At Lambda World, Robert M. Avram discusses effect models, as he exemplifies their joyful and unfortunate nature.
The very essence of functional programming relies on a fairly simple premise: write side-effectless programs so that you can reason about them equationally. Side-effects however are predominant in every program that . . . well, does something useful. In light of this, the modelling of effects in functional programs has become a fairly active domain of study. From stacked effects, to horizontal ones; from transformers to free monads and so on. Every effect model, though joyfully they may seem, have some rather unfortunate consequences.
This talk was given by Robert M. Avram at Lambda World.