Measuring the security of code obfuscation is difficult. A novel obfuscation
transformation is in some cases only measured in terms of code expansion and
speed, which are in fact only side effects of the transformation. A first step
to define a security value to an obfuscation transformation could be having a
look at what a cracker is able to reveal from an obfuscated program. This
abstract first of all gives a short overview of existing techniques to
obfuscate. Then, we describe existing techniques that can be used to
deobfuscate, which were sometimes originally meant for other purposes, and new
techniques which we are working on to deobfuscate.