Genesis of the Crystallographic Information File

Abstract

Efficient exchange of information and data within and across scientific disciplines is essential for progress in science. The International Union of Crystallography (IUCr) has tracked and encouraged efforts to design a formal description and exchange mechanism for crystallographic data. The result of this involvement has been the Crystallographic Information File standard described in this volume. Early approaches to the management of crystallographic data relied upon disk file formats derived from the input/output conventions of Fortran and associated 80-column punched-card images. Although such formats could be efficient and compact, they were inflexible and difficult to extend to embrace new types of data. The growth of computer networking made it important to design portable and extensible methods for exchanging data between different computer architectures and programming languages. After initial experimentation with an extensible pseudo-free format mechanism still modelled on card-image formats, the IUCr commissioned a lightweight free-format mechanism based on the STAR File. This is a plain-text file format where data values are tagged with character strings identifying their role. By extension, standard tags may be assigned, and the attributes of the data they describe may be recorded in separate reference compilations. This chapter traces the historical development of this approach: the crystallography community's choice of a restricted subset of STAR File functionality, its subsequent extension to describe large binary data sets, and, most importantly, the development of dictionary definition languages that allowed the description of data attributes using the same formalism as used for the data files themselves. Since 1991, several specialized collections of data attributes have been created for use in different fields of crystallography, and similar systems exist in a small number of related fields. This chapter is also available as HTML from the International Tables Online site hosted by the IUCr.