Richard,
The really nice thing is that we are not dealing with RF generated noise. If U-dUd keeps to a linear model he will not have to either. Sanyo provides lots of charge and discharge graphs for each of their cells. Have you played with the PIC processors? I'm not saying that circuit layout is not critical at times but it is a lot less than I expected after dealing with dedicated embedded PC's of the early 90's. Lot's of manufactured experimental boards are out there with an area for analog breadboarding with the PIC area already fully debugged and assembled.

One other point power supply design can be a real ***** because you don't know if the load is reactive or inductive in nature. Designing a NiCd or a NiMh charger is simply a lot easier because you know exactly what the load is and your can use a proven power supply to power it. The other simplifying factor is the environment these chargers operate in is not an industrial one with lots of dirty AC but one in a residential setting.
John