No comprehensive theory of development is available yet. Traditionally, we regard the development of animals as a sequence of changes through which an adult multicellular animal is produced, starting from a single cell which is usually a fertilized egg, through increasingly complex stages. However, many phenomena that would not qualify as developmental according to these criteria would nevertheless qualify as developmental in that they imply nontrivial (e.g., non degenerative) changes of form, and/or substantial changes in gene expression. A broad, comparative (...) approach is badly needed. In the Cnidaria, for example, even the boundary between generations is problematic. Describing their life cycle in terms of metagenesis (alternation between polyp generation and medusa generation) or in terms of metamorphosis (polyp as larva or juvenile) are matters of semantics more than biology. The life cycle of other metazoans, described in textbooks in terms of larva-to-adult metamorphosis, is hardly different from a typical metagenetic life cycle of cnidarians. This applies to holometabolous insects and to marine invertebrates like sea urchins, where most of the larval cells are discarded at metamorphosis. The uncertain temporal and spatial boundaries of individual development are also shown by the widespread lack of a strict correspondence between adult and mature. A comprehensive theory of development should start with a zero principle of “developmental inertia,” corresponding to an indeterminate local self-perpetuation of cell-level dynamics. Indeterminate growth, scale-invariance, segmentation, and regeneration provide examples of developmental dynamics close to that. (shrink)