Carnegie Mellon Libraries:

SMART WEB EXHIBITS:

Delivering Enhanced Library and Museum Collections
Online, On Target and On Time

NARRATIVE
The Problem
Collection holders assemble collections, not to gather dust or await the occasional scholar, but to educate and edify people. Physical space has traditionally constrained the presentation of collections, causing museums, libraries, and other collection holders to limit the size of their collections or store items off-site. Museums, for example, do not have sufficient space to present to the public all of the specimens, artifacts and information they retain about them in their collections. The need to design exhibits that are visually straightforward and interesting to visitors further minimizes the size of the collection exhibited. Focusing on key objects and facts about them, museum exhibits in effect stand alone, with no links to the wealth of related information and objects available within the museum and beyond. Libraries are in a similar bind, with insufficient space to organize, present and link their collections most effectively for their users. To exacerbate the situation, libraries and museums hold complementary collections of material. The books tell about the objects, and the objects tell about the books. But physical constraints limit or disallow the collaborative organization, presentation and linking of related materials in library and museum collections. Furthermore, some of the books and most of the objects are so fragile or precious that handling by users must be discouraged if not prohibited. The quality and depth of education would be enhanced if these physical limitations could be overcome and libraries and museums could collaborate in educating and edifying the community.

The "community" served by museums and libraries is really many diverse user communities. Users have different needs and expectations. For example, the information needs of university research faculty are significantly different from the needs of elementary, high school or undergraduate students. Users are constrained by many factors, including the time they can devote to information seeking, the purpose of their information seeking, and the level of information they can comprehend. Today computers are integral to education, library catalogs are available online, and museums have roving robots that answer questions about the exhibits. Access to the World Wide Web ("Web") is so widespread that businesses advertise their Web site addresses and institutions of higher education speculate that distance education will transform college campuses. Web television suggests Web access in every home.

On the down side, research shows that many people turn to the Web first when seeking information, and that most people get lost "surfing" the Web, have difficulty navigating the hierarchies inherent in the Web, and retrieve too much information using Web search engines (the "drinking from a firehose" metaphor). Education is in peril because much of the information retrieved on the Web is irrelevant, inappropriate, or dubious for the user's purpose. The gatekeeper function of the traditional print publisher is not found on the Web. To complicate matters, more often than not, today's information seekers are not in a library when they're looking for information. They have no access to a reference librarian who could assist them in finding appropriate, quality information for their needs. Users turn to the Web in search of quality information but there isn't much of it there. Because of economic, legal and security concerns, print publishers are hesitant to make their collections available online, and even when they do, the information typically is not indexed by Web search engines so users can't find it anyway.

The digital future provides new opportunities to solve these problems and develop new teaching tools and strategies dovetailed to the diverse needs and expectations of a fast-paced, on-demand Web-based world. Carnegie Mellon University Libraries, the School of Computer Science at Carnegie Mellon and the Carnegie Museum of Natural History propose to develop, document and disseminate prototypes for a new type of collaboration between libraries, museums and other collection holders. This collaboration will result in an improved educational outreach to the public in the form of "Smart Web Exhibits" (SWE) designed to deliver information online, on target and on time. The project will develop the content for two SWE and provide operational prototypes for library and museum experimentation.

To solve the problem of retrieving irrelevant, inappropriate or dubious material, the content of the SWE will be carefully selected, organized, and linked by experts. The success of the proposed project depends in large part on careful content development to meet different user needs. New software tools such as an interactive automated reference assistant will target the appropriate level and volume of information for the user. If a museum visitor requests a one minute introduction to the Diplodocus carnegii specimen on display, a five minute overview of the entire dinosaur collection, or a one hour tour of a paleontologist's field notes, reports and photographs, the SWE will deliver. If a library user sees a precious book or manuscript in a presentation case and wants to view the pages of the book online, order a facsimile copy of it, find articles about it or references to it, the SWE will deliver. In some cases, for example, exhibited rare books, SWE will reduce the need for handling the real objects while enhancing the user's experience of seeing them.

