I’ll start out by taking a brief tour that covers my past experience on various projects and how those experiences shape my current view of the importance and practicality of end-to-end system verification. My basic thesis is that end-to-end system verification is theoretically/technically possible, but often *economically* impractical. More importantly the marketplace for verified software systems is likely to significantly expand within this century.
I’ll will also discuss on going research in developing simpler and flexible semantic techniques for the verification of Foundational Proof-Carrying code systems. In particular, I will discuss how to build proofs of type soundness using higher-order abstract encodings of syntax without the need for meta-logical approaches to soundness.