Executive Summary *

Acknowledgments *

1. Description and Scope of the Workshop *

2. The Workshop *

2.1 Steering Committee *

2.2 Meeting Objectives and Pre-Meeting Preparation *

2.3 Meeting Participants *

2.4 Meeting Format *

2.4.1 General Format *

2.4.2 Presentations and Software Demonstrations *

2.4.3 Follow-Up Activities *

3. Themes Covered in the Small Groups *

3.1 Search, Browse, and Retrieval Session *

3.1.1 Setting the Issues *

3.1.2 Discussion in Small Groups and Plenary Session: Search, Browse, and Retrieval *

3.1.3 ADL Summary of Comments and Response: Search, Browse, and Retrieval *

3.2 Content and Processing Session *

3.2.1 Setting the Issues *

3.2.2 Discussion in Small Groups and Plenary Session: Content and Processing *

3.2.3 ADL Summary of Comments and Response: Content and Processing *

3.3 Interface and Navigation Session *

3.3.1 Setting the Issues *

3.3.2 Discussion in Small Groups and Plenary Session: Interface and Navigation *

3.3.3 ADL Summary of Comments and Response: Interface and Navigation *

4. Initial Summary by ADL Staff (Summing Up Session at the Workshop) *

4.1 Discussion *

5. Summary of Evaluation Comments on the Workshop *

6. Impact of the Design Review on ADL *

6.1 ADL Collection Development Strategy *

6.2 ADL User Profile Strategy *

6.3 ADL Human Support Strategy for Digital Library Users *

6.4 ADL Interface Design *

7. Future Plans the Design Review Panel *

Appendix A. Attendees *

Appendix B. Meeting Agenda *

 

This report is a summary of the Second Alexandria Design Review Panel. The workshop was held in Santa Barbara, California, 19-21 February 1997 to review user requirements for the Alexandria Digital Library and to discuss plans for testbed development and research progress in the coming eighteen months. The challenge to build an operational digital library testbed is improved by the research accomplishments of the ADL team. It is important to recognize that some research questions cannot be addressed in the absence of an operational testbed. This balance of emphasis will continue to be a concern in the coming eighteen months.

Thirty-two representatives from the public sector, private sector, and academia came from U.S. (one from Canada) for two-and-a-half days to iterate between plenary and small group discussion, software demonstrations, and informal discourse. The intention is that panel members will continue to be proactive in the coming year, through online activities (working with the Library testbed), through personal contact with ADL project staff, and by face-to-face meetings at national conferences in the coming year. The next Alexandria Design Review Workshop will be held approximately one year hence.

In preparation for the meeting, the Steering Committee agreed upon a set of objectives to guide the content of the agenda. A three-part focus was established to deal in turn with issues of search, browse, and retrieval; content and processing; and interface and navigation. These three areas encompass a number of the more difficult problems to be resolved immediately. Their selection was based primarily upon feedback gained in focus groups, survey questionnaires, and videotaped or recorded interactions with actual and potential digital library patrons over the first two years of the Alexandria Project. Many specific topics were discussed including ways for users to view the contents of the library (an overview of the holdings), giving the user a "road map" of the services of the library, visualization (presentation) of search results, iterative search processes, and customization of the user interface.

From the wide range of comments and recommendations contributed by the Panelists that are detailed in this report, some strong themes can be identified. First, the Panel is supportive of the efforts of the research teams working at Santa Barbara and at Colorado. Panelists and funding agency representatives agreed that keeping a careful balance between developing an operational digital library and contributing to theoretical and applied research is necessary. The importance of the content of the library was strongly supported, as were the enabling technologies for distributed ingesting of data with attention to the representation of data provenance and verification. Collection development criteria were discussed at length, with emphasis on focusing on particular geographic areas and collecting all forms of georeferenced information about that place. Panelists recommended that ADL develop enabling technologies and specifications for connecting local collections to ADL, with distributed searching capability.

In connection with ADL user interface design, the implementation of user profiles was discussed with ideas for multiple profiles per person and for the way in which the profiles could influence the interface presentation and conduct of the search. Support for a method of providing "something with a heartbeat" to be available to assist in the use of ADL was expressed in several ways, with the related recommendation that ADL aim to create a professional tool for information specialists as well as for other user communities. A discussion of measures of success for ADL focused on grounding those measures in the user's work environment and information need.  

The overall thread that runs through the recommendations is that ADL needs to focus better on what can be accomplished in the short time left in the current phase of the project. Most panelists reported that the format of the Workshop worked well, but they would like for it to have been more focused and they expect to see specific responses from ADL on the recommendations made.

The co-leaders of the Workshop wish to thank all the Design Review Panelists for their energy and enthusiasm in working with the Alexandria team members for this meeting. Steering Committee members gave generous amounts of time and thoughtful input. Those of you who served as small group leaders and presenters helped to make the meeting a success. We also extend a hearty thanks to the rapporteurs, including graduate students from the University of Colorado, for their hard work and patience in recording the discussion and to Mary-Anna Rae for videotaping the proceedings. For preparations behind the scenes, we thank Patty Towne, Teddi Lenard, Mary-Anna Rae, Jason Simpson, Randy Kemp, and Kevin Lovett. We are especially grateful to Bryan Chan for stepping in for the final month as major domo and otherwise helping to make the meeting run smoothly. We also received administrative assistance from Lanell Lucius and Sandi Glendenning at the NCGIA office at UCSB, and from Marcia Signer at University of Colorado-Boulder. This report was written with contributions from Barbara Buttenfield, Linda Hill, Randy Kemp, Suzanne Larsen, Mary Larsgaard, and Mary-Anna Rae and review and comment from ADL staff and Design Review Panelists.

The Alexandria Design Review workshop and this report are contributions to the Alexandria Digital Library Project, jointly sponsored by NSF, NASA, and ARPA. We acknowledge support from a grant by the National Science Foundation (NSF IRI-94-11330). Matching funds from the University of California and the University of Colorado are also acknowledged. Travel support for a number of participants to attend the meeting was provided by U. S. Geological Survey, the U. S. Navy, Xerox Corporation, Digital Equipment Corporation, and Hughes Information Technology Systems.

1. Description and Scope of the Workshop

The design of any system or product engineered by a multi-disciplinary team is complex. It is too easy to become so closely involved with a product as it comes together as to miss both obvious flaws and subtle errors. Discussion within a closed team tends to circle back upon the same territory over and over. Team members find particular problems intriguing, particular challenges especially thorny, particular successes particularly gratifying. In short, the adage that those who build a system are least qualified to critique it holds true in this project as in so many others.

The goal of the Alexandria Design Review is to ensure an outside perspective is drawn into the design process and that objective and informed critique helps to refine the system at regular intervals during implementation of the ADL testbed. Without such critique, the efforts of the Alexandria team may not yield a system that meets user needs. Following the first Design Review workshop held in Leesburg, Virginia in September 1995, we felt that a change in the scope of the panel efforts would be beneficial. To that end, we invited panelists who will commit not just to one meeting, but to multi-year participation.

Our invitation cited the following goals:

• to collect comments and recommendations from advisors who can benefit from the Alexandria Project;
• to gather a group with needed expertise to critique ongoing design and development;
• to ensure wide diversity of opinion and disciplines;
• to guarantee a commitment from participants to carry through the final eighteen months of the project.

In comparison to other digital library projects, the Alexandria testbed is perhaps unique in the following respect. We integrate a broader type of data into the national information infrastructure, which remains largely text-based. We deal with kinds of information that agencies who provide information to the public need to organize. These kinds of information are not traditionally stored in libraries. A digital library should be more than a physical library in electronic form, and as we design a software testbed with functionalities that may not have been completely predicted prior to implementation, the need for external and objective critique becomes paramount to our success. This then forms the scope of our intentions for holding the Alexandria Design Review Workshop.

