The paradigm for embedding an epidemic model within a probabilistic programming language should be sorted out. All conditional probability issues will have been addressed, and we will have a precise way of expressing all the information needed to construct a posterior predictive distribution from a given model and associated event data.