The book "Models for Discrete Longitudinal Data" offers a comprehensive analysis of modeling approaches for non-Gaussian repeated measurements that may be incomplete. The authors, Geert Verbeke and Geert Molenberghs, start with models for the complete marginal distribution of the outcome vector, which allows for model fitting based on maximum likelihood principles. This immediately implies inferential tools for all parameters in the models. Additionally, less complex alternatives such as generalized estimating equations and pseudo-likelihood methods are presented. The book also addresses conditional models and the family of random effects, including the beta-binomial model and the generalized linear mixed model. Various procedures for model fitting are discussed, and the differences between marginal models and models with random effects are highlighted.