# WG211/M22Kammar

From WG 2.11

Modern partial evaluation takes advantage of semantic invariants and properties to partially evaluate stuck open code fragments. In this talk, I will argue that the motivation for modern algebra is similar: studying actions on an unknown state on their own. We will implement a simple partial evaluator for a stateful, straight-line programming language, and draw connections to its associated monoid action. As a consequence, the free extension of this associated action is the data-structure needed for partially evaluating programs using semantic equivalences. Time permitting, I will outline our research programme in utilising modern algebra techniques to the development of modular modern partial evaluation libraries.