They prove that there can be no "Obfuscator"
which works for all functions.
essentially, they show that there are some functions which can never be obfuscated,
and hence no universal obfuscator can exist.
also, their model of obfuscation is rather strong.
they require that the obfuscated program be as good (or worse) as a black box.
the idea is that if you have a function which stores some kind of secret (otherwise why would you obfuscate) then however you obfuscate it,
you can always tell something more if you have access to the tapes(or memory) of the program works with, than if you dont.
Slashdot Mirror
This is exactly why i have an id called YouCantReallyTellWhoIam:)
They prove that there can be no "Obfuscator" which works for all functions. essentially, they show that there are some functions which can never be obfuscated, and hence no universal obfuscator can exist. also, their model of obfuscation is rather strong. they require that the obfuscated program be as good (or worse) as a black box. the idea is that if you have a function which stores some kind of secret (otherwise why would you obfuscate) then however you obfuscate it, you can always tell something more if you have access to the tapes(or memory) of the program works with, than if you dont.