Bryan Ford's 2002 Masters Thesis is remarkable in that it breaks decades of compiler-construction dogma with some simple principles and a compelling alternative to the complexity of parsing with context-free grammars (CFGs). He reveals a forgotten class of grammars-- top-down parsing language (TDPL), and some extensions known as parsing-expression grammars (PEGS) -- that directly correspond to the parsers that implement them. His primary contribution, however, is applying modern functional programming techniques of laziness and algebraic data structures to make TDPL/PEG parsers computationally efficient.
Presented at Papers We Love Chicago #1.