A modification of the Structure Alignment Program (SAP), combined with a novel automatic method for the definition of structural elements, correctly identified the core folds of a variety of small β/α proteins when compared with a series of ideal architectures. This approach opens the possibility of not just determining whether one structure is like another, but given a range of ideal forms, determining what the protein is. Preliminary studies have shown it to work equally well on the all dα-class and the all-β class of protein, each of which have corresponding ideal forms. Given the speed of the algorithm, it will be possible to compare all of these against the Protein Structure Database and determine the extent to which the current ideal forms can account for the variety of protein structure. Analysis of the remainder should provide a base for the development of further forms.