The SWE is a Web site specially constructed to accompany a museum or library exhibit. It invites users to explore the topic of an exhibit more deeply using a desktop or laptop computer running a standard Web browser. The web site can run on the computer or, to save expense, on a central server. An SWE has these features:

In a real sense, a Smart Web Exhibit takes the notion of a display of a collection to the next level of detail without cluttering the available physical space. With a few changes of introductory presentation, localization, and order taking, libraries or museums can make the same SWE available on the Internet. They can re-use the infrastructure and authoring tools to create new SWE by changing the pre-selected content.

The Plan
The purpose of the proposed project is to systematically develop and disseminate the know-how to collaborate and deliver quality information on target and on time to Web users. The project goal will be achieved through the design, implementation and evaluation of two Smart Web Exhibits (SWE) -- one for library exhibits and one for museum exhibits -- to ensure that all of the technical, content and usability issues have been addressed by the research. Library and museum representatives carefully selected the content for the proposed exhibits to emphasize collaboration.

The museum SWE will focus on text and images documenting the discovery of dinosaurs now housed in the Carnegie Museum of Natural History. Users will be able to search field photographs, correspondence of dinosaur hunters with Andrew Carnegie and Museum staff, field notes and other reports, and popular accounts and scientific publications of the period. Accompanying the "signature piece" of the dinosaur exhibit, the Diplodocus carnegii found in 1899, will be records and correspondence pertaining to its discovery, transfer and reproductions. The content of the museum SWE will be chosen from 1200 photographic negatives, 9 linear feet of archival material, and 0.5 linear feet of documents pertaining to the formation of Dinosaur National Monument.

The library SWE will focus on text and images documenting the evolution of artificial intelligence, specifically the Herbert Simon and Allen Newell Collections in Carnegie Mellon University Archives and related artifacts and information at the Carnegie Science Center. Users will be able to search photographs, correspondence, lecture notes, and published and unpublished papers of two of the early innovators of computer science, and trace the evolution of their work in artificial intelligence by interacting with a presentation on the development of a computer chess program that can beat a grand chess master. Artifacts in the exhibit will include the first typewriter fully assembled by a robot and the Nobel Prize awarded Dr. Simon in 1978. The content of the library SWE will be chosen from over 160 linear feet of archival materials, including over 200 photographs, thirty-two hours of videotape and 100 artifacts documenting the early history of computer science and artificial intelligence.

The revised library and museum SWE will be released for use within Carnegie Mellon University Libraries and the Carnegie Museum of Natural History. A slightly modified version will be "published" on the Web and maintained for use beyond the duration of the project. The SWE will provide a model of coherent and relevant organization for multimedia collections enhanced by links to additional resources. Because they can easily be adapted to new content, they will demonstrate the viable use of standard Web formats and technologies to expand access to information and achieve digital preservation goals. Most importantly, the "smarts" of the SWE -- the high quality of the content and the design and functionality of the interface -- will address the legitimate concerns educators have about the Web. By focusing on tools that can be widely distributed, the inevitable need to retain local creative control can be met without sacrificing the generality of the content and organization models provided.

Digitization / Likely uses of and demand for materials in digital form
The content of the proposed library and museum SWE will attract a wide range of users interested in dinosaurs and computers (the past and the future), from the casual museum visitor or curious web surfer to the academic scholar or field researcher. Most of the SWE content will be unique and precious in its own right: original archival materials and museum objects. With such content pre-selected, organized, integrated and enhanced by experts, the SWE will add value to the physical exhibit, the collection and the Web (where quality is hard to find). The Web itself enlarges the audience for the exhibit and enables navigation and manipulations of the collections not possible without computer technology. Though the SWE prototypes will be developed to accompany physical exhibits, they will be available to educate and edify users who do not see the physical exhibit, just the SWE on the Internet.

The museum and library SWE will provide access to unique materials that have never before been presented to the public as an integrated collection. For example, for the first time in history, museum "visitors" (who may or may not be in the physical museum) will be able to see illustrations of activity in the field, the removal of dinosaur specimens from the surrounding rock, and the installation of specimens in the Museum. Similarly, library users (whom research shows are not or prefer not to be in the physical library) will be able to see Simon and Newell's original research and manuscripts related to the development of software that can problem-solve and learn like a human being. Imagine seeing the incoming field photos, reports and correspondence of paleontologist Earl Douglass, who shipped 750,000 pounds of fossil remains from Utah to Pittsburgh over a period of sixteen years, or seeing Nobel Prize recipient Herbert Simon ponder the implications of his work on intelligent machines for our social institutions. The content of the museum and library SWE will have something to offer a wide range of users, from children to life-long learners.

