2005 Computerworld Honors Program

	Web Media Collective Ohio State University Columbus, Ohio United States Year: 2003 Status: Finalist Category: Media, Arts & Entertainment Nominating Company: Microsoft
	A collaborative effort creates six web-based digital media collections containing almost a million media assets that are used in more than one hundred university courses, dramatically enriching the student experience.
	The Web Media Collective of The Ohio State University: Digital “knowledge artifact” collections for research, teaching and outreach. http://wmc.ohio-state.edu/ The Web Media Collective is a group of faculty, staff and students in Humanities, Arts and Architecture who are working together to find cost-effective ways to make knowledge created at Ohio State available across disciplines and to audiences beyond the university. It was the proposal process of a 1998 internal funding competition that brought us together as we realized we were struggling with the same kinds of problems. We committed to cutting through the red tape, sharing resources and making things work on a grassroots level. The result: six digital media collections containing over 850,000 media assets that will reach over 20,000 students in 105 course sections annually. This has been accomplished on a relatively small budget. Our five-year funding (1998-2003) total for shared infrastructure, application development and support was approximately $260,000. BACKGROUND From the scenes depicted on ancient pottery to the messages communicated in today’s web sites, knowledge has been captured, preserved and passed along to future generations through a variety of media forms. Much of what we want to know about any society can be gained by exploring its media, or, “knowledge artifacts,” and deriving from them a sense of the meaning and values embedded within each one. For example, a colleague once showed me a photograph he had taken doing his research work in rural India. The subject of the photo was a rural house, mud walls, thatched roof, someone’s home. Crossing overhead were power lines, a normal scene in any neighborhood. However, the detail to be seen in the photo was that the power lines did not go to the home, it was without electricity…sitting beneath power lines. This was a photo that represented all rural life in India and the contrast between it and modern society. Modern society, as represented by the power lines, was passing rural life by, almost ignoring its existence. A photo OF a village house. A photo ABOUT the contrast between rural and modern society. Faculty have always relied on knowledge artifacts to help students learn particularly when talking about difficult concepts or exploring new areas. A photograph is not “fluff,” it is an example, or a non-example or it demonstrates critical attributes of something being studied. Audio, video, web sites provide prompts for discussion. Realia, excerpts from mass media programs, photos of another culture’s architecture, are all critical to understanding the world from a variety of perspectives. What prompted our collaboration was witnessing the struggle faculty endured in order to bring all kinds of knowledge artifacts to the teaching/learning setting. They spent incredible amounts of time trying to track down a single image, or fiddling around with a stack of video tapes in the classroom to show just a brief segment off of each one. They were frustrated because they knew some kind of visual example would help their students learn, but they didn’t have the equipment, or the time, or the expertise to actually make it happen. As a result, the Web Media Collective focused its efforts on helping faculty, students, researchers, take the media they use daily to explain or demonstrate their work and get it into a format that can be shared with the important contextual information included. A collection of knowledge artifacts in the form of shared, digital media assets. GUIDING RESEARCH QUESTIONS How can we most effectively capture, preserve and share the knowledge that is embedded within a knowledge artifact knowledge that isn’t necessarily captured at the time of cataloging (i.e., “OF” vs. “ABOUT” as posed by the example above)? How do we preserve knowledge that is important and represented in physical media assets but is not formally published (i.e., a researcher’s life’s work in 20,000 35mm slides)? How can we make this knowledge easy to find without adding more clutter to the current glut of information in our society? How can we integrate these solutions into existing budgets, workflows, and national and international standards, and how can we adapt existing technologies to new uses? PRIMARY GOALS To create collections of quality media assets for teaching, research or outreach that: 1. directly support the mission of Ohio State 2. are based on international standards so that the collections can be searched/found by anyone anywhere 3. are affordable to develop and maintain 4. can be accessed anytime anywhere 5. are easy to use and are useful 6. are clear of copyright problems 7. can be evaluated for return on investment MEMBER PROJECTS The Harvey Goldberg Program for Excellence in Teaching http://goldberg.history.ohio-state.edu/ The History Multimedia Database contains over 3,000 images, recorded speeches, animated maps, and video clips that cover a broad range of historical topics and periods. This was the first Web Media Collective project. Much of the collection was converted to a live web database from its ImageAXS Pro predecessor requiring significant data cleanup. It was through this project that we proved our shared infrastructure could work and that the collaboration was worth the time and investment. The Charles Csuri Archive http://wmc.ohio-state.edu/csuri/ Charles Csuri is an artist, a professor, but most importantly, a pioneer in the field of computer art and animation. In the mid-1960s, Csuri began experimenting with computers