I wouldn't dismiss this so easily. There are some serious problems with the target board circuit recommended by Microchip. Many people at Microchip don't seem to be aware of the issue. I've spent quite a bit of time trying to figure out why a simple target circuit wouldn't work with the ICD-2 and my own programmer. Everything was wired according to the Microchip recommendations, but it still didn't work.
I won't go into all the details here, but when the target chip drives the PGD line, that hard edge can couple onto the PGC line. There is absolutely nothing the programmer, ICD2, or whatever, can do about it at the other end of the standard RJ-12 cable. That's because the entire effect happens in less time than the propagation delay out the cable and back. Matters aren't helped by the fact that the PGD and PGC lines are next to each other in the cable.
The solution I'm using for all new target boards now is to put 22pF on both lines to ground right at the connector on the target board. In addition, I put a 100 ohms in the PGD line from there to the PIC. This seems to work reliably. The resistor and cap slow down the PGD edges produced by the PIC, and the cap on the PGC line further reduces its susceptibility to coupled transients.