Veit Elser (Cornell University)
Abstract. SFI is my spring break destination from the frozen northland where I occasionally teach the introductory physics course about waves. A standard part of the curriculum deals with computing a diffraction pattern, given a pair of slits of a given size and separation — but not the reverse, i.e. directly computing the structure of the slits from the diffraction pattern. This colloquium begins with a tutorial on solving the reverse problem and puts this in the context of current efforts to image molecules with free-electron x-ray lasers. The principle of the algorithm that solves the x-ray "phase problem," as it is usually called, is remarkably general and has proven to be a powerful strategy for solving a broad range of constraint problems. A sampling of these, from proteins to packings, are described at the end of the talk.