I went back to my computing roots. Once upon a time, I thought that there was no future in little computers, and the big money was to be made doing systems work on mainframes. I held a succession of jobs, growing in importance and responsibility, until I changed fields in 1996. I still remember a lot from my days as a systems type, though.
There's a program that measures how long it takes an IBM mainframe to execute each instruction in its set. Of course, there's no source code available, so I get to disassemble it (turn it back into readable text). I spent a happy couple of hours getting the disassembler installed and running, and made a first crack at the disassembly.
Unfortunately, the disassembler I'm using is old, and doesn't handle a lot of instructions. I get to enhance that, too. Whee.
At least this is stuff I know how to do, and enjoy...