2. The Workshop

The Workshop provided an opportunity to bring together representatives from academia, federal agencies, and industry to review user requirements for the ADL testbed and to recommend priorities for research and development that can be met given current technology and current understanding of digital library technology, recognizing an eighteen month funding cycle.

2.1 Steering Committee

The workshop co-leaders (Linda Hill and Barbara Buttenfield) drafted a Steering Committee to assist in organizing the meeting agenda, drafting the open call for participation, initiating discussions during the meeting, leading small group discussions, and providing follow-up activities. The Steering Committee was formed in Summer, 1996.

Steering Committee members included:

Linda Hill, Alexandria Project, Workshop Co-Leader
Barbara Buttenfield, U. Colorado-Boulder, Workshop Co-Leader
Terry Smith, Alexandria Project, Director
Mike Goodchild, Alexandria Project, Associate Director
Larry Carver, Director of the Map and Imagery Library, UCSB
Jim Frew, Alexandria Project, Manager and Senior Engineer
Mary Larsgaard, Head of Cataloging for the Map and Imagery Library, UCSB
Jason Simpson, Webmaster for the Project
Mary-Anna Rae, Alexandria Project, Administrative Liaison. 

The Steering Committee invited two individuals to serve as Co-Convenors of the Workshop:

Joseph Boisse, Head of Libraries, UCSB; and
James Williams, Dean of Libraries, U. Colorado-Boulder.

2.2 Meeting Objectives and Pre-Meeting Preparation

In preparation for the meeting, the Steering Committee agreed upon a set of objectives to guide the content of the agenda. A three-part focus was established to deal in turn with issues of search, browse, and retrieval; content and processing; and interface navigation. These three areas encompass a number of the more difficult problems to be resolved immediately. Their selection was based primarily upon feedback gained in focus groups, survey questionnaires, and videotaped or recorded interactions with actual and potential digital library patrons over the first two years of the Alexandria Project.

Shortly before the workshop, invited participants were encouraged to register via the Internet and to submit short statements about their background and interest that were available to other participants to read. This was intended to introduce participants to each other before the workshop, and to elicit from the registrants responses to an extensive list of user requirements collected from ADL user needs analysis activities.

2.3 Meeting Participants

Thirty-two external participants and twenty Alexandria staff attended the meeting. Panel members came from a diversity of disciplines representing academia, federal agencies, and industry. A directory of Panel members and ADL staff participants is included in Appendix A.

2.4 Meeting Format

2.4.1 General Format

The general format of the two-and-a-half day meeting was a series of group discussions, alternating between plenary sessions and small group discussions focusing on a single theme, as follows:

• Presentations on the status of ADL, the user and interface evaluation studies that have taken place, and the gazetteer component of the system.
• Discussion topic #1: Search, Browse, and Retrieval Functions
• Setting the issues: Terry Smith
• Small group discussions
• Plenary discussion: Bruce Gritton 

• Discussion topic #2: Content and Processing
• Setting the issues: Mike Goodchild
• Small group discussions
• Plenary discussion: Prue Adler
• Discussion topic #3: Interface and Navigation
• Setting the issues: Robert Nideffer
• Small group discussions
• Plenary discussion: Jim Frew
• Summing Up

There were also demonstrations in MIL on Wednesday evening. Continental breakfast, lunch, and dinners were provided to the participants.

A complete meeting agenda appears in Appendix B.

Assignments to the small groups were done with the goal of preserving some uniformity within disciplines (e.g., librarians, educators, and data managers) and keeping the size to about six people per small group. Small group membership did not change throughout the meeting to ensure a depth of interaction within the groups (two groups were combined into one on the first day.

The meeting began with presentations from ADL staff on the status of the testbed and the gazetteer, and the results of user requirement studies and interface evaluation research.

Three themes were covered during the meeting. For each theme, a plenary session presented and discussed the issues, ending with discussion points or questions as a basis for discussion in the small groups. The questions were intended to provoke discussion and to guide the potential scope of the session. Each small group selected a spokesperson to report on the discussions that took place. After participants reconvened as a large group, spokespersons presented summaries of their small group discussions. Graduate student rapporteurs were assigned to each small group to record discussions and assist in compiling summaries. Subsequent sections of this report will detail the points raised on each theme in the small group summaries.

2.4.2 Presentations and Software Demonstrations

On the first afternoon, the participants visited the Map and Imagery Library at UCSB to view demonstrations of projects demonstrating various implementation issues, and to see presentations of a number of ongoing research projects. Pete Fohl demonstrated work on how to present fuzzy footprints in response to user catalog queries. B.S. Manjunath showed results of the research on texture and color-based image searching. Doug Ramsey presented his spatially annotated bibliography on the Mojave Desert. Jason Simpson ran a live Alexandria Web session.

2.4.3 Follow-Up Activities

Following the meeting, each participant was asked to fill out a brief exit survey including five questions. Participants could fill these out on paper and hand or mail them back, or submit them electronically, via an email template distributed a few days after the meeting. Those surveys continue to arrive as this report is drafted, and are providing valuable feedback about the management of the meeting. This feedback will be used in planning for the next Alexandria Design Review Workshop.

3. Themes Covered in the Small Groups

3.1 Search, Browse, and Retrieval Session

3.1.1 Setting the Issues 

Terry Smith made the presentation on search, browse and retrieval functionality to initiate this discussion. In terms of desired functionality in this area, he said that we need to help the user to:  

Information organization issues include the issues of item classes (texts, photographs, etc.), item organization (ADL, federated collections, personal collections), and meta-information models for collections, items, and services. Interface issues include a choice of metaphor: desktop, traditional library, landscapes, and lenses or telescopes. 

System design must also take into account the users. Different user groups have distinctive information needs and expectation and individual users have specific objectives for using the system; e.g., to browse with no firm goal in mind, to find general information about a topic, or to find specific information. 

Terry posed four discussion topics:

1. Representation/display of collections: issue of aggregation levels

2. Which types of searches are most important for ADL to support

3. Ideas for the representation and display of result sets

4. Support for iterative searching

3.1.2 Discussion in Small Groups and Plenary Session: Search, Browse, and Retrieval

Themes from small group discussions are listed below. They are paraphrased from the small group presentations.

** column contains the number of times that the topic was reported from the five small groups

Bruce Gritton asked the group at the end of the plenary session to call out the top recommendations for this topic area. Here is what came out at that time:

• Option to get to human support to the use of ADL through the interface - more than just leaving comments as users can do now.
• Show the user the status of the process they are going through and give them a way to track their progress.
• Provide a way to customize (personalize) the search and retrieval process.
• Support search refinement (narrowing) and search enlargement (broadening).
• Show where the data is located (if it is not directly accessible online).
• Categorize the result sets: default groupings/rankings of search results.
• Treat result sets as subsets of the database with functions to process the set - e.g., custom sortings.
• Provide thesaurus support to translate or cross reference among terms.
• Provide glossaries to explain/define terms and processes.
• Allow easy switching between searching and browsing.
• Support guided searches.
• User profiles would be useful but shouldn't intrusively brand a person in any static way. Useful for system defaults. Possibility of dynamically collecting profile information as well as prompting for certain data.
• ADL should be able to pass queries beyond the local collection to other libraries and data centers. User should have choice of searching locally or distributing the query to other sites.

There was also some discussion of validation in the sense that items in the library should be screened (selected) and adequately described. In contrast to totally open contributions.