Standards for image quality, content descriptions, and system management / Plans for preservation and maintenance of digital files
One of the Principal Investigators, Robert Thibadeau, has headed the Imaging Systems Laboratory in the Robotics Institute at Carnegie Mellon University since 1982. He will oversee the digitization efforts to make sure that all digitization of documents and objects conforms to rigorous demands for preservation quality digital imaging. (See www.ecom.cmu.edu/preservation for a recent paper on this subject by Dr. Thibadeau.) Video will be encoded in both Quick Time and NetShow format. In the absence of standard content descriptions for digitized video, audio, photographs, and museum objects, the project will follow best practices. For example, the ISO 23950 (NISO Z39.50) profile for the Computer Interchange of Museum Information (CIMI) will be investigated for best practices regarding museum metadata. The Machine Readable Catalog (MARC) record will be the standard for content descriptions of books, monographs and journals. The Encoded Archival Description (EAD) will be the standard for content descriptions of archival materials.

Quality control and system management will be handled similar to the HELIOS project (1993-1997) at Carnegie Mellon University Libraries, which continues to provide Web access to 700,000 digitized pages of legislative records from the Heinz Archives. Web users will be able to view much if not all of the SWE content for free. Fees will be imposed for facsimile printing. Other restrictions may apply if the copyright holder imposes restrictions. The SWE content will be backed up so that it can be restored in case of emergency (e.g., data corruption, disk failure), and the system will be secured against hackers and intruders.

Carnegie Mellon School of Computer Science (SCS) will provide 45 gigabytes of server space for the project and maintain the files in perpetuity as part of the Universal Library Project (described below). Inevitably, the files will have to be migrated to new systems and technologies over time. SCS has considerable experience migrating data and developing software with this in mind.

Equipment
Contributed: The Imaging Systems Laboratory and the University Libraries at Carnegie Mellon will make existing equipment available for the SWE digitization work, including:

Carnegie Mellon also has a Xerox DocuTech scanner/printer that will be available for the project to enable facsimile printing from the SWE. In addition, the School of Computer Science will purchase and contribute the following equipment to the SWE project:

The cost of the new equipment amortized over two to three years (as appropriate) is approximately $58,000. The amortized value of the used equipment and the use of the DocuTech as provided by the Office of Property Management is approximately $18,000. Specifics are provided in the Detailed Budget.

Purchased with project funds: A digital spectrophotometer in a goniometer configuration will be purchased from project funding ($1500) to characterize the reflection characteristics of the various materials to be scanned and imaged. This will allow images of the material to be calibrated against their visual appearance under different lighting conditions and different angles of view. The spectrophotometer creates metadata for an image that constrains the faithfulness to the original object. The colorimeter, contributed to the SWE project, uses this metadata to directly transform the image display to an accurate color rendition of the object. The SWE project will characterize the reflection characteristics of sample objects to be included in the prototype exhibits.

Copyright Issues
Intellectual property issues are among the most difficult to resolve in creating relevant digital collections. The focal content in the prototype SWE project is archival in nature, with supporting content provided by published material. Ironically, there are fewer restrictions on the use of precious archival material, like the original papers of the Diplodocus paleontologists and the progenitors of artificial intelligence, than on printed (published) material. The SWE project entails contacting publishers to negotiate conditions under which they will allow their copyrighted information to be included in the exhibit. This project will test publisher attitudes and practices under the leadership of Principal Investigator Michael Shamos, intellectual properties attorney and computer scientist.

Management Plan
The Carnegie Museum of Natural History is responsible for selecting and organizing the content and providing appropriate metadata for specimens and artifacts included in the museum SWE. The museum will also provide video interviews with curators of Vertebrate Paleontology collections and experts to offer (digital video) tutorials to be included in the museum SWE. Museum partners will assist with scanning selected objects and designing the SWE end user interface. Principal Investigator Bernadette Callery will oversee the project at the Museum.

