The cradle now demonstrates a usable and coherent fraction of the semantic validation goals, specifically applied to tensors and expressions on them. Many things are not yet supported, and it's still not possible to get any concrete computation out of the system. (It's all semantic analysis still.) But this is showing promise.