The database is not a traditional relational database management system, but rather an advanced file management system. Clusters, but not stars, form the basic unit of the database and the structure has been designed to provide a natural working environment. The whole data set for a cluster forms a special relational set, because one key, the star designation, is common to all files.
The database structure uses the directory hierarchy supported by the Unix system. The main directory is the database itself. It contains several sub-directories: description of the database, help information, references, bibliography, programs, shell scripts. The clusters are collected in parent directories according to the source catalogues (NGC, IC or anon). Each cluster defines an independent directory identified by its name and containing the available data in distinct files, one for each data type. This structure allows easy inclusion of any new data type.
The present database structure is thought of as a first step in the organisation of cluster data and bibliography. When enough data analysis has been performed and only one set of data for each type will be available for each star, it may perhaps be more convenient to adopt another structure and collect all the data in one file for each cluster.