Structural genomics promises to deliver experimentally determined three-dimensional structures for many thousands of protein domains. These domains will be carefully selected, so that the methods of fold assignment and comparative protein structure modeling will result in useful models for most other protein sequences. The impact on biology will be dramatic.