Also a comment about evaluation: find a way to look at human performance and how well the system performs in helping the user do what they want to do.

3.1.3 ADL Summary of Comments and Response: Search, Browse, and Retrieval

Strongest support from the small groups and the plenary discussion was for the following:

• Human support available when using ADL
• Facilitation of iterative searching processes
• Make use of user profiles in various ways, including preferences for system settings, implicit query parameters, and linking relevancy ranking of result sets to user profiles.
• Categorize (group and/or rank) result sets in ways that are meaningful to users.

3.2 Content and Processing Session

3.2.1 Setting the Issues

Mike Goodchild gave a thought-provoking presentation on the criteria that could be used to select the items to put in the library and the possibilities for the manipulation of items in the digital library that are not possible in a traditional library. He proposed that selection should be driven by the search mechanism: it is impossible to browse through a large collection of items - if it can't be found through the search mechanism, it might as well not be in the library. The search mechanism of ADL is focused on the intersection of geographic footprints: the overlap of the user's target area with the footprints from the geographic representation of the items in the library catalog. This is a geoquery - it starts with a location, then searches and returns items with the same location. A geoquery may be subsequently refined by other criteria: level of geographic detail, time, and author/subject/title, for example. A geoquery can be generalized to any spatial frame: for example, the human body, an automobile, or a house. In each case, there are issues of fuzzy footprints both in queries and in the representation of items in the library and issues of degrees of matching between the target area and the item footprints. 

Geoitem can be similarly used to mean georeferenced items. Maps and remote sensing images are well-recognized in this group, but a georeferenced item could also be a piece of music, a photograph, or any "bag of bits" related to a geographic location. A geolibrary, such as ADL, is a library containing geoitems and using geoquery as its primary search mechanism. This approach is particularly well suited for

• Some items in a traditional library
• All items in a map and imagery library
• Some items in other kinds of special libraries
• Many items not traditionally stored in libraries for which there was no search mechanism 

Defining the content of ADL is hard: only limited analogs and metaphors exist - a geolibrary is only possible in a digital world. We need strategies for thinking about hard problems with no traditional metaphors or legacies. Negroponte suggested that we listen to children for ideas who are not constrained by the legacies of the old world. Or, we could search harder for metaphors. Thinking about things that are now technologically possible but impossible before is hard.

Consider two libraries: a geolibrary and a traditional library. What would be their comparative sizes and comparative levels of use? Would groups that have not used traditional libraries be users of the geolibrary? A speculation: a geolibrary favors information of local interest; favors multiple types of information where the geographic key may be the only attribute they have in common; and may be of optimal interest to local government and its citizenry. It would of most immediate use to organizations/activities whose interests also have footprints - for example, a resource management agency. 

Digital libraries can support the processing of items on-the-fly prior to retrieval and following retrieval. What do we mean by processing?

• The processing necessary in order to examine the contents: display content and browse, zoom and search within the content
• The processing necessary to integrate data: rubber-sheeting, conflation (visual fit), concatenation, schema merging
• GIS processing: overlay, buffering, simulation, simple modeling (through linked packages) 

A design consideration is the incorporation of distributed processing so that proprietary models can be used - sending the data to the model. 

ADL is a research project. Its value will ultimately be determined by the creation of new ideas and knowledge, expressed in the form of publications. It is not driven by consumer demand, but it would like to be popular. This suggests that our criteria for content should include

• Ways to demonstrate the unique, novel properties of ADL in a compelling way
• The concept of a geolibrary
• Integration of data based on a common geographic location
• Integration of the processing functions of a digital library
• Integration of distributed collections and services
• Use the limited resources available wisely (no immediate prospect of income from services)
• Provide a platform for the resolution of some basic research questions
• Avoid duplication of efforts 

These arguments suggest that we should include:

• Examples that incorporate processing
• Examples that show integration of data based on geographic location
• Examples that span the interests of target user groups
• Examples that illustrate the power of distributed collections
• Georeferenced items in preference to geographic items
• Many distinct information types (text, reports, photographs, images, maps, music, ...)
• Examples of current, compelling interest 

We should stay away from:

• Uniform coverage for its own sake
• Items whose cost of ingest in not justified
• Duplication of such efforts as EOSDIS, NGDC, GCDIS, NGDC, ... 

How to proceed? Small groups should identify compelling examples of content appropriate for ADL (consider ingest costs and implementation). Then, generalize from the examples to lists of content in the context of an operational library.  

We also need to consider experimental designs to measure the success of ADL in accordance with a definition of the operational mission of ADL (the site of choice for a particular type of information) and the research questions relating to becoming operational. How many ADLs should there be? One per major research university? Spatially clustered by demand? One depository per nation? Do these issues go away in a digital world? What can we say about this issue from the perspective of collection development?

 The main questions for discussion are

1. What is the set of selection criteria ADL should use for collection development?
2. What are the examples that illustrate these criteria?

3.2.2 Discussion in Small Groups and Plenary Session: Content and Processing

Themes from small group discussions are listed below. They are paraphrased from the small group presentations.

** column contains the number of times that the topic was reported from the five small groups

Other comments:

1. Panel members would like to know more about the ADL spatial search software. Is it portable? Is it interoperable? What is the status of the ADL distributed query capability?
2. Get focused on good rich deep catalog and don't get distracted by other things. Become a prototype for other nodes in a distributed system. 

 3.2.3 ADL Summary of Comments and Response: Content and Processing

The single most frequent comment was the need for local data, from collections outside UCSB - recognizing that no one place can or wants to be a physical Alexandria, but that ADL can be a virtual Alexandria for georeferenced information. These local collections, or "depth collections," should be multi-disciplinary/theme/temporal/format. It would be the responsibility of the relevant local agencies to develop these collections, creating metadata and digital data as needed; it would be ADL's responsibility to link to them, to provide an integrated path/navigation to these collections - distributed search as a primary function of ADL.  

There were several comments about who ADL's major user groups might be; some nominations were earth scientists, planners, historians, humanists, and architectural groups.

When it came to listing actual content that ADL should have, there were as one might expect many different opinions. The following were the favorites.

• ADL needs to select one geographic area (California, more narrowly southern California; national parks; major North American metropolitan area) and focus the UCSB data/metadata collection efforts on that. It was specifically stated that the focus area should be both for ingesting already-existing data and metadata, and for putting hardcopy data into digital form and creating metadata for, what is currently available only as hardcopy data, but this was implied.
• Another suggestion for concentrating focus was by theme - putting into digital form base maps such as medium-scale, uncopyrighted topographic-map series, which are used by many different user groups.
• The importance of gazetteer data was noted.
• Another suggestion was the need for ingesting already existing base-map-level data, such as DCW, DRG, DOQ, and TIGER. The Gutenberg collections were mentioned.
• Remote-sensing imagery (both satellite imagery and aerial photographs) historic datasets were considered to be important categories.
• There was a recognition that for ca. 90% of all spatial data, only hard copy, and therefore only metadata, would be available, and that this metadata needed to be added to (or if appropriate transparently searched) by Alexandria. It should be emphasized that this metadata needs to include where the user may obtain the data, either by going to a library or by purchase from a vendor. It is especially important that the Alexandria Catalog contain metadata for base-level spatial data, such as U.S. and foreign topographic series and remote-sensing imagery. 

There were what appeared to be some flat-out disagreements, with one group expressing the point that what would make ADL unique is for it to be a one-stop shop for finding spatial data. Another group stated that ADL should not aim for universal access to holdings. In actuality, there is probably not a disagreement, since the first group was thinking in terms of ADL being able to perform distributed search over many collections, and the latter was darkly contemplating the impossibility of solving the schema mapping problems that would enable comprehensive distributed search in the short time remaining for ADL Phase 1 funding. There is also the need to differentiate between ADL as a digital library for distributed georeferenced information and the ADL node at UCSB with its specific collection development policy. 

