When I read the article it reminded me of some ideas I had for for UI design. I get really frustrated in my IDE when I have to click on a folder or a file and i

When I read the article it reminded me of some ideas I had for for UI design. I get really frustrated in my IDE when I have to click on a folder or a file and it is so small. It seems that systems could build statistical models of what I do all day and resize icons based on what I am likely to click. There seems to be this area between 100% automation and doing things manually that isn't well designed for. Another example is when I have to go and look at log files, I have to click on these tiny folder icons but I do it every day ... why can't the system predict that and make the folder icon larger? Or how about while I'm in the IDE only show things that make sense at that point in time? You might say well there are thousands of things you can do in an IDE but I know that I do about 15 or 20 things day in and day out. I know I could use macros or something but that would be unique to each tool and would have a very stitched together feel to it. In my opinion the tab completion in the Bash shell is an example of game mechanics used outside of a game.