The challenges of protein engineering arise, in part, from the enormous number of possible sequences and the almost unimaginably small fraction of such sequences that can be studied experimentally or computationally. Fortunately, not all possibilities need to be considered because many different sequences can adopt the same structure. Of the vast number of sequences that fold into a given conformation, some are 'simpler' than the sequences of typical proteins. Studying protein sequences that are simpler helps focus attention on the principal determinants of structure. Recent examples of this strategy are the simplification of protein surfaces and cores, the use of a binary 'code' for protein design and the structural analysis of random simple sequences.