>>I have posted a message here some time ago about someone claiming to be
>>able to unprotect any '84 with a device called the picbuster II.
>>Now I read in alt.satellite.tv.europe that the picbuster III is coming.
>>I haven't heard any details about this device yet.
>
>I have a program called picbust. I don't have a pic programmer so I don't
>know if it works but if some one want it I could post it to the list.
>
The '84 is reportedly easy to bust, though I haven't tried it myself.  I can
think of a similar method for the EPROM models, by raising the Vdd voltage
until the unit senses the CP bit as unprogrammed, then programming groups of
4 zeros over the program and X-oring the verify codes at each stage to
determine what was programmed (this would destroy the program in the PIC,
but of course a new one could be programmed with the now cracked program).
Since reprogramming of the first 64 locations is not disabled by the CP bit,
these can be cracked immediately with nothing more than an ordinary
programmer.  At the very least, put sensitive code or data in the top part
of the EPROM which is somewhat more protected.

The only defense I've heard of (If Microchip is putting new features in the
PICs to defeat cracking, they aren't saying anything here) is to burn out
the pin buffers on the pins used for programming / verifying (B6 and B7 on
14-bit units, probably B0-B3 on 12-bit units would be sufficient) after
programming.  Of course the pins are then not available for I/O purposes
either, and the PIC may end up destroyed entirely.  This should defeat most
"garage" methods that use the existing program/verify paths, but more
advanced ones like opening the package up and scanning the chip will
probably never be stopped.

-Mike