While all of the groups recognized user interest in processing of search results, the practicalities of time frame and staff available probably mean that until the end of 9/98, only the most basic processing will be possible, specifically in the areas of presentation of search results (e.g., subsetting of search results by format or type). Processing functions considered to be of most utility to the largest number of users are given below, with most important listed first: 

• compression
• conversion of file type to user's file type of choice
• rubber sheeting; overlay; conflation
• data manipulation
• editing, annotation, layout
• encoding "geographer's rules" in expert system to guide users in the most efficient and appropriate use of georeferenced information. 

Linking users to sites that give examples of post-processing, or that perform post-processing, was recommended as a reasonable, temporary alternative to having this processing available on Alexandria. One suggestion was to create an architectural model like Stanford' Infobus to make ADL both a provider and server of georeferenced documents - could be called the Geo-Infobus. 

3.3 Interface and Navigation Session 

3.3.1 Setting the Issues

Terry Smith introduced Robert Nideffer as the latest addition to the project, brought in to conduct the systematic redesign of the ADL web interface. Terry made a point of recognizing that the last interface posed many problems, and that it was designed as a prototype to identify potential functionality, developed with advice from many people. Robert's background in sociology as well as design may help significantly in pulling the many different voices together into a coherent vision for the new interface.

Robert presented a variety of interface ideas, all based on his own work. These included:

1. His Ph.D. dissertation presented on interactive CD-ROM and supplemented with video, sound, and graphics.
2. Terminals, the Cultural Production of Death (history of computing and artists' relationship to technology). This was a collaboration among several UC campuses that raised the technological issue of how to create a distributed environment.
3. SPEED: Technology, media, and society (online magazine of architecture and design that Robert co-founded and edits). This online journal is designed to be dynamic--people can add to it over time, and navigate through the environment in exploratory ways. The website uses VRML and JAVA and allows links to information clusters.
4. Life in the Universe (based on Stephen Hawkin's Brief History of Time and five articles from a special issue of Scientific American). This raised issues of how best to present dense, sophisticated scientific material to a broad audience. It resides on CD-ROM with links to the web. It includes the presentation of three different terrains (cosmological, organic, mathematical), an intelligent agent which tracks the user's progress and allows one to return to any point (presented as a phage), a few games (one similar to Tetris), and various structure options.
5. Bodies Incorporated (makes a statement about bureaucracy and representations of self). This site experiments with the use of sound as an additional navigational tool. 

He continued by briefly explaining the workings of the Interface Design Team (weekly meetings, now every two weeks, to which people bring examples of their ideas to discuss; Robert also meets individually with people as needed). Robert introduced the three people working most closely with him to develop a prototype: Nathan Fratas (with VRML and JAVA expertise), Bob Steinke (a Computer Science graduate student with CGI and HTML expertise), and Kenny Fields (music department, sound incorporation expertise). Together they are working to create an interface that meets the following criteria:

• screen aspect ratio of 640 x 480
• reasonable download time with 100K images over a 28.8 modem
• multiple frames (if using a web browser)
• new components easy to port in

Robert spent the next several minutes presenting the most current mock-ups of the new interface and explaining how they would connect to ADL's functionality. Questions from the Design Review participants included:

• How can we link the user transparently to ADL's functions, without making them wade through the various applications?
• What happens when too many results appear in one place on the map?
• How can you inform the user of ADL's coverage?
• If you can only meet the needs of one segment of your audience (given your limited resources), which will it be?
• How are you addressing differences in resolution and scale? 

Just before the small groups went to their breakout rooms, the following key questions were identified: 

1. Think about the interface from your own standpoint as a user. How would you want it represented? What do you need, personally?
2. If you could concentrate all your thinking on pitfalls to avoid, what would they be?
3. What do you want in the first version? What can be relegated to later versions? 

3.3.2 Discussion in Small Groups and Plenary Session: Interface and Navigation 

Essential or Key Points for the new interface:

** column contains the number of times that the topic was reported from the five small groups

3.3.3 ADL Summary of Comments and Response: Interface and Navigation

Mentioned more often than anything else was the notion of "Keep it Simple." This came across in the repeated requests for:  

• tutorials that are necessary only when a user is lost
• visually appealing screens
• intuitive
• integrated components that show the user an obvious, seamless path
• clarity in the sequence of operations
• common menu bar
• avoid jargon; keep vocabulary simple

There were many requests for various kinds of help: 

• context-sensitive help with no dead-ends
• pop-up help windows that are easy to access and exit
• visible, ever-present help button
• user-choice for different degrees/levels of help
• CD-ROM or VHS tutorials and walkthroughs 

When asked to consider pitfalls to avoid, three of the four groups mentioned the importance of identifying the intended audience clearly, determining their requirements, and addressing that user level accordingly. Such clarity would allow outsiders to recognize what is needed to work with the system and allow them the opportunity to educate themselves.

Finally, the design review participants encouraged us to recognize what makes ADL unique, and how ADL is different in its use of available technology. The qualities mentioned in the plenary session included: 

• ADL as an enabling tool for Map Librarians and Earth Scientists to exchange resources and data
• functionality and infrastructure (content will come later if the service is usable)
• solid, modular core of enabling technologies for the ADL community
• access paths for other communities

The plenary discussion for interface/navigation quickly shifted its focus to the question of who is our intended audience. A couple people suggested focusing on map librarians and earth scientists, with the understanding that they work with similar kinds of data and that other communities could be addressed later once there were metadata ingest standards and a sizeable catalog. If the vast majority of users are at research institutes and work with earth science data, then we have a generalizable audience to address. Recalling that educators are college-educated people, too, they can use ADL if it is a well designed resource tool and decide for themselves how to incorporate it into various curricula. Roger Downs noted that ADL is its functionality, not the content to which it provides access, and added that it almost doesn't matter who the audience is as long as the power of the system is clear. Mary Larsgaard pointed out that there are two sorts of information most frequently requested-maps and aerial photographs-and that all three targeted groups use this information, though each uses it differently. Bruce Gritton suggested that the most compelling problem for ADL is multiple audiences and scalability to them, addressing the user issues that cannot be solved within the existing infrastructure. He suggested that ADL not worry about time constraints and not limit the audience to the scientific community, but rather maintain long-term objectives.

 If ADL is perceived as a resource, tool, or any other artifact to be read, understood, and used, then the importance of knowing our audience and writing to it is clear. Our continuing evaluation studies will inform us of the expectations our targeted user groups hold, and the new interface will be designed as a communication tool between library resources and the various communities who wish access to them.

4. Initial Summary by ADL Staff (Summing Up Session at the Workshop)

Four ADL staff members presented the main points that they had gathered from the Workshop as an initial feedback to the Panel members: Jim Frew, Barbara Buttenfield, Linda Hill, and Mike Goodchild.

Jim Frew noted that the prime message is focus and the end game should be to have a working system that lasts past the funding cycle. Instead of worrying about directions we are not going in, it's all right to pick a few. Everyone seems to appreciate the tension we are in to make something survive the initial funding cycle as an operation system and also to continue research in new areas. We need to focus precisely in both of these areas and not try to be all things to all people. There was some difference of opinion about who our target users should be and what level of service we should be providing, but I haven't heard any real arguments here. I feel better about things now then I did coming in.

