UBC Theses and Dissertations
A comprehensive investigation of data dictionaries Diersch, Thomas Karl Julius
This thesis reviews data dictionary systems, approaching the subject from four directions. First, the contents of dictionaries are examined. Second, the users and their interaction and use of the data dictionary are reviewed. Third, some popular commercial data dictionary packages are reviewed. Finally, data dictionaries are examined as a package to be implemented by a data processing department. Data dictionary entities can be grouped into four classes, each adding to the information stored by the preceeding layer. They are, the data dictionary which stores data entity data, data directories, which store processing entity data and process-data relationships, data resource dictionaries, which store data about processing environment entities such as hardware, and the metadata dictionary which stores data about conceptual entities such as events and functions. Using this structure, each class of dictionary is examined to identify the entities and their attributes. Users are examined in the following groups: the database administrator, systems analysts, programmers, the data processing operations department, and the user group. Five commercially available dictionaries are examined in detail, DATAMANAGER, DB/DC Data Dictionary, Data Catalogue 2, UCC-10, and IDD. In addition, the thesis examines two dictionaries in overview, highlighting their advanced thinking. ICL DDS, which allows the user to model both implementation and conceptual views of data, and CINCOM's DCS which includes the dictionary in a package containing a screen design aid, a programmer workstation, and security system all integrated with the dictionary. The data dictionary is examined from a system application view point addressing the questions to which any application is subjected when evaluated. Identifying the need for a data dictionary, three areas are examined. First the control of data as an organizational resource, second the use of system development methodologies and third, the management of change in data processing systems. The ability of the data dictionary to support a data processing environment which uses conventional file structures, a simple database or a complex environment with multiple databases, or distributed processing is addressed. Software selection criteria are examined, defining the dictionary content, and utilities which should be investigated, and criteria for environment such as data processing hardware and software, user-vendor relationships and cost and maintenance are proposed. Finally we propose some events in the data processing life cycle which may assist in the successful implementation of a data dictionary. Implementation in a complex environment where no other project proposed is critical, or before a major database management system conversion may be warranted. Or implementation may be undertaken before a large systems project, or as part of a data processing standards project.
Item Citations and Data