I have nothing against a graphic interface, I just question whether that is ergonomically efficient when you are processing, say, 1000 images. My bias in these situations is to emphasize a keyboard based solution.
In this case, I would prefer being able to apply frequently used keywords with a single keystroke. Unfortunately, this is not practical as most single keystrokes (e.g. 'o' or Ocean) are more often used for entering text into (surprise) text fields. Other key combinations like Ctrl+O may already be used.
I would suggest, at least for Windows user, that the Ctrl+Alt+Shift plus letter combinations be made available for custom keywording. I realize that actually typing Ctrl+Alt+Shift+O to keyword "Ocean" involves a lot of finger gymnastics, but if you can assign that keystroke to a programmable input device like a Belkin Nostromo, then it would be easy to quickly apply keywords during the editing process.
Deven