Chris Petrilli's second brain
I have been more frequently buying PDF books: not as nice to read, but in the future it is great to have indexed and online. Because I mostly use a MacBook with open connections to my customers' servers for most of my work, my MacBook is the home for my "second brain". I have experimented with doing my own indexing and document clustering (naive O(N^2), hierarchical, and k-means), using Spotlight, and using Google Desktop for the Mac. I have settled on using Spotlight, but using the text, HTML, and PDF's in my "second brain" for my own research programming for IR, etc.