Self-encrypting Code to Protect Against Analysis and Tampering (J. Cappaert, et al.) comes to mind.

A simpler way to provide strong security against reverse-engineering would be polymorphic encryption. Oddly enough it's a technique that originated from the malware hobbyists and although it's nowadays used by serious software developers a lot of good documentation and examples can be found on sites focusing on malware development and (software) security analysis.

It is disputed whether or not it is possible to create a software that can't be detected and therefore analyzed (see Blue Pill malware), although it is possible to protect the code good enough so that the effort of cracking it outweighs the worth of cracking it.


Although you could write every single function of your program so that it could be encrypted and loaded into memory only one function at a time (which makes both recursion and object-oriented programming impossible, by the way) it is highly unlikely that you'll find a graphics library that works with that kind of programming. It is also highly unlikely that you'll need that kind of security. As frob said, you need to consider who you're protecting yourself from and how much your data/code is worth to them.