I would change how you select and deselect, so it works like practically all other 3D software. That is, LMB-drag for box-select, and clicking in emty space deselects. Nothing wrong with ‘a’ button (select/deselect All), but pressing a,b,a,b,a,b,… all the time, is a little offputting, when it could just work like everything else. No reason to be different for difference sake.
If someone can tell me how the current selection method can improve workflow, over the de facto method, I will gladly stand corrected. Also, how does right-click select improve workflow over left-click?
Kill your darlings, maybe? (no offence, blenderheads)
when you quit, blender saves automaticly quit.blend to your temp directory.
if i could change something, i would make it possible for python gui scripts to use same widgets (and views) as blender does, instead of always building your own with ridiculous opengl commands…
(but i guess this even might be possible after event refactor…)
While that “works,” it forces you to use a separate lamp and different render layers (or lighting groups) in order to get correct shadows. also, you lose the light filtering/blocking effect from the halo lamp, which is one of the points of using them in the first place.
It would be much nicer, simpler, and mathematically sound to just implement it as in the original post.
@ san: you know you can change that behavior (left/right mouse select) in the user preferences, right? It would be worth considering whether that should be the default, though (I use right mouse, but for new users, it could be confusing)
I don’t have any really simple things… just complex things, and I don’t want to turn this into a feature request post.