Just a note for those who are interested in such things. Having said I couldn't find any reference before, both the 16F88 (18X) and 16F872 (28X) datasheets state an absolute maximum Input Clamp Current of +/-20mA. That probably applies to all modern 16F variants / PICAXE's, but always check the datasheet. I've read the 16F88 sheet numerous times and never noticed it there 
For the uninitiated, Maximum Input Clamp Current ( aka Maximum Injection Current ) is how much current can be pushed into / pulled from an input from a voltage below 0V or above +V before the internal diodes melt and the chip explodes. This is how the Serial Download Interface Circuit works with its 22K current limiting resistor. In the worst case analysis; RS232 at +/-15V, PICAXE PSU 0V/+5V, assuming input nominally sits at 0V, the 22K drops +15V (TX=+15V), from I=V/R, 682uA, assuming it sits at +5V the 22K drops -20V ( when TX=-15V ), -909uA. Well below the maximum allowed, and in practice the voltage drops will be lower.
In many cases the 22K can probably be safely reduced to 1K5 if that were necessary to resolve any interfacing problems.
For the uninitiated, Maximum Input Clamp Current ( aka Maximum Injection Current ) is how much current can be pushed into / pulled from an input from a voltage below 0V or above +V before the internal diodes melt and the chip explodes. This is how the Serial Download Interface Circuit works with its 22K current limiting resistor. In the worst case analysis; RS232 at +/-15V, PICAXE PSU 0V/+5V, assuming input nominally sits at 0V, the 22K drops +15V (TX=+15V), from I=V/R, 682uA, assuming it sits at +5V the 22K drops -20V ( when TX=-15V ), -909uA. Well below the maximum allowed, and in practice the voltage drops will be lower.
In many cases the 22K can probably be safely reduced to 1K5 if that were necessary to resolve any interfacing problems.