In the School of Computer Science (SCS) at Carnegie Mellon, Principal Investigator Michael Shamos will negotiate with publishers to secure copyright clearance (where applicable) to digitize content for the museum and library SWE. Discovering and documenting the conditions (if any) under which different publishers will allow geographically constrained or remote Web access to their materials in SWE will contribute new information to the ongoing dialogue about electronic licensing and possibly a new application of electronic commerce to education and lifelong learning.

Also in SCS, Principal Investigator Robert Thibadeau will oversee the development of the Web site infrastructure and authoring, indexing and analysis tools, production of digital video tutorials, and digitization of objects (including photographs, videos and dinosaur drawings) for both the museum and library SWE. He will ensure that only industry standard digital formats are used and that the color of the image is that of the original object.

Carnegie Mellon University Libraries are responsible for selecting the content, providing the metadata and organizing the library SWE. The Libraries will scan the document content for the library and museum SWE (the archival documents, books, monographs and journals), and assemble both SWE. The Libraries may digitize selected artifacts in the Carnegie Science Center. University Archivist Gabrielle Michalek will oversee the SWE project at the University Libraries and serve as overall SWE Project Coordinator to monitor the project budget, facilitate communication among project participants, ensure that milestones are met, and submit interim and final reports, including a financial report. Michalek managed the money, personnel, facilities, equipment, and supplies in the HELIOS project. It was in large part her rigorous management and fiscal responsibility that completed the HELIOS project under budget; sufficient funds remained after digitizing the materials and implementing the initial system to purchase hardware and license software to migrate the data to a new system.

Evaluation
By adhering to standards, following best practices, and applying the collaborative expertise of computer scientists, librarians, museum specialists and an intellectual property attorney to system design and functionality, the SWE will produce three clear and measurable outcomes. First, the SWE project will deliver Web accessible, high quality, high value content that can and will be preserved over time. Second, the project will produce the SWE organizational framework and authoring, indexing and presentation tools that add value to the content and can be re-used over time with other digitized multimedia collections. Third, the SWE project will conduct and disseminate the results of research in the areas of human factors, electronic commerce, digital imaging, and the collaboration of museums and libraries in the service of education. The results of this research should invite further research into ways to augment education and outreach on the popular and prevalent Web.

Carnegie Mellon University Libraries will conduct human factors research to identify usability problems in the prototype SWE. The design and functionality of the prototypes will be refined based on the results of this research. The Libraries will also monitor usage of the SWE to determine whether library users and museum visitors seek information differently. For example, how frequently do users use SWE associated with library and museum exhibits? How long do they spend at the SWE? Does time vary depending on whether they are viewing an SWE provided at the site of the physical exhibit or somewhere else (e.g., home, school or office)? Will the time increase as users become intrigued by the topic or familiar with SWE? What media are most popular in a digitized multimedia collection? Will the level of sophistication of user queries increase with exposure to the capabilities of SWE or the addition of content SWE (as more copyright permission enables more digitization or more related Web sites with quality information appear)? The Libraries have a decade of experience conducting this type of research. The usage monitoring and analysis tools are an integral component of the SWE, so anyone who builds an SWE can model use of the content.

Principal Investigators Michael Shamos and Gloriana St. Clair will conduct research in electronic commerce. Experience with electronic reserves indicates that publishers will allow use by a known audience (like an academic class identifiable by the Registrar's Office) at no charge or for a small charge. Whether publishers will expand this practice to users physically in museums and libraries, and the conditions under which they will allow remote Internet access to their materials in SWE is unknown. This work will answer such questions as: how difficult is it to clear copyright for old and new publications, and are small university presses more likely to grant copyright clearance than big commercial publishers? Caught in a paradigm shift, libraries at the turn of the century must decide whether to build new buildings to house physical collections or whether to digitize the collections instead. Digitization may be an affordable alternative and the preferred alternative for preservation purposes and Web users, but if the scholarly publishing culture is not yet read to provide quality content online, on target and on time, libraries and museums need to know that.

Principal investigator Robert Thibadeau will conduct research in digital imaging that preserves the appearance of the original object (see http://www.ecom.cmu.edu/preservation). The focus of the work will be on resolving unknowns in spectrophotometric and colorimetric reflectometry and developing simple heuristics to predict the appearance of an object under different lighting conditions and from different angles of view. The heuristics would take account of CCD imaging, low temperature lighting methods, and current practice in 3D-texture mapping.

