A short history of<i>SHELX</i>

<jats:p>An account is given of the development of the<jats:italic>SHELX</jats:italic>system of computer programs from<jats:italic>SHELX</jats:italic>-76 to the present day. In addition to identifying useful innovations that have come into general use through their implementation in<jats:italic>SHELX</jats:italic>, a critical analysis is presented of the less-successful features, missed opportunities and desirable improvements for future releases of the software. An attempt is made to understand how a program originally designed for photographic intensity data, punched cards and computers over 10000 times slower than an average modern personal computer has managed to survive for so long.<jats:italic>SHELXL</jats:italic>is the most widely used program for small-molecule refinement and<jats:italic>SHELXS</jats:italic>and<jats:italic>SHELXD</jats:italic>are often employed for structure solution despite the availability of objectively superior programs.<jats:italic>SHELXL</jats:italic>also finds a niche for the refinement of macromolecules against high-resolution or twinned data;<jats:italic>SHELXPRO</jats:italic>acts as an interface for macromolecular applications.<jats:italic>SHELXC</jats:italic>,<jats:italic>SHELXD</jats:italic>and<jats:italic>SHELXE</jats:italic>are proving useful for the experimental phasing of macromolecules, especially because they are fast and robust and so are often employed in pipelines for<jats:italic>high-throughput</jats:italic>phasing. This paper could serve as a general literature citation when one or more of the open-source<jats:italic>SHELX</jats:italic>programs (and the Bruker AXS version<jats:italic>SHELXTL</jats:italic>) are employed in the course of a crystal-structure determination.</jats:p>