From: J. Forster (jfor@quik.com)
Date: Fri Aug 05 2005 - 17:44:04 PDT
bruce C. Beattie wrote:
> I sure hope you are right about that John,
> Now we only have to deal with one problem instead of two,
> although like you said, it is not exactly a small problem.
Let me clarify a bit. It might be possible for an evil doer to build an undocumented
instruction into a chip, and such an instruction might cause a chip to go nuts. An
early example of this (not by evil design though) was the exp(ln(2.02))=2 bug in the
very early HP 35 calculators. But, it would require very clever SW modifications for
some baddie to cause the activation of such a bug.
IMO, it's much more likely that someone could get into the SW and cause a smart bomb
to self destruct or miss, if it were targeted within a certain Lat/Long box That
would only take a few (likely less than 5) lines of code. Chips are exhaustively
tested under actual working conditions. Bombs are not.
FWIW,
-John
This archive was generated by hypermail 2.1.4 : Fri Oct 28 2005 - 23:25:26 PDT