Dissemination, Information Access, and Sustainability
Information about the development of the SWE paradigm and the release of the library and museum SWE will be broadly disseminated and easily accessible. References to the project will be made in appropriate campus and Museum publications (online and in print) and on the Web sites of Carnegie Mellon University Libraries (www.library.cmu.edu) and the Carnegie Museum of Natural History (www.clpgh.org/cmnh). Articles about the project will be submitted to the Web-based journals of the Coalition for Networked Information and Digital Library Federation, and to popular and scientific print publications, for example, CRL News, American Libraries, Library Hi Tech, Library Information Technology, American Archivist, Museum Archivist, Archives of Natural History, Nature, Science, Discovery and Science News.

Not only information about the SWE, but the SWE themselves will be accessible on the Web. The SWE will continue to deliver enhanced library and museum collections online, on target and on time beyond the duration of the project and serve as a model for other educators and experts to add value to their collections and Web sites.

The use of software standards and "signature" collections will ensure that the SWE content is preserved for future use. The University and Museum will always be looking for ways to preserve, expand and provide access to the unique materials in the Vertebrate Paleontology and Simon and Newell Collections. The strategic plans of both organizations include efforts to meet or exceed national standards for computerization and access to their collections. The preservation and evolution of the SWE are further ensured by Carnegie Mellon's expertise and commitment to digital libraries and the development of intelligent software and efficient data storage and migration techniques.

The proposed SWE Project is in fact an effort in a much larger vision at Carnegie Mellon called the "Universal Library Project." A collaborative project of the Language Technologies Institute in the School of Computer Science (SCS), headed by Dr. Jaime Carbonell, and the University Libraries led by Dr. St. Clair, the Universal Library Project is a long-term (century long?) attempt to digitize the global multimedia history of human consciousness and culture (books, documents, audio, video, photographs, museum objects). Project Director Michael Shamos leads the efforts to secure the rights to digitize the information, efforts made difficult by copyright laws that vary from country to country. Scanning is underway, but efforts continue to secure funding to do the processing and research necessary to carefully digitize, preserve and provide user-friendly access to all that exists of human history and expression. Preservation is required because much of our history and culture is decaying or has been (or will be) destroyed. User friendly access to quality content and smart interfaces is required because of the problems on the Web that imperil education and scholarship today. The Universal Library Project vision includes migrating the collections to new technologies again and again over time to preserve them, free searching and discovery of content (if not free viewing), eventually automated summaries of the items retrieved (to deal with the "drinking from a firehose" metaphor), and ultimately machine translation of full text and discourse among many languages (to facilitate education and eliminate re-inventing the wheel in each culture). Principal Investigators on the SWE Project Thibadeau and St. Clair are also Directors of the Universal Library Project. If funded, the library and museum SWE project content will be available (preserved) within the Universal Library (www.ulib.org).

National Impact and Adaptability
The major goal of the SWE project is to produce an organizational paradigm and software tools for the low cost, rapid production of Smart Web Sites on a national basis. Low cost computers and the standardization of the Internet offer a straightforward means of enriching museum and library exhibits through the development of SWE. The SWE paradigm will contribute significantly to the process of providing more in depth educational opportunities to conventional museum and library exhibits, and offer Web site developers in any walk of life critical technologies for meeting the needs of diverse user communities. The organizational infrastructure and tools of the SWE can be re-purposed for future use with different content, thereby saving money on future exhibits.

The SWE Project will have national impact because it will produce four significant deliverables in addition to the SWE themselves:

This package of deliverables will enable others to create SWE at minimal cost.

Who will want Smart Web Exhibits? Museums and libraries will want SWE because they make more of their collections accessible, and because they enhance education by stimulating and rewarding interest, delivering appropriate information, and enabling users to view or print facsimiles of precious items. SWE will help users develop their own appreciation of what is on display in a physical exhibit; users won't have to find somebody to help them if they want more information. Library and museum SWE will support one other by inviting people to visit the other place to see the original specimens and gain different perspectives. Beyond museums and libraries, all Web users will want access to Web sites that organize pre-selected, quality information (documents, videos, images) for rapid navigation and delivery suited to their needs. Web users will gravitate to SWE because of the convenience of information delivered online, on target and on time. If the SWE prototypes successfully model and provide tools to rapidly and cost-effectively construct new SWE, the impact on education and life long learning could be far reaching as smart sites link to smart sites and more machine intelligence is integrated into SWE over time (for example, features to translate documents or automatically generate summaries of document collections).

