Implementation of a network database using a function graph language

Abstract: Most of the work on databases to date has involved approaches that deviate considerably from a purely mathematical (functional) viewpoint. The research reported in this thesis deals with an implementation of a network database using FGL to achieve this functionality. The amount of database copying resulting from functional updating, is the decisive factor in choosing the data structure of a network set. This mode of updating involves copying (just logical copying of updating and minimal physical copying) the database or portions thereof. Multiple interlinked btrees are found to be an advantageous data structure in our implementation, as the amount of copying resulting is acceptable in comparison to the principal alternatives, which involve physical or logical cycles. A manner of retrieving similar to that of CODASYL is pursued. However, the update primitives used put less restrictions on the extensibility of the update primitives used put less restrictions on the extensibility of the database schema in comparison with the standard CODASYL approach.