A model, schema, and interface for metadata file systems

Abstract

[Abstract]: Modern computer systems are based on the traditional hierarchical file system model, but typically contain large numbers of files with complex interrelationships. This traditional model is not capable of meeting the needs of current computer system users, who need to be able to store and retrieve files based on flexible criteria. A metadata file system can associate an extensive and rich set of data with a file, thus enabling more effective file organisation and retrieval than traditional file systems. In this paper we review a wide range of existing proposals to add metadata to files and make that metadata available for searching. We then propose a hierarchy of definitions for metadata file systems based on the reviewed prototypes. We introduce a data model for a database-oriented pure mdfss complete with operations and semantics. The model supports user-initiated instance and schema updates and file searches based on structured queries. We also explore the design space of a set of user interface operations intended to implement the pure model and facilitate the capturing of rich metadata. We argue that without such a simple method for users to create rich metadata, progress in this field will remain limited.