu-less main window)
If you put one main window on the top of another that implements a dock-area... well, they just stay on the top of one another.
What you need here is to have a look at window managers.
(the fancy beryl-compiz or those included in Windows/OS X/Linux(Gtk/KDE) http://en.wikipedia.org/wiki/Window_manager
That being said they are new kids on the block, like win8/metro and Qt/QML both designed for touch interfaces.
With metro and QML, you navigate on a canvas. Certainly, we are a few years away from those touch-based interfaces.
How long can you wait?