Section 2.4.2. Historical background

The Standard Molecular Data (SMD) format was initially developed by a group of European pharmaceutical companies in the mid-1980s. Draft documents were made available from 1987 and the specification was published (Bebak et al., 1989). A meeting in Frankfurt in 1988 established a series of technical working groups under the auspices of the Chemical Structure Association (CSA) to examine the format specifications in detail and to make recommendations for any revision. As a result, a draft form of a revised format, described as SMD Version 5.0, was published in February 1990 (Barnard, 1990). A document describing the core format, i.e. those data items regarded as essential in any exchange file, was prepared by one of us (JMB) for consideration by Subcommittee E49.51 of the American Society for Testing and Materials (ASTM).

In December 1993, the ASTM subcommittee E49.51 approved a standard specification for the content (i.e. recommended data items) of computerized chemical structural files (ASTM, 1994), although the subcommittee did not publish any proposals for a format specification. Later, the Chemical Abstracts Service (CAS) circulated a draft proposal for a connection-table-based exchange format for chemical substances and queries. It used some ideas that are similar to the 1990 SMD proposal and is expressed within the framework of the Abstract Syntax Notation 1 (ISO, 2002a,b). MDL Information Systems Inc. has also published a description of their proprietary formats (Dalby et al., 1992) and a number of other software systems now provide interfaces to these formats.

During this period, the IUCr Working Party on Crystallographic Information had commissioned one of us (SRH) to coordinate the development of a universal file to replace the existing fixed-format Standard Crystallographic File Structure (SCFS: Brown, 1988). As documented in Chapter 1.1 , the CIF approach was adopted as the international standard in 1990 and published by Hall et al. (1991). Although the small-molecule CIF is able to store a representation of 2D chemical topology, its data definitions do not meet all the needs of the chemical community. In 1991, the IUCr became interested in further extending CIF into the chemical arena and discussions took place between representatives of the CIF project and of the SMD Technical Working Group. These meetings decided that an integration of the SMD format and the STAR syntax was desirable because it provided a number of advantages over the existing SMD specifications (Barnard & Cook, 1992). In particular, SMD/STAR provides for a clearer separation of the data structure and the data content roles, together with more flexible data extensibility in future versions. In addition, automated data validation of STAR/SMD files is possible using electronic data dictionaries. In a wider context, there were obvious opportunities for integrating with other applications of the STAR File.