Abstract
George Sheldrick made his reputation by providing excellent software to solve the structures of a range of crystals which diffracted well and had fewer than 200 atoms. The techniques used included Patterson searches, direct-methods phase determination and least-squares refinement of the final model. By the 1990s it was possible to collect atomic resolution data from macromolecules, so he adapted the least-squares refinement program to handle the large numbers of parameters required for these cases. Experimental phases for large molecules could be calculated based on the anomalous signal from heavy atoms such as sulfur. The structure-solution software was adapted to find the positions of the anomalous scatterers using the small anomalous differences, and then extended to first generate and refine the phases, and then interpret the resultant maps.