"The problem is the stupid RC, not the bootvector"

You said it not I.

That too is from the datasheet. These mistakes will occur with any devp, but so late in the day is not acceptable, I mean they were not pioneering this concept in microcontrollers. To know, when I am in trouble that the only way to restore the boot vector is to get hold of a parallel programmer adds to my headache.
Also note that the local support from Philips was very good, but a couple of times they were also caught by surprise.