Barbara Buttenfield noted three major nuggets from the Workshop:

1. We do not have the resources or the time to add our own content. We need to build tools to let other people add content.
2. Issue of the focus and the target groups and the interface: we have however inadvertently set up three target groups whose interface requirements are somewhat incompatible and we need to rethink that very soon.
3. The work we have done with actual use patterns through session log analysis needs to be coordinated with the perceived use patterns to give us some sense of what users think they are doing.

Linda Hill noted the following points as coming from the Workshop:

1. The call for a human component to the user interface: I made note of the fact that having the human connection in the interface came up strongly both at the last ADR and this one. We need to begin talking about how to actually implement something like that. Because of some of the things said here, I have been thinking more about the idea of having a reference librarian or a subject specialist actually assist a user in using the ADL (behind the scene) - as part of the session. I think we would learn a lot about the type of intelligence we talk about putting into the system. Just an idea - not sure what to do with it.
2. Perceiving ADL as a tool for information professionals and designing to that end: I picked up the new term information specialists instead of information intermediaries, and got the message at several points during the workshop that we should be developing ADL as a tool for these professionals to do their job better. This is very important both in ingest and as a reference tool. This has some implications for various design components of the system.
3. The importance and promise of user profiles: There was a lot of talk here about user profiles. I think we should raise this up in the priorities and do an early implementation of some of these ideas.
4. Focussing on a particular target area: I got a great deal out of Mike Goodchild's presentation and the subsequent discussions of fleshing out a local collection. It makes a lot of sense to me and what I'm wondering is what are the implications of this for the users we go out and find, the evaluations we do because inevitably part of user evaluation responses is based on the content that are able to access. We should also think about the gazetteer in relation to chosen target areas and think extending the gazetteer information for those areas - more place names and more extensive footprints.
5. Confirmation of the value of including non-digital data in ADL through metadata: I think I heard you say - although there may have been a mixed message - that putting in metadata for items that are not actually digital is a good idea. This is good because we have some possibilities of doing that and we can take the idea of target geographic areas and then load the pointers to in-depth information about those areas that is in non-digital form.
6. Follow up on the issue of developing measures of success for ADL: I would like to follow up on Mike's idea of measures of success for ADL and Bruce's recommendation to focus on the user's problem environment: What are good measures of success? Focussing on users and their needs, their problem environments, what the gaps in their information are, what are the bridges they are trying to build for a better understanding and how successful we are in helping them.

Mike Goodchild was the last one in the group to give his summary of what he learned at the Workshop:

He said that the term operational library has rather more baggage than he originally thought. Perhaps we should be using the term Phase 2 Testbed to indicate that at the end of the project we don't mean that the research effort is over even if we have an operational system. We have to think about that quite seriously.

He picked up the message in a couple of places that if we can afford to build it, then the content would come. In a sense, the point is that content will follow good applications or good technology or good interfaces. That's worth keeping in mind.

The concept of geographic basis of demand seemed to strike a receptive chord. How do we assess the likelihood that someone located at point x is interested in information about point y, assuming that the information is stored at place z. Can we model that? How can we determine the best place to go for the needed information? Eventually, we will have to - both in terms of collection development and in terms of successful searching. There seemed to be support for our thinking about suitable models for this interaction.

He picked up the notion that the collection should reflect the interests of sponsors because this is related to funding.

There seemed to be some interest in the notion of annotation. In a digital library, we are dealing with two-way flow of information and that is a message that echoes all around the geospatial data community. One area where it is particularly compelling is in farming where information used to be from the U.S. Department of Agriculture to farmers. Increasingly, it's the other way because farmers now have much better technology in many cases than the USDA. A two-way view of geospatial information is becoming increasingly appropriate.

There seems to be a lot of interest in interoperability, taken to mean the ability to generalize the interface. For example, there are many applications where someone else's gazetteer would be more appropriate than a standard one. Is there any point in mounting the 7-million item standard gazetteer hoping to serve everyone when in many applications the gazetteer needs to be specialized? What is really needed here is the ability to operate between gazetteers. That applies also to the base maps, data formats, and numerous elements in the design.

He made the point that the assumption seemed to be uniformly made that ADL is in the public sector and that the norms of library access and library services apply. But, ADL is not necessarily a creature of the Internet. It could also be an Intranet creature. There are many applications that have massive internal information management problems that could use technology like this. It is conceivable that within a few years there will be commercial products of this nature. Our discussions about how to serve a broad audience will be usurped by a different set of criteria. What if ADL becomes a commercial product in a few years? This isn't unreasonable. He estimates that there is a market for a thousand copies in simply the petroleum industry, utility industry, etc. Our thinking needs to reflect this. Measures of success in a commercial environment are much clearer than in the public sector.

He noted a lack of much discussion of the question of verification. This is a very general and important issue for digital libraries in the two-way flow of information no longer verified by the gatekeeper function of the traditional library.

Several persistent tensions were reinforced by this meeting. One is between research and services - different measures of success. Another tension is between the library paradigm of a top-down, gatekeeper environment and a more volunteered one in which one of the things that is critically important is the amount of work it takes to document a piece of information because that essentially is volunteered. The tension between using technical terms and dumbing it down - where do we find a suitable compromise in that tension? The tension between legacy ideas as a desirable thing because they provide landmarks for navigating in a new environment and their constraining effects in preventing us from thinking outside the box. The tension between providing traditional services and providing new services which push the envelope much more aggressively.

He posed a number of strategies for dealing with these tensions. One of these is multiple views - you can create alternative views of the library along a continuum. You can compromise by finding a mid-point which doesn't satisfy anybody. But a much better solution is to admit that in many cases there is no need to resolve the tension. A strategy can deal with both ends simultaneously.

He ended with a proposal and invite comment on it. What if ADL chose the following example - in what way would this be compelling and which ways problematic? The situation I'm looking for is an example where there is no existing information commons - where new functions are necessary because the demands require it. Where there are multiple stakeholders who have typically not talked to each other. One thing that glues a community together is a shared information commons. Would a suitable example of that be a community such as Goleta and a local planning issue - like old town Goleta and its rehabilitation. There is no one place you can go to get information about that problem. Can ADL address an issue like this one by aggressively addressing an area where an information solution does not exist.

4.1 Discussion

Bill Anderson: A tremendous amount of engineering work is being done here. At the end of the project, in terms of software choices, architecture design, etc., there will be a terrific amount of reflection that could be done. From a testbed research project like this, we can learn what it takes to do it next time; i.e., how much does it cost to build and maintain a digital library. If someone wants to build a library, you can sort of tell them. This is something that fits into the future.

Jim Frew: I have always said that what ships at the end of the project is a spec. A spec driven by whatever testbed we build. I would like to deliver to our funders something that they could embed with minimal effort into an RFP --or an RFQ if you will - for a digital library. Now we know how to write an intelligent bid spec to ask someone to build a library for us. We have provided a lot of raw material that would let someone write an intelligent application.

Bill Anderson: Some engineering specs are extremely good - most of them are extremely difficult to put into practice and they don't really show anyone what the end application is and how it fits in. The Alexandria Project can show us what's new and what can be effective.

John Dullnig: I'm a little disparaged by your (Mike's) presentation of librarians as gatekeepers. That's a very old vision. Librarians are facilitators - reference librarians - and librarians are builders - catalogers. Catalogs don't spring up "from the head of Zeus." Librarians are the ones who build the catalogs. If you are going to be totally open to take in "stuff" from everywhere, there is a huge verification problem. You need someone who can arbitrate between different spellings (for example); there has got to be a referee there somewhere and this has traditionally been the role of the librarian.