as an artistic medium and has made many important contributions to the field. In fact, one of Charles Csuri's computer films is in the collection of New York's Museum of Modern Art. In subsequent years, Csuri developed a large graduate research program that ultimately informed the development of flight simulators, computer-aided design, visualization of scientific phenomena , magnetic resonance imaging, education for the deaf, architecture, and special effects for television and films. Graduates from his program are employees of Industrial Light and Magic, Pacific Data Images, Metro Light, Pixar, Rezn8, Silicon Graphics Inc., USA Today, Rhythm and Hues, Xaos, Walt Disney Productions and others. His former students have worked on such films as Star Wars, Terminator 2, Lawnmover Man, Jurrasic Park, Casper, and Toy Story. The Charles Csuri Archive is a collection of still and animated works that the university has been given permission to use for education, research and outreach purposes. This was the second Web Media Collective project, one that built upon the successes and best practices of the History database. One unique aspect of this collection is that we designed an interface for Charles Csuri to add his own comments to the works. An excerpt from the collection is below: Title: Ribbon Romp Primary Artist: Charles Csuri Media Type: Algorithmic Paintings Year: 1998 Algorithms Used: Ribbon tool Artist's Note: The function to create ribbon like lines was used. Once again, the same figure is used but in different positions and orientations. I simplified the figure before I applied the ribbon function. I jumped back and forth with the thickness of the ribbon before I made up my mind... ...My young granddaughter Hannah was the influence on this piece. I love her work and I just wanted to see if I could achieve something of the feel of her work. For more information about Charles Csuri, please visit the SIGGRAPH site at: http://www.siggraph.org/artdesign/profile/csuri/ The John C. and Susan L. Huntington Archive of Buddhist and Related Art http://kaladarshan.arts.ohio-state.edu/default.html The John C. and Susan L. Huntington Photographic Archive of Buddhist and Related Art contains nearly 300,000 original color slides and black and white and color photographs of art and architecture throughout Asia. Countries covered in the collection include India, Afghanistan, Pakistan, Bangladesh, Sri Lanka, Nepal, China, Japan, Thailand, Indonesia, and Myanmar (Burma). Works range from approximately 2500 B.C.E. to the present, and documentation includes contemporary religious activities in various parts of Asia. The Archive documents the art and architecture of these countries in situ, as well as works of art found in most major Asian, European, and American museums. This broad, yet detailed, collection contains predominantly Buddhist material, but also includes Hindu, Jain, Islamic, and other works. Many of the historical sites and works of art photographed by the Huntingtons have been lost, stolen or destroyed, leaving the Archive’s slides as the only source of documentation. Such is the case with the world’s largest Buddah statues (150-175 ft tall) that, until very recently, stood in the Bamiyan Valley region of Afghanistan. In addition to being the most comprehensive collection of its kind, The Huntington Archive includes the largest photographic archive of Nepali art and architecture in the world and represents the only formal collection that photographically records this country's artistic heritage. The Huntington Archive represents the efforts of over twenty-five years of field documentation photography by John and Susan Huntington, professors of Asian Art History at The Ohio State University. Taking this vast resource and converting it into an on-line database and image delivery system was no small task. This was the third project tackled by the Web Media Collective, one that provided us with the opportunity to explore more complex data structures, national and international data standards, controlled vocabularies and the vast amounts of information about each image housed in the archive. Of significance was converting the cataloging information from flat-file 4D databases into a relational system. The site was released in December, 2002, with a public interface to be available in 2003. The History of Art Visual Resources Library http://www.history-of-art.ohio-state.edu/pages/vrl/index.html The VRL contains over one-half million images, making it one of the largest such facilities in the United States. Over 325,000 slides, 215,000 study photographs, 400 VHS tapes, 150 16mm films and over 10,000 digital images representing the history of world art, archeology, and architecture from prehistoric through contemporary periods are housed in the VRL. The VRL is also responsible for designing and maintaining OSIRIS which allows History of Art faculty to post study images, in digital format, to a secure site for OSU students to access. Between 8,000-10,000 new items (slides, photographs, videos, digital files) are added to the collections each year. The VRL circulates approximately 90,000 items annually. The majority, 85% of these items, meet the daily teaching needs of the History of Art faculty and teaching associates. The other 15% circulate to faculty from a variety of departments campus-wide. As the fourth Web Media Collective project, this one takes our understanding a step beyond relational data structures and the common needs of cataloging a collection. The focus for this application is on instructional needs, particularly the ability to compare and contrast images as is currently done with physical slides, and the ability to zoom into details as questions arise in class. Before embarking on this project, we added an image server (TrueSpectra Image Server) to our infrastructure. This was the best decision we’ve made to date. The zoom capabilities have been a hit, and the file