Sustainability = separation: keeping database structure, domain structure and interface independent

Ian Johnson


The conventional approach to relational databases implements the knowledge domain structure directly as database structure; tables represent entities and specific relationships between them, forms represent views which depend on specific joins, workflows and interfaces are customised through programming. This approach is inherently unsustainable due to dependence on specific software/versions and essential knowledge embedded in the custom programming rather than the database.

A number of archaeological databases, from IDEA in the 1990s through to FAIMS and Heurist today, use a fixed database structure across all applications, and represent the knowledge domain structure as data within the database. They generally incorporate documentation of the knowledge domain structure directly within the database.

In this poster I demonstrate the advantages of such an approach for long-term sustainability, since the only requirement for data accessibility is a database which can manage a set of SQL tables, against which a set of standard SQL queries can be issued to retrieve the knowledge domain structure and content. This requirement should be satisfied as long as SQL servers exist which, despite fierce competition from other models, show no signs of fading away.

The poster will also argue that the original user-interface - so often confused with the database itself, due in part to desktop software where the interface is tightly integrated with the data storage infrastructure - is merely an end-user convenience and should be regarded as transitory; its value can be sustained with simple documentation such as an explanation of workflows accompanied by screenshots.