Hand-written notes don't need to be necessarily 'well-designed' in the same way a website does; you aren't going to be using them for the first time, so it's sort of like code you're writing for yourself that you know you'll never have to use after tomorrow.
But it's important that they be usable, in that they should answer "where-am-I-going-with-this, how-does-this-tie-together" questions for you. Because I participated in that tactile process of writing the notes, they can do that more effectively than something I typed; an interesting squiggle can express my intention to myself better than any comparably terse character on my keyboard.
I'd bet we take and make ugly, idiosyncratic notes for the same reason that we start out writing ugly, idiosyncratic code: it's memorable. (In lisp, 'idiosyncratic' code is close to idiomatic code ;)).
When I'm trying to think about something and am stymied, I need to doodle as well. Of course, there's very little transference with something that's ugly and idiosyncratic, so his notes do nothing for me and mine, likely, would do nothing for him.
The imagination is too elusive for the rigid construct of a word processor. I can get into the zone without drawing. However, the more difficult or creative problems always solve better on paper. How is it that a supercomputer cannot help solve these problems as well as a piece of paper?
All the qualities of both?
