SIGNIFICANT EVENTS:
TESTBED
DeLIver is distributed campus-wide
The final phase of the design and implementation of the Web client (DeLIver) has been completed and campus-wide distribution commenced October 15. DeLIver was tested by select user groups to insure the usability of the interface and the performance of the system. We are actively promoting the system and soliciting responses from users. http://dli.grainger.uiuc.edu/deliver.htm
Access to Ovid from DeLIver
We are also in the process of testing an Ovid proxy system which provides links to full text articles in DeLIver from Ovid bibliographic databases. This will provide the advantage of providing access to over 4 million citations from 4,000 journal and conference titles. Items available in full-text format from the testbed can be retrieved and displayed from an Ovid citation short record.
Testbed (DeLIver) Status:
As of this month, we have approximately 33,000 articles from five of our publishing partners processed and indexed and we are adding about 2,000 articles per month to the system. For a current list of what's in the system, investigate: http://morrigan.grainger.uiuc.edu/pubs/
EVALUATION
During the past quarter, the Social Science Team has been involved in the iterative design and testing of DeLIver; new analyses of data collected earlier in the project; the design and implementation of user registration, authentication, and transaction logging procedures for DeLIver; the development of a summative user survey instrument; ongoing research related to work practices in the digital era; and the redesign of the Teams web pages.
Team members re-worked the registration form that all public users are required to fill out before using DeLIVer. Because people expect the web to operate on an open and, largely, anonymous basis, crafting a user registration form that would collect demographic data as non-intrusively and accurately as possible has become a delicate balancing act. Team members also worked with the Testbed Team to develop user authentication procedures for DeLIver and to set up analysis and reporting procedures for DeLIver transaction logs.
The Social Science Team implemented an online usability survey meant to provide user feedback during the trial rollout period from August to October. When the user pushed the submit survey button, email with the results of the survey was sent directly to the team member responsible for analysis. The email message held an attached MIME file with the survey results formatted for direct import into a database. Only several DeLIver users submitted surveys. We do not know how many people aside from DLI project members used DeLIver during this trial period, so we have no way of estimating a survey response rate. Nonetheless, we feel that we will need to be more inventive and careful if we expect to get good compliance with optional user surveys from our anonymous user base in the future.
The Social Science Team conducted usability tests of DeLIver, and participated in a radical redesign of the systems interface based on these results. The usability test represented a new approach -- subjects were instructed to use DeLIver on their own, in the course of their work over a two-week period. They kept a log of their impressions the first time they used the system, and participated in a group interview to discuss and demonstrate system strengths and weaknesses.
In addition, two researchers have begun their own user studies as independent affiliates of the DLI Social Science Team. Michael Twidale, a professor in the Graduate School of Library and Information Science and an acknowledged leader in the study of collaboration in online system use, has initiated a study that draws students from courses in which DeLIver has been demonstrated. Students are invited to sign up for individual help sessions with Twidale, during which they try using the system to retrieve literature for some current assignment. These usability sessions reveal basic problems with DeLIver, which are forwarded to DeLIver designers, but they will also support Twidale's own development of new interfaces that include visualizations of search. Najmuddin Shaik, a PhD student in Educational Psychology who previously has worked on evaluations of the Museum Educational Site Licensing (MESL) project, has begun designing a series of usability tests for the DLI Iodyne client. The aim of this study is to do a close task analysis of the search process in Iodyne in order to produce a new mock interface that explicitly represents typical search paths in the system.
Team members developed preliminary drafts of user survey instruments that will provide summative data on the extent and nature of testbed use, the level of user satisfaction with the system, and reasons for use and nonuse. The complexities of developing a reliable survey instrument for a system that is in flux and designing survey delivery mechanisms that will encourage response are proving challenging.
Finally, Social Science Team members continue work on their studies of work practices and the changing nature of digital information infrastructure. These include research on the use of document components, the organization of office workspaces, the manner in which people deal with new computer-based information systems, and information convergence.
