|
| "nemo_outis" <abc@xyz.com> writes:
>Moreover, even TOTAL open-source is not invincible. The classic
>exposition is Kernighan's "Reflections on Trusting Trust" which is widely
>available on the net. (PS Kernighan is one of the co-originators of the
>C language, etc. so that gives you some idea of his heavyweight
>credibility). Read it and weep.
This was essentially a joke. Withe the precognitive abilities required to
know how compilers would be organized in the future, the writer of this bug
could make himself far richer and influencial playing the stock market.
>However, for those even more theoretically inclined, it can be shown
>(using reasonably broad definitions, constraints, etc.) as a corollary of
>the "halting problem theorem" with respect to a Turing machine that there
>CANNOT BE a universal process that can show an arbitrary program to be
>free of "malware." The demonstration is left as an exercise for the
>reader :-)
Same is true of recognizing all future compilers and how to insert malcode
into it.
|
|