"What are the best practices for shipping high-quality software?" This question is an example of a question that I will not answer in my talk. Instead, I will describe why any software project that contains more than 10 lines of code is guaranteed to fail. Using my fragmentary knowledge of the laws of thermodynamics, I will explain why code is like a proton that must ultimately turn into a crystal or, uh, whatever it is that thermodynamics says will happen. I will demonstrate how unit tests, functional programming, and UML diagrams fail to address the primary source of software failure (namely, that software is an inherently bad idea because our brains evolved to hunt giant sloths with primitive stone tools, and MongoDB only partially resembles a giant sloth).
I will conclude the talk by luring a group of agile programming experts into a large cardboard box using a collection of buzzwords like "evolutionary development" and "cross-functional team;" once captured, they will be forced to implement obscene, poorly-specified COBOL algorithms as I laugh maniacally and disable my compiler warnings.