RESEARCH
Interspace
We have begun work on the development of automatic techniques for validating concept spaces and category maps. The preliminary version of the technical report of our research can be found at: http://www.canis.uiuc.edu/~zelenko/comparison/. At a recent DARPA IC&V/IM Principal Investigators meeting, we also established a working group to develop techniques and metrics of evaluation for information retrieval systems (dlib-metrics@cnri.reston.va.us).
Seminar Series Initiated: Information Retrieval in Digital Libraries: The Technologies of Scalable Semantics. Our goal for the seminar series is to bridge the gap between digital library applications and CS&E technology by presenting several information retrieval problems and 'brainstorming' how Computational Science & Engineering from the Computer Science department can help solve problems in information retrieval - a classic case of "Computing the Future". http://dli.grainger.uiuc.edu/canis/seminar.htm
Geoscience Tool evaluation:
User testing has begun on the Visual Thesaurus. Several different types of test have been designed to determine if computer image partitioning is comparable to that done by a human subject.
User analysis is also being performed on a Multi-dimensional Scaling analysis, and 1-D, 2-D, and 3-D (VRML) SOM displays for a collection of group systems comments.
Geoscience collections:
Vegetation and Temperature Query: AVHRR data is now available for California. The new system allows users to select temperature and vegetation patterns of interest and see them displayed on a map of California.
Textual Knowledge Source: The recently developed Textual Knowledge Source allows users to simultaneously search Petroleum Abstracts, GeoRef Thesaurus, and the GeoRef and GeoAbs collections. User may see output for these concept spaces and selected terms from each to redefine their search.
Tile Knowledge Source: The Tile Knowledge Source allows the user to access an aerial photo by selecting a pattern of interest from a visual thesaurus.
The Knowledge Source Manager: The Knowledge Source Manager integrates the above techniques and allows the user to search all three applications simultaneously and uses co-occurrence and similarity weights from all three to weight and rank the output.
IODyne:
Development of IODyne continues for Windows95/NT platforms. A Java version is still on hold until Java can demonstrate a sufficient level of maturity in terms of performance and user interface tools such as drag and drop. In the meantime, continued development in Visual Basic will still provide a basis for improved design that can be ported to Java when the appropriate time comes.
Development of IODyne is now focused on improving its core functionality in 4 principal areas.
1. Extending the record display language (RDL) and making it useful in more display tasks than at present. In particular, more sophisticated tagging is needed to make the subject navigator and other term suggestion tools more useful, especially in the concurrent display of multiple thesauri and classification systems. RDL must also handle conditional display of fields and conform more closely to XML syntax.
2. Reimplementing offline processing of thesauri and classification systems to generate hierarchy records that conform to XML syntax and break up large hierarchies into smaller subtrees for faster transport across networks and faster rendering by the IODyne client. This requires that parts of the IODyne subject navigator be reimplemented to process the new format.
3. Making the IODyne client multithreaded.
4. Developing a scheme for distributed configuration of repositories.
Related to further development of IODyne is development of a new information retrieval protocol (IRTP) to work with IODyne as well as other information retrieval clients. It uses some of the insights made during development of IODyne and is intended to be an easy-to-implement alternative to Z39.50, using XML-based encoding and modular, hierarchical attribute sets.
PUBLICATIONS
"NCSA Claims Parallel Code Breakthrough" High Performance Computing And Communications Week, Volume 6, Number 28, July 28, 1997.
H. Chen, T. R. Smith, M. L. Larsgaard, L. L. Hill, and M. Ramsey. "A Geographic Knowledge Representation System (GKRS) for Multimedia Geospatial Retrieval and Analysis. International Journal of Digital Libraries," 1997, forthcoming.
D. Roussinov and H. Chen. "A Scalable Self-organizing Map Algorithm for Textual classification: A Neural Network approach to Thesaurus Generation," accepted for CCAI Communication Cognition and Artificial Intelligence, Spring 1998.
M. Ramsey, H. Chen, B. Zhu, and B. Schatz. "A Collection of Visual Thesauri for Browsing Large Collections of Geographic Images", JASIS, Perspectives Issue on Visual Information Retrieval Interfaces, forthcoming.
