A few miscellaneous items accomplished:

• Finally got the interrupt handler working properly on the microcontroller; now I have just about everything needed to get the main sensor/effector loop implemented.
• Discovered that the TPU function I am using for the accelerometer is not really exactly what I want; the accelerometer's output pulse period can drift over time and the TPU function cannot really detect that.  So now I have a new item:  write custom microcode for a new TPU function for the accelerometers.  I needed another steep learning curve, I guess!
• Spent some time learning the "freeware" version of the Eagle circuit/PCB design program.  The circuit is complete for the contact sensors, but the board isn't correct yet.  Worked on defining the new circuit for the vision system.  I was planning to just have the MPC555 do the vision loop along with everything else, but I was concerned that the 16,000 interrupts I would have to handle per image would load the CPU too much (purely because of the state-saving and restoring necessary in an interrupt handler).  So I have gone back to the idea of using the external Atmel AVR board to do the data collection.  I'll store the image in a dual-port SRAM which the MPC555 can access at leisure.  This will be a rather complicated PCB for an amateur like me, though, but it should be an interesting challenge.

Most of my "robot time" this week was spent using my CNC mill -- learning all kinds of stuff.  It is quite time consuming, though.... I had expected that maybe I could sort of push a button and out would come a part.  But it turns out that there are reams of "g-code" (CNC control code) to write; and the process has to be stopped quite often to rearrange clamps or change machining bits.  Plus the bits sometimes break or don't behave that well, and there is a lot to learn about feed rates and backlash compensation, and so on.

I did an experiment in cutting a PCB board as a test (I had hoped that machining with a tiny bit would be a good way to make boards).  So far the results are promising but I have some tuning to do as I broke two bits trying to make a small board.  I'll continue these experiments; I think it will work well eventually.

And I cut two parts of Bing's right foot.  The first didn't come out that well because I am just learning (not to mention the small pile of aborted attempts), but it is usable.  Someday I'll redo it but first I want to get more stuff done.