Mark Wedel wrote: > > One request that I've made before would be for the map windows to be > their own top level windows managed by the OS instead of subpanes within > the editor. Alternatively, if the layout could be a bit different so > that the selected object information on the map was all done in a second > column to the right of the one that is used for selecting stuff to > insert in the map. > > The odd geometry of my system (3200x1200) means that even if I maximize > the window size of the editor, I effectively lose a lot of editing > space, as the pane for selecte object spans the entire width of the > screen, which isn't very useful. I actually did some quick hacks, and at least have my first point working (individual top level managed for each map). I need to hack around some more - the layout of the main window is now less optimal, because the pane area that used to be used for that information is completely unused. I need to poke around some more and see about moving over the map information to that pane - then I could have a 400x1000 (or thereabouts) window for that, and devote the rest of my screen for the map stuff. The problem with this is I'm not sure how to deal with this as a configuration option. I'm a pretty java newbie, and the change I basically did was change the map stuff from inheriting from the JinternalFrame to JFrame. I have no idea how that could get done as a runtime option (maybe inheriting from both, and using the correct constructors depending on the desired behaviour)? There is also what is perhaps some other behaviour I haven't figured out - using the close (delete) feature of the window manager on these new map panes is ignored - I'm sure there must be some setting someplace to handle what to do on a requested close event. The close from the editor menu does close the window. I'm not sure how it figures out which one would get closed - I'm guessing the one last selected. The easiest solution may be to just include the diffs within the editor distribution and have people just apply the diffs if they want this different behaviour. I suppose for a 'binary' distribution, both versions could be supplied.