size/image quality concerns have been resolved. Knowlton School of Architecture Visual Resources Library The Visual Resources Library's state-of-the-art digital images collection supports the Architecture, Landscape Architecture, and City and Regional Planning curricula in the Knowlton School. Through online resources, faculty and students are able to access high quality images and multimedia files for use in research, study, or course work. The Library's exemplary teaching and research collections will help to establish the School as an international resource for the study of design and planning. This is the fifth project of the Web Media Collective and, at the time of writing this case study, is still under development. The primary challenge here is to attempt to create one robust infrastructure that can support the knowledge artifacts of three different but related disciplines: Architecture, Landscape Architecture and City and Regional Planning. We have already learned many lessons from this project, lessons that will force us to stop and reflect deeply on our experiences thus far before moving ahead. The Humanities Digital Library The sixth and final project seeks to bring together knowledge artifacts from a broad array of disciplines in the College of Humanities. The College consists of 22 departments and centers, teaches over 29 foreign languages, and enrolls approximately 22,000 students per quarter (that’s per quarter – some 80,000 students annually!). With the previous five projects, the Web Media Collective has refined and streamlined the crucial cataloging process. In this project, however, we are going to try to capture metadata a bit differently, in small “chunks” based on intersections of use, user and the asset itself. That is to say, we will try an approach that “pecks away” at the metadata rather than trying to capture it all in one cataloging form. This project is in the early stages of prototyping and research at the time of this writing.
	The World Wide Web makes it possible to develop, manage and deliver digital media collections that are accessible, searchable, easy to administer, and flexible to meet a variety of needs. A collaborative approach to developing on-line multimedia collections makes taking advantage of this medium feasible. The Web Media Collective has demonstrated that such a collaborative approach works. Faculty and students using the collections have demonstrated that our efforts are well received. In fact, students enrolled in the American History Survey at Ohio State have explicitly commented on how paintings, political speeches, statistics, dynamic maps, household objects and other knowledge artifacts used in class brought them closer to the subject. Other students have credited History’s instructional approach with giving them learning strategies they will use throughout their lives. Below are some of the key benefits of our collaborative approach to developing and delivering shared collections of knowledge artifacts: 1. Appropriate Content Faculty spend years collecting knowledge artifacts as a part of their teaching and research work. To this end, they have gathered what is crucial for instructional purposes; their collections go beyond what one might find in a commercial slide collection. For example, in the Huntington Archive there are photographs of buildings that have either been demolished or been left to deteriorate over time. In some cases the photographs taken by John and Susan Huntington are the only detailed record left of a building or significant cultural site. In the Department of History, Dr. K. Austin Kerr manages a collection of important political cartoons from the Prohibition Era. His electronic version of the collection is the only contact many will have with these valuable historical documents. 2. Ownership of the Content Because faculty either own a majority of the knowledge artifacts they use or have documented the sources, departments and colleges have a better chance of negotiating usage rights. 3. Support of Cross-disciplinary Approaches The ability to search knowledge artifacts from a variety of disciplines at the same time often results in making discoveries about one's own field that would not have been apparent otherwise. It is important, then, to develop collections of knowledge artifacts that can be cross-referenced or searched simultaneously. 4. One artifact, many uses Once a collection of digital media assets is organized in a database, it becomes possible for different groups to point to individual assets for different purposes. This has been the overwhelming benefit of employing image server technology—collection managers only have to maintain one file, yet faculty, students and the larger community may reference that file at a size that fits their needs. Impact Perhaps the most overwhelming benefit of building web-accessible collections of media assets that represent our very knowledge is in the transformation of those most responsible for creating and sharing it. When faculty have access to collections of quality, relevant media assets, they USE them and they want more! There are two primary reasons for this. First, for many this is the first time they have seen their areas of research from a new perspective – a multimedia perspective. A professor of Asian history remarked that his use of photographs in teaching permanently changed his approach to research and he has gained significant new insights as a result. Second, many faculty and students have remarked that the use of images, animated maps, audio and video in class helped them learn the subject. This is supported by current research in educational theory about visual versus aural delivery of information and learning. New Challenges for Society INFORMATION GLUT Information glut is a problem that already has many of us skimming through media as quickly as possible focusing on only those details that jump out as important. The consequence