The Holy Grail of the formal semantics community is a semantic description formalism that is readable, writeable, usable, powerful, and has a rich theory. Two recent approaches (modular monadic semantics [LHJ95, LH95, Esp95] and action semantics [Mos92]), both coming out of denotational semantics, try to modularise a semantic specification in order to enhance its readability and modifiability. These approaches are quite different, and yet have some intriguing similarities. In this paper, I detail the two approaches and compare them.
As it appears that the modular monadic approach is more general than the action semantic approach, I then propose an investigation into the representation or implementation of action semantics in the modular monadic semantic framework.
In Section 2, I consider the modular monadic semantics based on lifting used by Liang, Hudak and Jones [LHJ95]. In Section 3, I consider another more general modular monadic semantics based on stratification and invented by Espinosa [Esp95].
In Section 4, I consider Mosses' alternate approach, action semantics [Mos92]. Then in Section 5, I compare this approach with the modular monadic one, and consider how the former might be represented in the latter. Some of the difficulties involved are considered in more detail in Section 6; and I conclude my thesis proposal in Section 7.