In 1991, we published a paper in Development that proposed the ABC model of flower development, an early contribution to the genetic analysis of development in plants. In this, we used a series of homeotic mutants, and double and triple mutants, to establish a predictive model of organ specification in developing flowers. This model has served as the basis for much subsequent work, especially towards understanding seed plant evolution. Here, we discuss several aspects of this story, that could be a much longer one. One surprising conclusion is that materials and methods that might have led to similar work, and to the same model, were available 100 years before our experiments, belying the belief that progress in biology necessarily comes from improvements in methods, rather than in concepts.