Mike Goodchild: This is exactly the point I was trying to make - that the function of the library in verification is tremendously important. I was trying to contrast that with a world in which the custodian of information provides it on a voluntary basis and contributes the metadata record. In that world, there is no verification - pure caveat emptor.

Jason Simpson: I don't think we can verify all of this information. The best we can do is provide enough information for the user to make a judgement about the source and the information.

John: Compare information from a CIA page on Bosnia to information on Bosnia from someone's personal homepage. Another example is that if you buy a book from a university press, you can have some confidence in it because it has gone through the peer review process versus trade shelves. Those are different clues as to the veracity of the information.

Jason: This all goes back to what you know of the source of the information and how this affects your evaluation of the information for your own purposes - the confidence you have in the provider.

Joy Colucci: I found it a little disturbing, Mike, that you picked up on the notion that "if you build it, the content will come." The first day of this workshop we were rather vague and unfocussed and frustrated but the second day we got really excited because we were talking about content. Three of the five teams were in pretty strong agreement that content is important. And by the way your advisory board said the same thing. I think it's rather weird that you are coming away with the notion that content is not important.

Mike: It was just a quote from previous discussions.

Joy: You have a lot of people here, and there were comments like that, but that one was a minority opinion.

Barbara Buttenfield: We are walking away with the idea that content is important and that we can solicit help from the community to make that happen.

Joy: That bothers me too. In my experience, in actual operating systems where we did provide tools for people to add information, it didn't work out that way. I would love to have you show me a system like that did work. People don't like to archive the data. They don't like to publish when they are not getting peer reviewed status. It's a good idea, but it doesn't work.

Mike: Let me try to clarify. Of course, content is important and absolutely content drives the entire library. It determines ultimately every use. But in order to think about the content of this container, we have to think outside the walls of any particular model of content.

Joy: If it is so important, then you don't want to relegate it to just the notion that willy-nilly it will come or people will just put it in there.

Mike: I agree.

Michael Keeler: This refocuses back on the ingest process. The tools and capability there can serve ADL well in decentralizing the process of content input - not necessarily lowering the quality also. If you have specs in place as well as enabling tools for ingesting a variety of analog data, people have a pathway. It's not peer review but you are articulating what quality means which is needed.

Jason: Content is what the library is - the underlying backbone of the library. We were asking for guidance on what the content should be. The spin I heard was that we need to focus on what your user community here specializes in - what your library here specializes in. We need to develop the technologies to ingest our own stuff; refine the tools. We need to provide enabling tools that can be distributed to other collections that have the desire to provide access to their information and enable them to communicate with ADL for distributed searching. Through developing our own collection, we can create a generic tool that other institutions can use to participate in an exchange of information.

Ken Foote: It seems to me that there is a need for an example that demonstrates the depth of the library, as Mike was talking about in an area where there is a lack of an information commons. It makes a lot of sense to do this to demonstrate how this system can be used to integrate information about a place from many sources of information. And show how the system, as it expands, can work with existing information commons. Maybe it is already happening with the collections centered on Santa Barbara and those centered in Colorado. With those two areas, you have demonstrated most of the functions.

James Boxall: What is really exciting us (non-US observers), is the model of what you are doing and looking toward doing in 10 years hence and its overall promise rather than the specific technological developments. (Extended thanks for being invited and being able to participate).

5. Summary of Evaluation Comments on the Workshop from Attendees (responses from 12 people)

Participants asked for more clarity of purpose, a precise preview of the goals and desired results of the meeting, and focus in the meeting. They wanted the focus to be about the specifics of the project and not about the overall vision and goals. Also, a better overview of the project, accomplishments to date, and remaining issues in the beginning or before the workshop would have been helpful. A brief discussion of what a library is, what a digital library is, and what a spatial digital library is would have brought participants closer to the same shared understandings as a foundation for discussion. It would have been helpful to have the opportunity to "play" with the system while on campus before the meeting or to have more demonstration of the system.

The participants liked the structure of the presentations on the topics, followed by small group discussions and then sharing of the main points of the discussions in the large group setting. More time for unstructured discussions was requested. But, on the other hand, several thought that the meeting was too long and should have been shortened to one or two days.

A couple of comments on logistics: it would have been better to have a single point of contact before the meeting. Room setup should include tables for note taking.

Participants requested that they receive a report on the results of the workshop, particularly the specific changes that will be made as a result of the workshop as well as project updates. This report should be made available to participants for comment to ensure that it accurately reflects the content of the workshop; in particular, that the mix of opinions and suggestions from participants is indicated and that ADL doesn't report just on what they wanted to hear. To encourage continuing participation by the reviewers, ADL should show how reviewers' suggestions are taken up by the project.

Some participants felt that ADL needs a new concrete plan document that specifies needs, requirements and time goals - or the project's mission, goals, and scope and an implementation plan for going operational.

The small group discussion structure received many favorable comments as a way to fully explore the issues and interact with fellow participants. Keeping the same groups together throughout was seen as a good decision by most, but some felt that the arrangement limited exposure to differing viewpoints. Large group discussions were appreciated as giving the opportunity to hear interesting views from different backgrounds. The "framing" presentations received mixed reviews, with some appreciating the information content and others seeing some excess in the "slide show" approach.

Interaction with other panel members and with the ADL staff was cited several times as the best thing about the workshop, including the social events and the helpful, accommodating, and receptive attitudes of the staff.

Some said that nothing needed to be changed - that the workshop was well done and that we should keep the basic format.

Better focus was cited again as something that needs to be improved, with specific tasks for reviewers. The focus needs to be on issues that the panelists can have an impact on, not a rehash of issues already decided by ADL staff. The request for input should be made more specific and it should be demonstrated that the advice was heard and given serious consideration.

ADL staff should listen more and talk less and not step in to defend the project in response to critiques and criticisms. They should play more of a facilitation role and not be as active as they were in the discussion.

Individual participants should not be allowed to dominate the discussions.

The diversity of participants was cited as a strong component of the workshop.

Add more hands-on use of the system and discussion of what is and what is not happening.

There was mixed comments about the small group structure: some said keep the structure as it is but there were also suggestions for improvement: having five parallel groups discussing the same issue could be a luxury and create too much overlap; include a mix of interests in each group and make the first small group session longer so that the group can get a feel for the concerns of others in the group.

Add a session at the end where each small groups presents their ideas on "If we built ADL, it would ...."

Some felt that the Pacifica Suites was not the best choice since it was isolated from Santa Barbara.

There was a suggestion to "require" participants to go through the beta test questionnaire before they were allowed to come.

There were very favorable and complimentary comments made about the workshop and appreciation for being included.

Concern was expressed that some of the project leaders were absent from much of the workshop.

Concern was expressed that the time was not well-spent because there was no confidence that ADL will really take up the comments made by the participants.

For the next workshop, pick a few key questions where feedback is needed and make these as sharp and focussed as possible.

6. Impact of the Design Review on ADL

The total impact of the discussions and recommendations of the Design Review Workshop on the focus of the Alexandria Digital Library will not be entirely clear immediately and will be influenced by continuing contact and dialogue between Panel members and ADL staff. Four specific impacts have been identified as of the writing of this report: (1) on collection development strategies, (2) on user profiles as components of the user interface and system functionality, (3) on providing human support to ADL users, and (4) on interface design for ADL. The current directions of ADL planning are described below; these plans incorporate ideas from the Design Review Workshop.

6.1 ADL Collection Development Strategy

A high priority activity since the Design Review Workshop has been the formalizing of collection development guidelines for ADL. These guidelines are being adopted within a distributed library model where multiple nodes would build collections about their own local areas and thematic interests. Access to multiple collections would be supported through distributed search and retrieval protocols.

