This paper extends the discussion on computational approaches towards modelling emotion antecedent appraisal processes, based on Scherer (1993a) and Chwelos and Oatley (1994). In the first part we briefly discuss the different goals that motivate such modelling as well as potential pitfalls. We then take up some of the critiques of GENESE expressed by Chwelos and Oatley, and examine the proposed alternatives. The paper conclude with an outline of perspectives for future work in this domain.