I've mentioned Karl Voit's Memacs project before. I'm really interested in its ideas and have been implementing various manual procedures to capture some of the same data. My end goal is to implement Memacs on my own machines but it may take some work as Memacs is implemented in Linux and some of the data gathering modules may have to be reworked to function on OS X.
If you're interested in more information about Memacs, Voit published a very nice description of it on arXiv.org, which you can get for free. The paper goes into the rationale for Memacs, its basic architecture and then a use-case story that illustrates how it's helpful.
It's an interesting paper and not too long so you should give it a read if you have any interest at all in personal information management.