of this behavior is that we miss the “sub-texts” of media we are skimming, and we shut ourselves off from finding those knowledge artifacts that we DO want. Throughout our work we have struggled with the notion of “formality” in a collection. Most traditional libraries (and there are some stellar exceptions) do not meet the need of managing or housing collections of non-print media. Over time, groups and discipline areas (History of Art, for example) have established their own ways to meet this need, and these groups are well known by their constituents for providing quality, reliable, and relevant media assets – knowledge artifacts. Where do we draw the line between that which qualifies as collection-worthy and that which does not? Is an individual’s life’s research work—his or her knowledge gained—worthy of being preserved and shared if it contains, say, photos that are a bit blurry yet prove a point exceedingly well? This vast divide between “formal” and “informal” is analogous to the gulf that separated traditional libraries from organized media archives many decades ago. And, it is this divide that is now adding to the information glut. That is to say, researchers who have something to share yet are not part of any formal collection end up creating a number of one-off web sites that are used by only a select few who know they exist. Because they include little or no metadata and cannot be found very effectively, they ultimately end up adding more bulk to the sea of information. METADATA This is where metadata and search engine technology come into play. The more we can learn about concept mapping, thesauri, metadata harvesting, and the notion of the “Semantic Web,” the better we will be able to find important and relevant knowledge. But cataloging requires skilled people and lots of time, and it limits the relevance of a particular knowledge artifact to the pre-conceived ideas of the cataloger. What we hope to test with the Humanities Digital Library project is the notion of “pecking away” at metadata through intersections of the use situation (an American History Survey lecture or a Women’s Studies graduate course or an advertisement for a product/service, etc.), factual information about the knowledge artifact itself, and characteristics of the individual making use of the media asset. If it is possible to “follow the asset,” (or to have the asset report back in some way) and determine where and how it has been used, then it is possible to add layers of metadata in each use case. Over time these rich layers of metadata will make knowledge artifacts much easier to find by those who seek them, without placing unnecessary burdens on existing budgets and workflows. “BRIDGE PEOPLE” One important part of addressing this problem is what we call “bridge people.” These are people who have both sophisticated technical knowledge AND discipline-specific knowledge. They help individuals with knowledge artifacts to share take that first step. In our case, they take what faculty say is important about their media collections or how a faculty would approach a search for a particular kind of “example,” and convert this into data models, controlled vocabularies and business rules. They are part programmer, part librarian, part discipline-area researcher, part next-generation media consumer. There is no formal training program that prepares one for a career in this area.
	How did IT contribute to this project? Without Information Technology, the Web Media Collective would not exist. When we began this project, our primary goal was to share the technology side of media asset collection delivery. This included servers, server software, custom programming, backup systems, etc. Of primary importance to our work are: database technology, dynamic web applications, image and other media server technology, robust server infrastructure, digital rights management, backup systems, authentication systems, and off-line tools for preparing media assets. All six collections reside on three servers: a production web server, an image server and a database server. We anticipate this infrastructure will grow significantly as more items from each collection are digitized and added to the on-line system. Why was IT particularly important? Out of all of the advantages information technology provides to a project like this, there are a couple that are worth highlighting. First is relational database technology. We designed these systems to be maintainable by collection staff rather than technology staff. If controlled vocabulary changes, collection managers have the ability to edit entries that appear as controlled vocabulary options on cataloging forms. In the Huntington Archive application this is particularly important. The concept of historical period is a difficult one to resolve when trying to classify works from Nepal, India and other countries of the region. As scholars’ understanding of the complex historical and cultural relationships evolves, so will the cataloging system without the need for re-programming. Second is image server technology. One of the best investments we made was in the TrueSpectra image server (www.truespectra.com) because it gave us a broad range of image display options while significantly simplifying the image asset management process. Now, instead of maintaining various resolutions and sizes of image assets, we only need one file and it is from this one file that thumbnails, low-res study copies, hi-res research copies and zoomable detailed views are generated all on-the-fly. This technology has recently become much more affordable and we look forward to similar solutions for other types of media. Throughout this project we have focused on adapting existing technologies to our uses rather than developing things from scratch. The image server is a good example of this. The product we are using was originally developed for on-line retailers and e-commerce applications.
