>There is a way of rendering any PIC secure against all non-invasive attacks

>[2] Nothing else on the chip SEEMS to be damaged.

Good point. Best put some type of test routine in to test all the
instructions and features. Even then, Who knows what chemical reactions,
thermal stress or lattice metal migration changes have taken place?

>Note that without the ability to electrically tell the chip to read the
>program data, the crackability of the code-protect fuse will be a moot issue;
>if the i/o transistors for PB6 are blown, it's a bit hard to "unblow" them.

Be carefull. I've heard you can recover data from a hard drive that's been
over-written. The magnetic media is not 'digital', its analog. If you read
it using an analog head, the level of magnetization will vary according to
the previouse data. I've read to be realy safe you should overwrite with
random data at least 5 times!  Like a desk blotter or typewriter ribbon, you
can read a faint analog imprint of what was there.

Has anyone taken an analog scope and seen if any weak signal leaks past a
blown driver?