Fear and desire in systems design : negotiating database usefulness

Abstract

Databases are ubiquitous. They are used for a host of functions including coordination, decision making, and memory archiving. Despite their importance and ubiquity, databases continue to frustrate us, often departing from the goals originally set for them. If databases are such essential ingredients for organizations, what diminishes their usefulness? Besides the nature of the data that is entered into the database, usefulness is also shaped by the fields, features, and functionalities that the database designers originally construct that then shape the kind of data that can be entered into the system. This dissertation examines the process of database design and the assumptions and concerns adopted by the stakeholders involved in it. I conducted a year long ethnographic study of a university that has been engaged in creating a self-sustaining Environment Health and Safety system to manage research related hazards and to ensure regulatory compliance. The integrated database system was envisioned as a tool that would allow the university to observe and improve compliance practices while keeping records that would be available for self-auditing and government inspection. My research observations suggest that actors imagine diverse purposes that the database, when complete, should serve. These purposes - entailing the three themes of accountability, efficiency and comparability - appear to guide the design process. As these imagined purposes gain momentum, they translate into both desires and fears for the features of the database. For instance, when efficiency is imagined as a purpose, it creates a desire for features such as drop-down menus that are easy enter information into. The inclusion of such features, however, creates a fear of oversimplification.

(cont.) Through a negotiated process, features such as text boxes are added to address the fears. Yet, every design change negotiated within the database system creates ripple effects with regard to other purposes, generating the need for still further changes. The process of database design becomes highly dynamic and the final database system is a negotiated compromise between multiple trade-offs over time. By juxtaposing these fears and desires, and through the use of causal-flow models, I articulate the process by which databases depart from their original goals.