[SoC] Proposal for wxGUI layout modifications

[Hamish]

Michael Barton wrote:
> Actually, I think what (I hope) was
> envisioned was not a true MDI--which has lost considerable
> popularity--but attaching the layer manager to the display
> window. Ideally, this could be done by docking or
> undocking.
> 
> That said, Markus and I had some offline discussions about
> this since he was the proposer, in response to some user
> comments and his own experiences. The main complaint is
> having too many windows open in GRASS.
> 
> After some talking about it, the real problem may not lie
> primarily with the layer manager/display windows but with
> all the module windows which a lot of people don't realize
> that you can set to close automatically when you push OK.
> Some possible solutions to this would include having all
> modules open in a single window with tabs or modally in a
> single window or modally in a window attached to the layer
> manager to keep down clutter. By focusing a bit much on the
> initial display configuration of a couple of other GIS
> programs (ArcGIS and QGIS, but not in a lot of others by the
> way), we may have missed the real cause of the perceived
> screen clutter. Anyway, it's something to think about.

[we should probably move this to the grass-dev ML]

The Output window in the wxPython gui is now a tab of the main control
GUI, so that's one down. Yeah, nice idea- all module GUIs open in a
single window, adding and removing tabs for each at the top as they
live and die. Then (assuming the terminal window is minimized) you have
a max of 3 open windows, which isn't so bad.

it is nice to keep a module gui open so you can tweak one parameter and
run it again. it is nice to keep a couple open and switch between them
if you are doing something on a multi-processor machine (if doing so
within a single mapset, at your own responsibility and risk that the
computation region etc is not changed of course :) for that reason you
might want to have e.g. 3 r.in.xyz's running at the same time, each
crunching along on a different statistical method.


Hamish