Rand Corporation

Stephanie Cammarata

Darrell Shane

Prasad Ram

This report describes a research project whose objective is to improve the interaction between a database user and a relational database management system. It discusses the "data flow" gap between (1) external databases acquired from public agencies and (2) datasets derived from these databases to be used as input to specific simulation models. This gap is the lack of interoperability between external databases and simulation databases and the transformations that are necessary to derive compatible datasets. It also presents the computer software system the authors developed to bridge this data flow gap--the Intelligent Information Dictionary (IID) software system, which serves as a semantic-based interface between a database user and a relational database management system. IID extends the traditional roles of a data dictionary by enabling a user to view, manipulate, and verify semantic aspects of relational data. IID operates as a domain-independent kernel augmented with domain-specific knowledge bases. IID represents and maintains these knowledge bases as semantic metadata necessary for correcting external databases and deriving required abstractions and aggregations from these databases

An Intelligent Information Dictionary for Managing Semantic Metadata

Iris M. Kameny

Corinne Replogle

This report describes the RAND Metadata Management System (RMMS), a system that manages metadata--definitional & descriptional information about databases, simulation models, & procedures--for relational databases such as those maintained in INGRES. Many of these databases have little documentation or other descriptional information to accompany them, making it difficult for users to understand the definitions, abbreviations, acronyms, & descriptions of the data elements stored & maintained in a database management system. The authors developed the RMMS with five major goals in mind: (1) provide complete, thorough, & standard database documentation; (2) record & manage information about different versions of each database; (3) maintain a history of the changes made to database tables, schema, or data values; (4) facilitate deriving databases for input to simulation models & for sharing among models; & (5) standardize the names of data elements that are conceptually the same but named differently or are named the same but are conceptually different. This report should be of help to users of other relational databases who would like to develop a similar metadata repository for their own set of databases.

Stephanie Cammarata

Books

Your next favorite book is out there