For my undergraduate senior thesis, advised by Drew Hilton, I set out to solve the following problem: Verilog and other HDLs lack the expressivity and modularity of modern software programming languages, due to a lack of features such as a strong type system, recursion, pattern-matching, polymorphism, higher-order functions, and more.
Specifically, I aimed to answer the following two questions:
Can I design a programming language that combines the powerful features of software programming languages with the ability to describe electronic circuits?
Can I develop a compiler that accepts a program in this language and produces an optimized Verilog module?
Over the course of a year, I determined the answer to both of these questions to be yes, and the result was a programming language called Gemini.
Along the way, I demonstrated with my prototype that novel type-theoretic concepts could be translated into concrete implementation. These concepts included a type system with multiple kinds, dependent types, multi-phase compilation, and "time as a type".
My research earned me the Charles Ernest Seager Memorial Award for most outstanding undergraduate research, and the inaugural spot in Drew's Hall of Fame.
An extended version of my thesis can be found online, and my thesis presentation can be found here.
Further, one can also look at developer-oriented documentation and the source code for more detail.