I'm sure there's no 'hard and fast' rule, but I gotta ask...

Assume your PIC has been programmed, and it's I/O pins are doing their thing. Is there any rule about what happens to these I/O's when you reprogram the PIC? To my surprise mine seem to hold their state during the programming, and then do what they're suppose to again once the new program has started. The reason I mention it is this PIC is controlling a large security system, and holding things locked. I was afraid that during programming an update, the pins might/would go high or open collector or ??? and unlock the locks, but it isn't doing that, and I'm pleasantly surprised.

Are you talking about programming with MCLoader, or with a device (like an IC2/3 or PICKit3) that pulls MCLR to a high voltage for programming?

circuitpro, Alot has to do with the design of the particular drive circuit. I always design my interfaces to have the state of the particular driver to maintain an OFF state when the processor is in a RESET state as the output pins of the processor present a tristate level. As far as I'm concerned, any good engineer should think about this as it is the first state the processor comes up in when power is applied or when the processor is placed into a RESET state whether remotely or thru a local reset condition.....