The draft Content Selection Guidelines include:

1. All items must have geographic footprints; i.e., latitude and longitude representations of the geographic coverage of the item.
2. Selection should be driven by the ADL user community, through knowledge of their research interests.
3. The processing burden of ingest should be minimized through selection of items that are already in digital form, that are unencumbered, and that already have metadata.
4. Unique collections that are not otherwise available will be given preference over collections that are available through other convenient channels.
5. ADL is aiming for a large number of items rather than a smaller number of large items.
6. The collection should be built to showcase ADL functionality, such as heterogeneous data types, heterogeneous search capability, post-retrieval processing, and distributed search and retrieval.
7. The UCSB node of ADL will focus on the local area perceived as a geographic cone centered on Santa Barbara County and expanding out to the Southern Coastal Region, Southern California, California, Southwestern U.S., etc. The scale or resolution of coverage will decrease as the area footprint increases.

Candidate collections are being identified according to these criteria and prioritized for loading into ADL. For example, collections that include both data and metadata for the local area such as AVHRR, Digital Elevation Model data, Digital Raster Graphics, scanned air photos, and Thematic Mapper data are on the list. Collections that have metadata only include GeoDex, GeoRef records for California that contain footprints, and UCSB library catalog map records for the area. A special georeferenced bibliography collection covering the Mojave Desert will be loaded.

6.2 ADL User Profile Strategy

The User Interface and Evaluation Team is beginning the design of the user profile component for the new ADL web-based interface, keeping in mind the ideas for user profiles that were expressed at the Design Review Workshop. In particular, we note the consideration for multiple profiles for one person and the consideration of how a profile can influence interface presentation, query parameters, and system functionality (e.g., influence the way in which results sets are ranked).

So far, user profiles have been defined to consist of both user provided information and selected systems settings and system acquired patterns of usage. User provided information will consist of a careful chosen set of demographic information useful for (1) characterizing the user community, (2) providing contact information, and (3) potentially providing authorization for access to controlled collections. User-provided demographic information is assumed to be relatively stable information. User preferences and system settings, on the other hand, are more dynamic and it is here that multiple profiles would be supported. Preferences would be set to default values with additional settings available for choice. The list of customizable parameters will be tied to the interface design.

System provided information could include logs of session activities, state of the interface which can be used for resumption of a session, stored result sets, stored queries, and other passively recorded usage patterns that could be used to automatically customize user/system interactions.

Under discussion is how user registration will be implemented for ADL, given that it is to be an open system for anyone to use. Registration will be by user choice and it is still to be determined what, if anything, a registered user will be able to do or have access to that is different from a "guest" user.

Key to the design of user profiles is to have a clear plan for using the information collected. The User Evaluation Team, the Interface Design Team, and the Implementation Teams will all be working on various aspects of the design. Implementation of some form of user profile is planned for the new interface. Research aspects of user profiles include characterizing the user community, interpreting patterns of use and user evaluation and feedback on the basis of demographic attributes, and using profiles as a factor in system presentation and query analysis.

6.3 ADL Human Support Strategy for Digital Library Users

The Design Review Panel highlighted yet again the need for digital libraries to provide personalized help to online users, emulating the reference and technical assistance that is available from libraries and data centers. ADL is developing a strategy for providing such help, initially with some degree of "help desk" support through email in the new web interface and eventually as a component of an operational library which could potentially include both human assistance and the help of "wizards" specifically designed for particular functions.

Operational issues center on the management model and the infrastructure that would be required to provide live support to online users. Components of this are: (1) the range and breadth of expertise that would be needed by the help desk staff, including knowledge of the data and its appropriate use; software, hardware and networking knowledge; and knowledge of the metadata and information resources; (2) the software support required for logging, tracking, archiving, and searching the user contacts and the issues that are raised; (3) searchable FAQs for users; and (4) service considerations such as hours of operation, turn around time benchmarks for responses, and quality control measures. Many of these issues have been dealt with in libraries, data centers, commercial operations, and other service desks - these models will be investigated for application to ADL as an operational library component.

Research issues include technology developments and the use of user feedback through help desk interactions to characterize the user community and to inform interface and system design and collection development. Interesting technological research areas that could be pursued include the use of live video links between ADL users and help desk staff; a way for users to "share" their ongoing sessions with help desk staff for interactive assistance; and expert system development to automate an increasing portion of the support function so that support can be given to a larger number of users than human support alone can ever hope to do. Incorporating help desk interactions into user evaluation studies will be a natural expansion of ongoing studies.

The Design Review Workshop made clear that augmented help services must be part of ADL. This is recognized as a high priority and is being addressed by the systems implementation, interface design, and user evaluation teams and by a committee formed to plan for an operational digital library. Most likely, a phased approach will be tested starting with some combination of email and phone assistance and expanded as technology and resources permit.

6.4 Interface Design for ADL

ADL is currently designing a new user interface. The comments that were made about interface design at the workshop are being fed into the design process along with the other sources of user comments on the beta web interface. The direction to "keep it simple" is a goal that we are trying to meet, but this is a tricky thing to do when the information sources and the processes of search and retrieval are themselves complex - especially since geospatial representations must be smoothly integrated with textual and quantitative representations to support users who are working with georeferenced data, images, and text.

The interface is being designed to support iterative searching based on the review of initial and interim results. The following are some of the iterative processes that we hope to support in the interface:

• Portions of images can be used to find other image patterns that are similar.
• Footprint clusters from result sets can be used to identify areas of interest (target areas) for further searching.
• Place names can be used to locate target areas and then additional place names can be identified in the same location.
• Users will be encouraged to search first by geographic location. The system will then display the types and numbers of things that are relevant to the target area(s) submitted. The user can then narrow the search by additional parameters or choose a different geographical area, using the information already retrieved to guide the search process.

A user will be able to review the things that they have done during a session at any time and revisit an earlier point in the session history. The ways in which "help" will be available are being developed with the goal of not making it "necessary" for a user to study the tutorial before they are able to do basic operations.

Thesaurus and concept space expansion will probably not be in the first release of the new interface but will be part of future plans.

Ways of visualizing results sets are being considered. Initially, the user will be able to sort the results by type of material (texts, maps, images, etc.) and by date.

An activity indicator is being included in the new interface that shows that the retrieval process is still active.

The Design Review Panel will be asked to review and comment on versions of the new interface design as it develops.

7. Future Plans for the Design Review Panel

Each member of the Design Review Panel will continue to receive information from the ADL Project and will be asked to participate in design decisions. Each member has been assigned an ADL staff member as a principal contact to facilitate communication. We have established an email distribution list for Panel announcements and discussion: alex-adr@alexandria.ucsb.edu. A final Design Review Panel workshop will be planned for 1998; the agenda for this session will be worked out in advance with Panel members.

