Sources of Change for Modern Knowledge Organization Systems

📝 Abstract

Knowledge Organization Systems (e.g. taxonomies and ontologies) continue to contribute benefits in the design of information systems by providing a shared conceptual underpinning for developers, users, and automated systems. However, the standard mechanisms for the management of KOSs changes are inadequate for systems built on top of thousands of data sources or with the involvement of hundreds of individuals. In this work, we review standard sources of change for KOSs (e.g. institutional shifts; standards cycles; cultural and political; distribution, etc) and then proceed to catalog new sources of change for KOSs ranging from massively cooperative development to always-on automated extraction systems. Finally, we reflect on what this means for the design and management of KOSs.

💡 Analysis

Knowledge Organization Systems (e.g. taxonomies and ontologies) continue to contribute benefits in the design of information systems by providing a shared conceptual underpinning for developers, users, and automated systems. However, the standard mechanisms for the management of KOSs changes are inadequate for systems built on top of thousands of data sources or with the involvement of hundreds of individuals. In this work, we review standard sources of change for KOSs (e.g. institutional shifts; standards cycles; cultural and political; distribution, etc) and then proceed to catalog new sources of change for KOSs ranging from massively cooperative development to always-on automated extraction systems. Finally, we reflect on what this means for the design and management of KOSs.

📄 Content

Sources of Change for Modern Knowledge Organization Systems

Michael Lauruhn and Paul Groth {m.lauruhn, p.groth}@elsevier.com Elsevier Labs 1600 John F. Kennedy Boulevard, Suite 1800, Philadelphia, PA

Abstract Knowledge Organization Systems (e.g. taxonomies and ontologies) continue to contribute benefits in the design of information systems by providing a shared conceptual underpinning for developers, users, and automated systems. However, the standard mechanisms for the management of KOSs changes are inadequate for systems built on top of thousands of data sources or with the involvement of hundreds of individuals. In this work, we review standard sources of change for KOSs (e.g. institutional shifts; standards cycles; cultural and political; distribution, etc) and then proceed to catalog new sources of change for KOSs ranging from massively cooperative development to always-on automated extraction systems. Finally, we reflect on what this means for the design and management of KOSs.

1. Introduction Knowledge Organization Systems (KOSs) such as ontologies, terminologies, data dictionaries, and classification schemes provide the foundation for a variety of applications. These applications range from classification of objects, indexing processes, and traditional information retrieval (IR) systems to semantic web applications, question answering, and rule-based systems. While the core goal of many KOSs is to resolve entities and concepts for the applications they serve, newer functionality includes reasoning and discovery. Traditionally, KOSs have depended on manual processes that were largely akin to an editorial process based entirely on human supervision, recommendation, and decisions. These were typically done by, or with input and influence from, subject matter experts or collection experts. As KOSs creation and maintenance processes evolve, one significant change is the level of human input that can effect change.

We note how sources of change differ as KOSs evolve across various models over time. We begin by looking at issues related to evaluating and correcting bias in early KOSs (Library of Congress Subject Headings and Dewey Decimal Classification) that were designed for library cataloging. These models had a wide breadth in scope and regularly scheduled distribution processes, initially exclusively in print. These earlier KOS rely heavily on full-time editors and dedicated subject matters and the development and dissemination process can be equated to an editorial system the produces editions on a regular schedule. Over time, KOSs with scopes that are more niche became more prevalent. In this paper we will examine a pair that are created for indexing biomedical literature and databases. The editorial processes are still based on human decision making, but begin to come from a more distributed group of experts who are both users of and contributors to the model. Likewise, the sources of change also begin to become more application focused. As the contributor model becomes more distributed, the release of updates and change evolve into an ongoing process with more frequent versions that start to resemble a software release cycle.

In this work, our aim is to provide a catalog of sources of change that designers and users of KOSs need to be aware of. Figure 1 depicts the nine source of change we call out in this paper and how they map into the major constitutes that impact the maintenance and creation of KOSs.

Figure 1: Sources of changed mapped to the major actors and entities within KOS creation and maintenance processes.

To substantiate this list, we begin by a review of the existing literature. We then proceed to catalog new sources of change with appropriate exemplars. This is followed by a discussion about the implications all these sources of change have on KOS design and management.

2. Sources of KOS Change: The Existing View 2.1 Change as an editorial process One of the major drivers behind change in KOSs, particularly ones that are largely-distributed across disciplines, is a desire to remain up to date in terms of cultural sensitivities and perceived bias. In his 1971 book, Prejudices and Antipathies: A Tract on the LC Subject Heads Concerning People, Sanford Berman outlines many of the subject areas in which the Library of Congress Subject Headings (LCSH) are insensitive and out of touch with the society that libraries operated in (Berman 1971). These include areas such as race and ethnic groups, gender roles in society, politics, sexuality, and others. In the 1971 introduction, Berman cites previous literature which identifies that earlier justifications for the bias that was prevalent in the LCSH. The author then encourages others to join him in seeking to “remedy long- standing mistakes and to gain for the profession a genuine, earned respect among people who read and think.”

In his subsequent 1993 Preface

This content is AI-processed based on ArXiv data.