Following is a brief summary of activities and progress of the University of Michigan Digital Library ( NSF -UMDL) project from February, 1996 through May, 1996.
The organization of the project has continued to develop, allowing investigators and team members to form more "natural" links, which have fostered better communication and task/goal coordination. The Advanced User Interface group (see below) was formed as a result of this process.
In addition, the project team participated in the recent annual site visit on May 6. Both principal investigators and graduate students participated heavily in this review which showcased the depth of the UMDL project. The semi-annual all project meeting is being hosted by the UMDL on May 16 and 17. This meeting includes a special pre-conference on User Needs and Evaluation. Information for site visit, pre-conference and the all-project meeting are available on the recently re-designed UMDL web site at http://www.si.umich.edu/UMDL/HomePage.html.
During the past quarter, the architectural group has done the following:
An updated specification of the UMDL protocols was written. This specification defines the legal protocols and messages for all agents participating in UMDL. This specification clarified how error conditions are to be handled (e.g., how the "sorry" KQML performative is interpretted), and eliminated several redundant protocols.
A draft specification for introducing security into UMDL was drawn. These specifications describe security issues particular to digital libraries in general, and UMDL (e.g., authenticating users, protecting against rogue agents) in particular. In addition, the specification suggests both protocols for UMDL agents that achieve specific security objectives and a secure "substrate" to add to the architecture. The security specification will receive considerable effort this summer.
Informal plans were made for incorporating auction agents into the UMDL architecture. Changes to the testbed to improve its performance and reliability forced significant changes on many of the agents. In addition, several new agents were added to the UMDL testbed. We are pleased to report that these agents were added without impacting the rest of the testbed, or forcing changes to the architecture. Thus, we successfully passed our first, albeit very minor, scaling test.
The newly launched Advanced User Interface (AUI) group has as its charter to design and begin prototype implementation of a visually rich paradigm for the extended information gathering and structuring task. In the past quarter we have been undertaking two parallel activities, User Centered Design (UCD) and Technology Exploration.
UCD is design method that focuses on real users -- identifying what functionality they really need, and how they can effectively access that functionality. So far we have completed the User Group Description and first Task Analysis stage. We conducted about two dozen in depth interviews of people in their current information work situations, identifying their information needs and how they currently go about satisfying them. The user groups studied include researchers, university professors, physicians, consultants, editors, scientists, consumers, administrators, students. The tasks studied include humanities teaching, humanities research, science research, policy creation, editing and writing, communication, exploration, business development, everyday tasks. We then conducted lengthy in depth discussions of the individual task interviews, trying to understand the dimensions of the task space, and of the corresponding procedures that people use in the various regions of the task space. We have developed a preliminary model of the suite of tasks, trying to capture both the common activities, and the variations, and using representations based on both a procedural description and a general model of the anatomy of the general information world and how the various procedures move information around within that world. Our next phase will be to try to design new procedures and computer support for them that will be more effective.
The Technology Explorations involved developing more local design expertise on relevant technologies. This has included getting some students familiar with PAD++, an infinite zoomable worksurface that we plan to use as a substrate for our explorations, and educating students on information visualization in general. Exploration have also included, as part of a related course, having students build some "Micro Prototypes" - small student projects meant to inspire later design, e.g., search+browse applications, graphical query histories, tools for early-structuring, and shared workspaces.
Michigan's collection testbed continues to grow, reaching over 150,000 journal articles from our publisher partners. In addition, significant effort has been put toward collaboration with reference publishers. In particular, Grolier's Americana Encyclopedia has just been released, following extensive co-development activity (Groliers provided raw SGML used for print production which Michigan structured for electronic publication). Similar to McGraw Hill's Encyclopedia of Science and Technology, our multimedia Americana implementation represents a unique publication not commercially available. Groliers has been structured to provide differential representation of the content based on the size of the retrieval-- useful functionality given the length (over 100 pages) of a large number of the 40,000 articles.
Locally developed curricular units have been added to the testbed, providing topic-specific inquiry units and a mechanism to explore "driving questions" for the inquiry process of our high school user community. Librarians with subject expertise in the testbed domain have evaluated hundreds of web resources and identified, described, and registered over 200 collections of relevant resources within the Conspectus.
Current attention is focused on numeric and spatial data, with the University Library supporting licenses of resources such as Michigan land use and school district data, world mapping information, and city GIS data.
Solicitations with new publishers continue, with Cambridge University Press having recently provided verbal confirmation of their intent to contribute. Negotiations are also underway with several other scholarly and curricular publishers.
The Collection Search & Retrieval Working Group has optimized Z39.50 Agents so that they take less time and use less memory. In fact, processing speed is now almost twice as fast as it was during the first demonstration in November 1995. We also separated most of the interoperability agent code from the Z39.50 Agent code. This gave us the opportunity to use most of the interoperability code for a separate FTL Agent.
Working Group members have incorporated the Broad System of Ordering ( BSO ) as an agent for use by the UIA and Task Planner. Because of the need to have agents for linking various specific subject vocabularies, e.g., NASA Thesaurus and Sears Subject Headings, we created one Thesaurus Agent to accommodate many different vocabularies.