Appendix A. Attendees

Prue Adler
Assistant Executive Director
Association of Research Libraries
21 Dupont Circle, Suite 800, N.W.
Washington, D.C. 20036
prue@cni.org
(202) 296-2296
Fax: (202) 872-0884

Alper Naim
Computer Scientist
SAIC/NRL Monterey
585 Hawthorne #204
Monterey, CA 93940
alper@nrlmry.navy.mil
(408) 656-4756
Fax: (408) 656-5769

William L. Anderson
Xerox Corporation
100 WillowBrook Office Park
Fairport, NY 14450
Band@wc.mc.Xerox.com
(716) 383-7983
Fax: (716) 264-2489

Christopher Baruth
Library, University of Wisconsin-Milwaukee
P.O. Box 399
Milwaukee, WI 53201-0399
cmb@gml.lib.uwm.edu
(800) 558-8993, (414) 229-6282
Fax: (414) 229-3624
http://leardo.lib.uwm.edu 

Cecil Bloch
Multimedia and User Interface Consultant
49 Showers Drive #E-249
Mountain View, CA 94040
DoctorMax@aol.com
(202) 395-7600

Joseph Boisse
University Librarian
UCSB Library
University of California, Santa Barbara
Santa Barbara, CA 93106
boisse@library.ucsb.edu
(805) 893-3256
Fax: (805) 893-7010

James Boxall
Map Curator
Dalhousie University
Halifax, Nova Scotia
CANADA B3H 4H8
jcboxall@is.dal.ca
(902) 494-3757
Fax: (902) 494-2062
http://www.library.dal.ca/science/ mapcoll.html

Frank Celona
DEC
fcelona@ljo.dec.com

Joy Colucci
Hughes Information Technology Systems
702 Blossom Lane
Redondo Beach, CA 90278
jcolucci@eos.hitc.com
(310) 372-0795
Fax: (310) 372-0725 

Roger Downs
Pennsylvania State University
302 Walker Building
University Park, PA 16802-5011
RD7@PSUVM.PSU.EDU
(814) 865-1915 or 865-3433
Fax: (814) 863-7943

Beth Duff
Product Development Branch Chief
U.S. Geological Survey
12201 Sunrise Valley Drive
Reston, VA 20192
bduff@usgs.gov
(703) 648-4621
Fax: (703) 648-5939

John Dullnig
Integrated Library System Migration Program Manager
NIMA
3415 Elenoir Court
Bowie, MD 20716
dullnigj@nima.mil
(202) 863-3030
Fax: (202) 863-3617

Ken Foote
Associate Professor
Department of Geography
University of Texas at Austin
Austin, TX 78712-1098
k.foote@mail.utexas.edu
(512) 471-5116
Fax: (512) 471-5049

Christian Foster
Director
K-14 Community Outreach Coordinator
Graduate School of Education
University of California
Santa Barbara, CA 93106-2010
cfoster@education.ucsb.edu
(805) 893-7966
Fax: (805) 893-7264 

Alan Gaines
Head, Research Grants Section
NSF Earth Sciences
4201 Wilson Blvd., Room 785
Arlington, VA 22230
againes@nsf.gov
(703) 306-1553
Fax: (703) 306-0382

Bruce Gritton
Monterey Bay Project
grbr@mbari.org

Tim Gubbels
Hughes EOSDIS Cores System Project
1616 McCormick Drive
Landover, MD 20774
tgubbels@eos.hitc.com
(301) 925-0452
Fax: (301) 925-1151

Phil Hoehn
Acting Map Librarian
Stanford University Libraries
Branner Earth Sciences Library
Stanford, CA 94305-2210
(415) 725-1103
Fax: (415) 752-2534
http://www-sul.stanford.edu/depts/branner/

Carol Hughes
Member Services
Research Libraries Group
1200 Villa Street
Mountain View, CA 94041-1100
bl.cah@rlg.stanford.edu
(415) 691-2293
Fax: (415) 964-0943

Michael Keeler
President
ECOlogic
19 Eye Street, NW
Washington, DC 20001
keeler@ecologic.net
(202) 218-4100
Fax: (202) 842-5088

Edward H. Liszewski
Chief Librarian
US Geological Survey Library
101 National Center
Reston, VA 22092
eliszews@usgs.gov
(703) 648-4305
Fax: (703) 648-6373

Karl Longstreth
Map Librarian
University of Michigan
825 South Hatcher
Ann Arbor, MI 48109-1205
karleric@umich.edu
(313) 764-0407
Fax: (313) 763-5080

Clifford Lynch
Directory Library Automation
Office of the President, University of California
6400 Christie #5406
Emeryville, CA 94608
Clifford.Lynch@ucop.edu
(510) 987-0522
Fax: (510) 839-3573

Patrick McGlamery
Map Librarian
University of Connecticut
MAGIC U-5M, Homer Babbidge Library
Storrs, CT 6268
pmcglamery@lib.uconn.edu
(860) 486-4589
Fax: (860) 486-3593
http://magic.lib.uconn.edu

Roberta Balstad Miller
CIESIN
2250 Pierce Road
University Center, MI 48710
roberta@ciesin.org 

Eugene Miya
Initiative Reviewer
NASA Ames Research Center
MS 233-10
Moffett Field, CA 94035
eugene@ames.arc.nasa.gov
(415) 604-4407
Fax: (415) 604-6999

Loren Nibbe
Santa Barbara City Schools
lnibbe@aol.com

Lola Olsen
Global Change Master Directory Manager
NASA Goddard Space Flight Center
Code 902
Greenbelt, MD 20771
olsen@gcmd.gsfc.nasa.gov
(517) 797-2727
Fax: (517) 797-2622
http://gcmd.gsfc.nasa.gov

Lisa Ritter
Research Cartographer
National Geographic Society
1145 17th Street, NW
Washington, DC 20036
lritter@ngs.org
(202) 775-7834
Fax: (202) 429-5704

Gary Strong
Program Director
National Science Foundation
4201 Wilson Blvd., Room 1115
Arlington, VA 22230
gstrong@nsf.gov
(703)306-1928
Fax: (703)306-0599
http://www.cise.nsf.gov/iris/ISPhome.html 

Nancy Tosta
Director of Forecasting and Growth Strategy
Puget Sound Regional Council
1011 Western Avenue, Suite 500
Seattle, WA 98104-1035
tosta@psrc.wa.com
(206) 587-5665
Fax: (206) 587-4825

James F. Williams II
Dean of Libraries
University of Colorado Libraries
Campus Box 184
Boulder, CO 80309-0184
James.Williams@colorado.edu
(303) 492-7511
Fax: (303) 492-1881

ADL Staff Participants

Barbara Buttenfield
University of Colorado
Babs@colorado.edu

Larry Carver
UCSB
carver@alexandria.ucsb.edu

Jim Frew
UCSB
frew@alexandria.ucsb.edu

Mike Goodchild
NCGIA
good@ncgia.ucsb.edu

Linda Hill
UCSB
lhill@alexandria.ucsb.edu

Randy Kemp
UCSB
kemp@alexandria.ucsb.edu

Mary Larsgaard
UCSB
mary@alexandria.ucsb.edu

Suzanne Larsen
University of Colorado
suzanne.larsen@colorado.edu

Dan Montello
UCSB
montello@geog.ucsb.edu

Robert Nideffer
UCSB
nideffer@arts.ucsb.edu

Mary-Anna Rae
UCSB
mrae@alexandria.ucsb.edu

Doug Ramsey
UCSB (visiting from Utah State Univ.)
doug@nr.usu.edu

Keith Rokoske
University of Colorado
keith.rokoske@colorado.edu

Mike Rock
University of Colorado
rockm@ucsu.colorado.edu

Jason Simpson
UCSB
simpson@alexandria.ucsb.edu

Andrew Smith
University of Colorado
andrew.d.smith@colorado.edu

Terry Smith
UCSB
smithtr@cs.ucsb.edu

Ming Tsou
University of Colorado
tsou@ucsu.colorado.edu 

Appendix B. Meeting Agenda

ADL Design Review Meeting, Number 2

Goals of the Meeting

To collect comments and recommendations from advisors who can potentially benefit from the Alexandria Digital Library Project and who can contribute needed expertise to the ongoing design and development of the digital library. Results will be captured in a report which will be distributed to the Design Review Panel members and which will serve as a guideline document to the ADL staff. Evaluation of the meeting will be used to plan the next meeting of the ADR Panel early in 1998.

Agenda

Wednesday, February 19th