From obfuscation to comprehension

Code obfuscation techniques are widely used in industry to increase protection of source code and intellectual property. The idea is that even if attackers gain hold of source code, it will be hard for them to understand what it does and how. Thus obfuscation techniques are specifically targeted at human comprehension of code. We suggest that the ideas and experience embedded in obfuscations can be used to learn about comprehension. In particular, we survey known obfuscation techniques and use them in an attempt to derive metrics for code (in)comprehensibility. This leads to emphasis on issues such as identifier naming, which are typically left on the sidelines in discussions of code comprehension, and motivates increased efforts to measure their effect.