Since digital-library users are likely to enter terms that are more specific than BSO terms, the UMDL needs to link their terminology with the more general BSO terms that encompasses their topics of interest. To accomplish this, the Working Group plans to link several specific, in-depth subject vocabularies to BSO . Working Group members have begun a literature review on linking multiple thesauri in information retrieval systems and a survey of available subject vocabulary files. The Group intends to start with the "Sears Subject Headings" which is a subject vocabulary licensed by the H. W. Wilson Company and used in many school libraries throughout the country.
The Collection Search & Retrieval Working Group has built prototype agents in C++ for storing and indexing UMDL-registered web books. (Web books are HTML pages that constitute a web presentation on a particular theme, idea, person, organization, etc.). The Web Book Creator Agent sends a completed list of URLs that comprise a web book to the Pat Agent via the UMDL's ILU communication interface. The Pat Agent then fetches the web book pages over the Internet and applies full-text indexing to them. The Pat Search Agent searches UMDL-registered web books and retrieves web books' URLs. Search results are flowed into a HTML page for display with a web browser like Netscape. Current prototype agents have been applied to a single web book. Working Group members plan to apply indexing and searching across all UMDL-registered web books in the coming months.
During the past quarter the Education Group engaged in the following activities:
Design of the Inquiry Interface: Students who will use the UMDL will be engaging in inquiry -- searching and re-searching to explore a driving question that they created, e.g., what does it feel like to be in an earthquake? why can't you see a black hole? The current search engines are built for one-shot queries, however; they do not support sustained, repeated searches, nor do they support the synthesis of the information that is found into some coherent argument. The Inquiry Interface does both: it helps to scaffold repeated searches (e.g., by allowing a student to easily see past searches, by employing thesauri to help suggest other search terms) and it helps to integrate the information found (e.g., each driving question has it's own workspace where found items are stored and organized.
Collection of data in 3 high schools: We are working in three different high schools (traditional, focused science; interdisciplinary science, English & social studies; and project-based) to see the different issues involved in using on-line materials. We are collecting a range data on these experiences, e.g., process information from groups of students (i.e., recording keystroke interactions plus an audio track of a group's discussion plus a screen capture), classroom notes (e.g., what instructional strategies are used by the teacher and their impact). Once analyzed these data should give us a better picture of how on-line materials are actually being used in science instruction.
Development of curriculum modules: The use of on-line materials is novel; there are no published guidelines of how they should be used in classrooms. By and large all the curriculum on the network currently is meant for teachers to use; the material itself is not meant to be used by students. In contrast, we are developing curriculum units that use on-line material quite explicitly. In particular, we have developed the Inquiry Wheel model for using on-line materials, e.g., students ask a question, plan a search, analyze the information found on the network, etc. The Inquiry Wheel serves to structure the activities as the students use the on-line materials. We have designed and classroom curriculum units for geology (e.g., volcanoes & earthquakes), as well as astronomy (e.g., black holes).
This next quarter we plan to analyze the data collected and revise the curriculum materials based on findings from the former activity. In addition, we will modifying the curriculum units to take advantage of the new Inquiry Interface that will be available for students this fall. In addition, we will be filling out the UMDL's collection of educational on-line materials.
The IPE group has implemented a market facilitator (auction) agent that determines prices for information goods and services using UMDL protocols. A simple scenario where agents bid for notification and search services was demonstrated at the May site visit and DLI conference.
In the Testbed Construction group, our main emphasis this quarter has been to add considerable new user functionality to UMDL by creating a totally new user interface. This Java-based interface has been designed based on analysis of user needs to be more than a simple query system. More explicitly than other searching tools, the new interface is designed to allow users to organize their thoughts and queries in a manner centered around their research interests. This is accomplished by introducing the concept of a user workspace where the user can formulate queries, search UMDL, view items retrieved by the search, organize and share these items, and keep on-line notes.
To support this interface we have greatly expanded the capabilities of the user interface agent (UIA). We have set up a Java API to access the UIA functionality and added multithreading to the UIA so that collection searches are performed in parallel, greatly improving performance. We have provided support for user workspaces by adding persistence to the items returned by the collection search and retrieval protocol. This allows results to be saved for later use. New data structures were implemented to organize those items in ways that are meaningful to the user. Recently, we have started adding facilities for browsing general thesaurus agents, such as the Broad System of Ordering ( BSO ) and NASA Thesaurus Agents, to aid in the query formulation process.
Our work on other agents has included integration of redesigned Thesaurus and BSO agents into the testbed. We greatly improved our access to our locally owned collections by developing fully integrated collection interface agents (CIAs) for our collections which use the FTL search engine. These CIAs allow FTL collections to be searched and documents returned using the UMDL collection search and retrieval protocol. We are currently using these CIAs to access collections of earth and space science journals and magazines from Elsevier and UMI.
Meetings and Visitors
Below is a brief list of some of the meetings and visitors with which NSF -UMDL personnel have been involved during the past three months.
