The Digital Reference Electronic Warehouse Project: Creating the Infrastructure for Digital Reference Research through a Multidisciplinary Knowledge Base

Modeling the Complex Digital Reference Landscape

One area of research stemming from the use of complexity theory is modeling the digital reference transactions within DREW as a complex adaptive system. Once the digital reference transactions have been cleaned, an inductive system of clustering can be utilized to examine the self-organizing nature of digital reference knowledge bases. Each transaction will be modeled as an autonomous agent with a set of attributes (the proposed DREW element set). Some of the attributes are static (such as the text of the transaction), but some are dynamic (such as the time since the transaction was closed, or the number of times the agent is referred to by other transactions). By placing these transactions in an n-dimensional space (two or three dimensions for visualizing the space for example), pair-wise comparisons between the agents can be conducted (in essence determining how similar any two agents are). Agents will move “closer” or “farther” apart based upon these comparisons. It is anticipated that these agents will inductively cluster. It is also hypothesized that these clusters will change over time, as not only the dynamic attributes change (a transaction ages for example), but the agents themselves change (new questions or new references added).

Creating an Infrastructure for Virtual Collaboration

One of the exciting possibilities of a DREW schema is that it empowers the infrastructure to allow for virtual collaboration between researchers and practitioners. Services will start by providing records for DREW. Researchers will then use these records to develop tools across different services. These researchers will then be encouraged to prepare their models and tools using the DREW schema so that the services participating in DREW can apply these research results to their own services. Practitioners can immediately benefit from research and will be encouraged to not only continue their involvement in DREW but also to improve their management of the digital reference service. Researchers can then test the difference these new tools and models make on reference service, and the cycle continues.

This model is currently in use in the open-source community. As infrastructure and data schema are created, programmers use this information to develop tools. As tools are created and released, other programmers improve on the code, and the result is that the users have a much better experience. This virtual collaboration will allow digital reference to rapidly improve as a service.

Conclusion: The DREW Research Agenda

The process of creating this digital reference archive introduces a set of questions that power a research agenda. Each of these questions stems from a challenge (or opportunity) in the process of creating, implementing, and using this warehouse of digital reference transactions. Some of these issues have been previously addressed in this paper.

• What would an archival schema for digital reference transactions look like? Will one schema work for all communication mechanisms used by digital reference? What minimum subset of these fields is needed to be useful?
• What tools are needed to extract these fields from digital reference transactions? How complete of an archival record can be automatically recreated from a chat or e-mail reference transaction?
• Is there a thesaurus that would be useful in linking subject lists from different services? Can this assignment of subject headings be done inductively?
• What subset of fields will maintain the information needed for research and discovery while still protecting the privacy of patrons? What policies are needed to balance keeping a data-based history of the service with the need to protect personal information of patrons?
• How can the information space within DREW be explored through bibliomining and visualization tools? What patterns can be discovered about the process of answering questions? Can the changing space of reference transactions be demonstrated through animated visualizations?
• What is the life of a reference transaction? Are there facets that can be used to predict how long a question will be useful in a question archive? What indicators can be used to detect questions that have outdated information?
• How can digital reference be rapidly improved through the virtual collaboration of researchers and practitioners? What management tools are most effective in helping digital reference services improve? What measurable differences do these tools make?

Through reference authoring via human intermediation, libraries have the ability to produce large amounts of high-quality information. To understand this information and create tools that allow for the rapid creation of knowledge bases, as well as advance our conceptual understanding of the changing face of reference, researchers need a cleaned collection of transactions from a wide variety of services. The DREW project will supply researchers with this data source, as well as making it possible for participating services to quickly benefit from the results of the research.

Scott Nicholson is an Assistant Professor at Syracuse University’s School of Information Studies. R. David Lankes is Director of the Information Institute of Syracuse and an Associate Professor at Syracuse University’s School of Information Studies. Submitted for review May 2, 2005; accepted for publication July 28, 2005.

References

1. R. David Lankes, “The Digital Reference Research Agenda,” Journal of the American Society for Information Science and Technology 55, no. 4 (2004): 301.
2. Karen Markey Drabenstott, “Classification to the Rescue–Handling the Problems of Too Many and Too Few Retrievals,” Knowledge Organization and Change: Proceedings of the Fourth International ISKO Conference, 15-18 Jul. 1996, Washington D.C., USA (Frankfurt: Indeks Verlag, 1996): 107-18.
3. Chris Sherman and Gary Price, The Invisible Web: Uncovering Information Sources Search Engines Can’t See (Medford, N.J.: CyberAge Bks., 2001).
4. Michael K. Bergman, “The Deep Web: Surfacing Hidden Value,” Journal of Electronic Publishing 7, no. 1 (2001), (accessed Dec. 1, 2006).
5. Danny Sullivan, “Intro to Search Engine Optimization,” Search Engine Watch (accessed Dec. 1, 2006).
6. “What Is a Knowledge Base (KB)?” (accessed Dec. 1, 2006).
7. P. Rumbaugh, personal communication, Jul. 6, 2004 and Jan. 9, 2007.
8. M. Mitchell Waldrop, Complexity: The Emerging Science at the Edge of Order and Chaos (New York: Touchstone, 1992); John H. Holland, Hidden Order: How Adaptation Builds Complexity (New York: Addison-Wesley, 1995); R. David Lankes, Building and Maintaining Internet Information Services: K-12 Digital Reference Services (Syracuse, N.Y.: ERIC Clearinghouse on Information and Technology, ED 427778, 1998).
9. Holland, Hidden Order .
10. Library of Congress, NetRef: NISO Committee AZ: Networked Reference Services, 2004, www.loc.gov/standards/netref (accessed Dec. 1, 2006).
11. Joseph Janes, “Question Negotiation in an Electronic Age,” in The Digital Reference Research Agenda, eds. R. David Lankes, Scott Nicholson, and Abby Goodrum (Chicago: ACRL, 2003): 48-60.
12. Rumbaugh, personal communication.
13. Ibid.
14. Library of Congress, NetRef.
15. Marcia Lei Zeng and Lois Mai Chan, “Trends and Issues in Establishing Interoperability among Knowledge Organization Systems,” Journal of the American Society for Information Science and Technology 55, no. 5 (Mar. 2004): 377-95.
16. Dennis Nicholson, Ali Shir, and Emma McCulloch, HILT: High-Level Thesaurus Project Phase II, 2004, available as an E-print (accessed Dec. 1, 2006); National Institutes of Health Fact Sheet UMLS Metathesaurus (accessed Dec. 1, 2006).
17. ALA, The USA PATRIOT Act (accessed Dec. 1, 2006).
18. Scott Nicholson, “Avoiding the Great Data-Wipe of Ought-Three,” American Libraries 34, no. 9 (Oct. 2003): 36.
19. Workgroup for Electronic Data Interchange. De-identification and Limited Data Set White Paper, 2003.
20. Nicholson, “Avoiding the Great Data-Wipe of Ought-Three.”

Pages: 1 2 3 4 5 6