Technical Knowledge
Carnegie Mellon University Libraries and the School of Computer Science have considerable experience and expertise with digitizing technologies and workflow, including the development of scanning hardware and software and procedures for verifying image quality and metadata. The high-value content of the library and museum SWE will be matched by high-quality digitization to preserve the appearance of the original objects. See, for example, the Web sites for antique books (www.antiquebooks.net), the Heinz Archives (HELIOS Project, diva.library.cmu.edu), and the Hunt Institute for Botanical Documentation (huntbot.andrew.cmu.edu/HIBD/huntinstitute.html). The Principal Investigators and Project Coordinator will follow existing and evolving standards and best practices in their fields. Their commitment to research and education is apparent in their track record of presentations and publications in professional forums. (See the attached resumes of the SWE Principal Investigators and Project Coordinator.)

The SWE project will use existing technologies whenever possible and develop new ones when it's not. Many of the existing technologies that will be used or adapted for the SWE were developed at Carnegie Mellon. For example, technologies developed in the InfoMedia project will be used to digitize video and, if necessary, capture textual transcripts to enable retrieval of video segments. Funded by the National Science Foundation's Digital Library Initiative, the InfoMedia project is expanding their research in the areas of automated summaries and video recognition. The Just-In-Time video lecture technology (www.jitl.cs.cmu.edu) will be used to answer (pre-selected and loaded) frequently asked questions in the museum and library SWE. The syntactic search engine SynSearch, developed by Dr. Thibadeau, will be used to search the SWE content. The prototype Automated Reference Assistant (ARA) software for the SWE is currently being developed at Carnegie Mellon with funding from a private foundation; the two-year ARA project (1999-2000) is another collaboration of the University Libraries and the School of Computer Science.

Personnel
Carnegie Museum of Natural History (CMNH):

Carnegie Mellon School of Computer Science (SCS):

Carnegie Mellon University Libraries:

Contributions
The total cost of the proposed SWE project is $838,334. Carnegie Mellon University is asking for $388,517 from IMLS and is contributing $449,817 or 54% in in-kind and third party contributions. The total equipment cost for the project is $88,807. Carnegie Mellon is contributing new or used equipment valued at $71,307 or 80% of the total equipment cost. Details are provided in the Detailed Budget.

IMLS funding is requested for the salaries of permanent staff. All of the Carnegie Mellon staff engage in research and development efforts as part of their normal duties. For example, the SCS Principal Investigators are research scientists, not teaching faculty. Three of the library staff are in the Library Information Technology R&D department; the University Archivist led the way in digital archives with the HELIOS project. The Carnegie Museum of Natural History may hire temporary staff to perform or assist with some of their regular duties during the grant period.

Budget
Carnegie Mellon University has experienced people, leaders in their fields, who will accomplish the SWE project work more expeditiously than people with less background. The workflow for digitization and the production of user friendly Web interfaces and efficient system architectures will be cost effective, for example, based on experience gained in the TULIP, HELIOS and Universal Library projects. Using available existing equipment, existing technologies (like the Just-In-Time lectures) and developing technologies (like the Automated Reference Assistant) likewise will produce results at minimal cost. In addition, the archival collections have already been processed, which further reduces the cost of the SWE project. For example, the Simon and Newell Collections have been organized in archival folders, finding aids are available, and the audio and video materials have been remastered for preservation and transferred to standard technologies to facilitate use and reduce the cost of digitization. In addition, whenever possible, quality control (such as scratch removal, contrast enhancement and image restoration) will be performed with available software tools rather than using manual preservation techniques.

CMNH is a full partner in the SWE project and will assume responsibility for managing IMLS funds. If CMNH was a subcontractor for Carnegie Mellon, the university's overhead rate of 56% would be applied to the first $25,000 of CMNH project work.

Duration of grant: 24 months

Project start date: October 1, 1999

  Contact Us | Site Map | Comments

  January 23, 2004 -- http://www.library.cmu.edu/Libraries/IMLS/narrative.html
  webmaster@www.library.cmu.edu
  © 2004 Carnegie Mellon Libraries