	Creating a digital repository of media assets is not a new thing. There are many superb examples of digital media collections, both public and private, that informed our work. What makes this project unique has more to do with the workflow and administrative problems we’ve tackled rather than technology, and the fact that we’ve accomplished so much on a relatively small budget. The collections themselves are unique, as outlined above. The Csuri Archive captures a single artist’s and researcher’s accomplishments with a new medium over time. The Huntington Archive contains photographic documentation of human history that, in some cases, is now the only such documentation in existence. The History Multimedia Database touches thousands of students each year by enabling faculty to bring knowledge artifacts to the classroom for investigation and consideration. Studying a piece of Shaker furniture says a lot about the Shakers and their place in American history. Photographs CAD drawings and other resources in the Knowlton School of Architecture collection document the world’s planned spaces in ways that help us understand the human experience, something that is also captured in the History of Art Visual Resources Library. Finally, a wealth of knowledge about ourselves, our histories, our cultures is a part of every faculty member’s private collection in the Humanities, knowledge that we will soon be able to share with those who seek it.
	Since 1998, we have accomplished what we set out to accomplish and then some. The History Multimedia Database, Csuri Archive and Huntington Archive are all fully operational. The History of Art Visual Resources Library and the Knowlton School of Architecture Collection will both be released in 2003. What we have learned from these projects has made us change our approach with the Humanities Digital Library application, one that will work from “informal” to “formal” rather than the other way around. What is most promising is the change we see in our faulty and their students by making these knowledge artifacts available to them. Faculty are now able to take teaching and learning to a new level by showing the message rather than just describing it. Students gain new ways of understanding the world in which they live by seeing the layers of meaning embedded in things one would normally take for granted. In an information society, these collections and their use in the classroom not only help communicate a subject matter more clearly, they teach the next generation how to be responsible information consumers and to probe for the layers of meaning in the knowledge artifacts they encounter. Future plans include enabling cross-collection search capabilities, planning for shared infrastructure growth and expanding the number of projects and collections participating in the Web Media Collective. With a solid and broad core of knowledge artifacts we can begin tackling the difficult metadata issues associated with concept mapping and linking, multiple languages, tighter digital rights management and tracking and expanded user authentication. LONG-TERM GOALS • Register collections with state, national, international clearinghouses • Start competition to build media assets course-by-course • Solicit projects that have external funding • Integrate search with local, state and national library systems • Follow the assets and the users: What’s being used the most? By whom? How? What kinds of knowledge objects are being generated? What is needed but not currently part of the collections? • Publish case studies, data models, workshops, workbooks, and other documentation that details what we’ve learned and best practices we follow. • Devise business model for long-term sustainability, one that supports a research mission as well.
	Lessons Learned Work in narrow-and-deep mode rather than wide-and-shallow mode. Staying focused and getting something accomplished that might have to be remodeled down the road costs less in the long run than trying to take into account every possible usage scenario or type of asset upfront and getting stuck in analysis paralysis. It is important to consider the broader context in the early planning stages, but focus, focus, focus when moving into design and production. Always build ways to “follow the asset” and “follow the user” into a media management system and periodically review the logs to determine future directions. Be nimble, flexible and non-bureaucratic if you want to get things accomplished. This is why we moved the oversight to Humanities from OIT. The grassroots level has been extremely productive and it would be best to keep things operating on this level if possible. Build trust. The most difficult thing to share, difficult from an emotional or administrative perspective rather than a technical one, was media file storage space. Individuals are attached to their knowledge artifacts and it is difficult to let them go. What surprised us the most was the realization that it was more important to share workflow processes, metadata strategies, controlled vocabularies and tracking capabilities than it was to share a server infrastructure. Content must drive the technology infrastructure. Trying to squeeze rich metadata about content into a “standard” metadata core dilutes it and makes it less accessible to those trying to FIND it. However, it is important to use a metadata set as the base—you can always “boil down” back to the core. Thank you for taking the time to read this summary of the Web Media Collective project at The Ohio State University. Please visit us on the web at http://wmc.ohio-state.edu/