Data management of musical information
Brad Rubenstein, Michael R Stonebraker · University of California, Berkeley eBooks · 1987
This dissertation explores various issues related to the application of computer data management techniques to musical information. The contribution of this work is twofold: (1) It extends an existing data model (the entity-relationship model) to support a database schema for musical information. (2) It develops particular data management access methods to effectively manipulate information in the musical database. Various types of musical information are analyzed, focusing particularly on the conceptual representations necessary to formally encode musical scores expressed in Common Musical Notation (CMN). The entity-relationship model is taken as a starting point for a CMN database schema. A new feature is added to the entity-relationship model to represent the notion of ordered sets of entities. This property is called hierarchical ordering. Such a relationship occurs, for example, when an ordered set of notes constitutes a chord, or an ordered set of measures forms a movement of a composition. A method is proposed for representing attribute inheritance among entities, and various approaches to the problem of managing this inheritance among entities within the music database are considered. Several examples demonstrate that inheritance under hierarchical ordering is more complex than that supported by standard generalization hierarchies. Using these two data modeling tools, hierarchical ordering and attribute inheritance, a schema is developed for CMN. Entities from the CMN schema are divided into groups according to various aspects of the information: temporal, timbral, and graphical. To implement hierarchical ordering using existing relational database technology, extensions to relational access methods are developed. Entity ordering is supported by the introduction of ordered relations. A data structure is developed for ordered relations which provides a general solution for the support of user-defined aggregate functions computed over ordered sets, and user-defined orderings, including hierarchical orderings. The dissertation closes with a summary of questions left open by the present research, particularly those remaining issues that need be addressed in order to develop a functional database system appropriate to the management of musical information.