For advanced users the area operations invoked at the corner action zones are (an obscure but) efficient way to manage areas. But I propose a small change that can make it a bit better...
Currently if you drag out of an area a join is proposed. It is not done immediately, but waits for your confirmation. And before this confirmation you can change your mind and even join in the opposite direction if desired. However, when you drag into the area a split is completed almost immediately. The only interaction is after the fact in that you can continue to drag and set the new edge location where you like.
This means that these two operations do not have equal chances of success and failure. While learning, a new user is far more likely to create multiple unwanted splits since this happens without confirmation.
This patch adds that confirmation, and with it more features. And does so without changing the existing mouse operation. It still starts with a mouse down, drag, and ends with a mouse up. But simply changes the moment the split happens to the mouse up time. Instead of the split happening immediately you instead see a highlight line showing you a proposed split location.
Because the mouse operation does not change at all it still feels quick and natural. But the timing change instructs the user a bit better since nothing happens without confirmation.
And there are side-benefits....
This is more consistent with how it works when invoking with the edge context menus.
While dragging the mouse (and showing the proposed split line) you can actually move your mouse to ANY area and therefore split ANY editor, not just the one you started from.
While dragging your mouse you can press TAB and change the direction of the split.
So, for example, you can start with a downward drag from the Properties action zone. You will see a proposed horizontal split of that editor. Then drag to the 3DView and see that you could split there instead. Hit TAB and you can now split the 3D view vertically. Or just hit ESC to cancel and do nothing...
Not sure what others might think of this, but I find it only a positive change without any negative impact. Doesn't describe well though, so has to be actually tried in order to evaluate it properly.