So, how to apply computers to computing...and really both of my points previously (the id of the computer and software visualization) are finding ways to make the computer more useful to itself. Why is it so hard for a computer to self-recover? IBM's autonomous computing is drilling at that idea. But are they going in the right direction?

Ooh, what if we think about a dual processor, not as things to co-process, but rather as two semi-independent hemispheres. Or, alternatively, what if we used one processor as the 'concious' element of the computer and dedicated the second to be the 'subconscious'? Dedicated processing could go on that would help put context and add the relationships to the events happening in the other hemisphere....

Interesting thought...