January 1996
9 Dr. Helen Shen, Hong Kong University of Science and Technology
Dan Atkins, Bill Birmingham, Randy Frank, Wendy Lougee, Greg Peters
9 Bob Rossbach, WTVS Detroit (television station)
Wendy Lougee, Elliot Soloway
10-11 Defense Information Enterprise Technology (DIET) Program PI meeting
Sunnyvale, CA
Bill Birmingham
16-19 Association for Library and Information Science Education Annual Conference
San Antonio, Texas
Dan Atkins, Olivia Frost, Laurie Crum
19-25 ALA Midwinter
San Antonio, Texas
Dan Atkins, Olivia Frost, Laurie Crum
29-2/2 Patrick Finn, La Plaza de Taos
Semnar on the Le Plaza de Taos (community network)
Dan Atkins, Joan Durrance
February 1996
MAME (Michigan Association of Media in Education)
Grand Rapids, Michigan
Keynote Presentation: Elliot Soloway
The Role of Digital Libraries in Supporting Inquiry Learning
1 Carlene Ellis, Intel
Dan Atkins
5 Roland Dietz, Senior VP, Elsevier Science
Wendy Lougee, John Price-Wilkin, Dave Rodgers
9 Ontario Library Association
Toronto, Canada
Speaker: Dan Atkins
16 Carol Moore, Director University of Toronto Library
Wendy Lougee, John Price-Wilkin
2 Sang-Wan Han, Sung-Hyuk Kim, Sung-Bin Moon
Korea
Dan Atkins
29 Academic Press Library Advisory Group
Orlando, FL
Wendy Lougee
March 1996
3-8 OMG (Object Management Group)
David Richardson
5 Committee on Institutional Cooperation, Chief Information
Officers and Library Directors
Wendy Lougee, Randy Frank, John Price-Wilkin
4-5 ERCIM meeting
Nice, France
Bill Birmingham, Kathy Willis
Presentation: Bill Birmingham
Rank Xerox
Grenoble, France
Bill Birmingham, Kathy Willis
5-7 Netscape Developers Conference
Greg Peters
8 George Shipman, University of Oregon
Dan Atkins
15 School of Information is officially launched
15 MACUL (Michigan Association of Computer Uses in Learning)
Lansing, Michigan
Keynote Presentation: Elliot Soloway
The Role of Digital Libraries in Supporting Inquiry Learning
Speaker: Raven Wallace
18 National Digital Library Federation, Washington, DC
Lougee
19 Research "Science Fair"
Capitol Hill, Washington
Raven Wallace
19 Ken Perlin, Director, Center for Advanced Technology in Digital Multmedia
Media Research Laboratory, New York University
AI Seminar: Two research Projects: PAD and Improvisational Animation
20 Jerry Butters, A T & T
Wendy Lougee, Randy Frank
21-22 7th Annual IPoSCE Review
The University of Michigan
Ann Arbor, MI
Speakers: Jose Vidal (Building and using agent models in an agent-based digital
library); Tracy Mullen (Exchangng goods and services in a digital library); Ansoara
Nica (Resource discovery in large-scale distributed library systems)
26-4/5 Joerg Haake, Anja Haake
GMD, Darmstat
UMDL team
Seminar (4/4/96): Hypermedia-based Collaboration Support
31-4/1 OCLC Authority Control Conference
Dublin, Ohio,
Speaker: Karen Drabenstott
"Beyond Online Catalogs: The Role of Authority Control in Digital Libraries"
April 1996
1 Leif Hansen, Copenhagen Business School
Wendy Lougee, John Price-Wilkin
1 Chichi Tanaka, Takanori Hayashi, Hatsuhito Mtsuhashi
Japanese Ministry of Agriculture, Forestry and Fisheries
Dan Kiskis, Wendy Lougee, John Price-Wilkin
2 IBM Steering Committee
Speaker: Dan Atkins
14-15 Amy Friedlander, Editor, D-Lib Magazine
UMDL Team
14-18 CHI '96: Common Ground
Vancouver, British Columbia, Canada
George Furnas
14-18 NAB '96
Las Vegas, NV
Bill Birmingham
22 Sue Corbett, Blackwell Publishers
Wendy Lougee, John Price-Wilkin
25 W3C Workshop on High Quality Printing from the Web
Spencer Thomas
May 1996
2 Lu Kabir, Oracle
Wendy Lougee, Randy Frank
5 Tom Kalil, White House
Dan Atkins, Mike Wellman, Wendy Lougee, Randy Frank
6 NSF / ARPA/ NASA Site Vist
Ann Arbor, MI
UMDL Team
8 Maria Luisa Arenas-Franco, Pantificia Universidad Catolica de Chile
Wendy Lougee
13-16 7th Joint European Networking Conferencee
Budapest, Hungary
Bill Birmingham
16-17 All Project Digital Library Initiative Workshop
University of Michigan
Ann Arbor, MI
UMDL Team
Comments or questions may be sent to: UMDL.INFO@umich.edu