PRESENTATIONS
Bishop, Ann. "DLI Social Science Research," Fall Program for the Chicago chapter of the American Society for Information Science, University of Illinois at Urbana-Champaign, Sept. 27, 1997.
Chen, Hsinchun. "Technology Showcase", hosted by the Artificial Intelligence Lab at the University of Arizona. Attendance was estimated at over 150 people. The "Showcase" consisted of demonstrations and poster sessions highlighting the groups most recent accomplishments in information retrieval for several different domains. September 15, 1997, Tuscon, AZ.
Dr. Hsinchun Chen and T. Dorbin Ng attended the DARPA ITO Principal Investigators Meeting in San Diego, California October 15-17. They demonstrated the latest techniques developed by the AI Group to integrate the retrieval aerial photo images, concepts from geoscience text sources, and AVHRR data using the Knowledge Source Manager.
Gross, Ben. . "Overview of UIUC DLI Project," Fall Program for the Chicago chapter of the American Society for Information Science, University of Illinois at Urbana-Champaign, Sept. 27, 1997.
Johnson, Eric. "IODyne: Dynamic, Object-Oriented Information Retrieval", Technologies of Scalable Semantics seminar series, Urbana, IL, October 7, 1997.
Mischo, Wm. "DLI Testbed", UIUC Library Faculty, Urbana, IL, October 17, 1997.
Pape, Daniel. "Categorization and Visualization of Information Spaces with a Self-Organizing Map", Technologies of Scalable Semantics seminar series, Urbana, IL, October 14, 1997.
Pottenger, Wm. "Computational Approaches To Semantic Retrieval: An In-Depth Look at Concept Spaces", Technologies of Scalable Semantics seminar series, Urbana, IL, September 23, 1997.
Schatz, Bruce. "Information Retrieval in Digital Libraries", Technologies of Scalable Semantics seminar series, Urbana, IL, September 9, 1997.
GRANTS AWARDED
Ann Bishop received the Centennial Scholar award from the Graduate School of Library and Information Science at the University of Illinois. The award provides $4000 for continuing her study of document disaggregation and the use of journal components.
VISITORS
August 5, 1997
Martha Richardson, Associate Head, SOLINET
August 12, 1997
Donald Bitzer, Department of Computer Science, North Carolina State
August 22
Bob Michaelson, Head, Science and Engineering Library, Northwestern University
August 15, 1997
Dr. David Nichols, Lancaster University
September 5, 1997
Dr. Lucy Lee, Professor, Department of Library and Information Science,
National Taiwan University
Philip Tsai, Head, Library System Office, Loyola University
From Saint Louis University:
Dr. Frances Benham, University Librarian
John Waide, Archivist/Associate Professor, Pius Library
Georgia Baugh, Science Librarian, Pius Library
Doris Beeson, Circulation Librarian, Pius Library
Shirin Jamasb-Doares, Reference Librarian, Pius Library
Win Shih, Reference Librarian, Pius Library
September 19, 1997
Richard Jonson, AT&T
September 26, 1997
Professor Edward A. Fox, Ph.D, Department of Computer Science, Virginia
Tech, Blacksburg, VA
October 3, 1997
Susan Gravenhorst, UIUC Board of Trustees
October 6, 1997
Chris Hagar, Liaison Librarian, The Robinson Library, University of Newcastle,
UK
October 20, 1997
Rob Bouzon, Hewlett Packard UIUC campus manager
October 24, 1997
Sun-Mi Kim, Researcher, Information Systems Division, Technical Information
Center, Agency for Defense Development, Korea
Mi-Ra Kim, Librarian, Information Systems Division, Technical Information
Center, Agency for Defense Development, Korea
October 28, 1997
Hsiu-ju Wu, Cheu-sheng Memorial Library, Tamkang University
Chiu-hsia Shih, Cheu-sheng Memorial Library, Tamkang University
Junfang Zhang, Bejing University, Beijing
Ed Hueckel, Elsevier
Appendix A: UIUC Digital Libraries Initiative Partner Update, November 3, 1997
Appendix B: DeLIver press release for NCSA (http://www.ncsa.uiuc.edu/Indices/Spotlight/Press_DeLIver.html)