The Crystallographic Information File (CIF) [3] format has become a standard interchange format for the description, archiving, and publication of crystallographic experiments on small organic and inorganic molecules. The remarkable feature of the CIF interchange approach is the use of an electronic dictionary to provide a detailed description of each item of data in a CIF. The task of developing a dictionary of terminology for macromolecular crystallography and structure has been underway for several years [4] and is in near final form [5, 6].
The syntax of mmCIF is derived from the Self-defining Text Archive and Retrieval (STAR) format proposed by Hall [7]. mmCIF data files employ a simple grammar in which each data value or list of data values is accompanied by a data item name. Each item of data has a detailed definition, and these definitions are collected in the mmCIF dictionary. The prescription for specifying each mmCIF data definition is separately defined in another dictionary. The latter dictionary defines a Dictionary Description Language (DDL) which provides the framework for constructing the mmCIF dictionary definitions. Both of these dictionaries are expressed using the simple mmCIF syntax. Figure 1 shows an example of a fragment of a mmCIF data file, a mmCIF dictionary entry, and a DDL dictionary entry illustrating the uniform mode of expression used in each of the cases.
Figure 1:
Abbreviated examples of CIF data specifications, CIF dictionary definitions,
and DDL definitions.
The uniform appearance of mmCIF data files and mmCIF dictionaries as collections of name and value pairs is dictated by STAR syntax rules; however, the underlying organization of the data and definitions is determined by the DDL. The DDL used in developing the mmCIF dictionary [8, 5] is an extension of the DDL proposed by Cook and Hall [9, 10] for the core dictionary of crystallographic definitions. These DDL extensions were designed primarily to make the content of the mmCIF definitions more accessible to software that might use this information for data integrity validation.