Is it worthwhile to develop this collection into a digital project? This is a most question to ask right at the beginning, especially if the library has not yet moved towards planning a digital library and thus has no obvious technical support. The library first needs to assess the value and uniqueness of the collection, and then consider whether the collection is worthy of the extra effort required to make it available on the Internet. And, if so, it needs to seek a grant to support the project. Another important consideration is whether the library can recruit experts from campus and develop a collaborative and working relationship for the project. Collaboration with subject experts is an important component of a project dealing with primary sources in the humanities if it aims to develop a successful "content-based user interface".[2]
Initial interest in the digital project at the Chester Fritz Library emerged from a discussion with Prof. Jim McKenzie, the current Director of the University of North Dakota Writer's Conference, who pointed out the significance and value of the Writer's Conference video collection.[3] The collection includes readings, talks, and panel discussions by many notable contemporary American and international novelists, poets, playwrights, essayists and filmmakers who have participated in the annual UND Writer's Conference over the past 20 years.[4]
In the past two decades, the annual UND Writer's Conference has established itself as a rare public forum for the University and its surrounding community. In more recent years, the Conference has drawn a diverse, educated public from the region and beyond to the campus. Each year, the week long Writer's Conference is organized around a particular theme, issue, or literary movement, and features six to ten writers. During the Conference week, each writer gives a reading and participates in several panel discussions. These sessions have become most interesting forums capturing each participating writer's perspectives on his or her own work and the wider personal and cultural contexts from which that work arises. The Writer's Conference at the University of North Dakota has created a number of rare occasions to observe memorable interactions among participating writers, reminiscent of the eighteenth- and nineteenth-century literary cafe culture.[5} The Writer's Conference archival video collection captures many unique moments--reminiscent of such cafe culture--that may never be recreated.
The collection is a valuable primary resource for literary students and researchers, as well as humanists and social scientists whose interests range from American intellectual history to American social and cultural studies. It is worth the effort to make the materials available on the Internet and thus to a wider audience. Some highlights of the archival collection include a historic 1974 gathering of the Beat poets, entitled, "City Light in North Dakota;" 1976’'s "New Journalism and the Novel" brought, among others, Tom Wolfe and Truman Capote to the campus; and several international Writer's conferences in the 1980s that featured such authors as Alain Robbe-Grillet, Joseph Brodsky, Derek Walcott, and Czeslaw Milosz. The Conference's panel discussions have referred to a wide variety of significant social and cultural events in American history: Watergate, the Vietnam War, and the nuclear arms race, to name a few, over the years. The collection has captured the evolution of various cultural discourses on gender, politics, and ethnicity, especially as they relate to language and cultural representations in literary works. The connection between a Writer's work and the social and cultural milieu from which the work arises is one of the important contextual features of the archival collection.
Aigrain's paper, presented at the 1995 International Symposium of Digital Library, has a theoretical relevance to the UND Writer's Conference project. [6] Aigrain outlined key theoretical and technical issues in the practical development of image and sound digital libraries. He emphasized the importance of "content- based processing and interaction” for sound and digital projects, and directly addressed the key issues as they relate to technical solutions, such as better digital compression methods and standardized content- description codes programming. Aigrain thus delineated needed technical improvements that in turn constitute important considerations for the delivery of "content- based processing and interaction,” as well as "content- based user interface”. Many of the technical issues, however, can be overcome by increasingly user-friendly multimedia development tools appearing in the marketplace, such as Progressive Network, Inc.'s "RealAudio" and Silicon Graphics, Inc.'s comprehensive WebFORCE MediaBase.[7] These pre-packaged multi- media development tools would make the audio delivery of primary humanities sources on the Internet much easier, and increase access by researchers and students to resources otherwise not easily available.[8] The UND Writer's Conference project selected RealAudio as its primary audio delivery mechanism, based on its real- time, on-demand audio streaming technology, its availability to a wide variety of computer platforms, its established popularity on the Internet, and its indexing capability.
The marriage of intellectually dynamic audio content and the current web multi-media technology requires that the project development team pay extra attention to and put significant efforts into the overall research tool design and organization. The issues surrounding content-based interaction or user interface have particular relevance to a project dealing with humanities primary sources. The issues of design and organization have to be addressed in the process of working together with subject experts who will evaluate and analyze the contents of the primary sources.
The vision for the UND Writer's Conference Project is to create a well-organized electronic reference tool that facilitates user's browsing and skimming through the contents by means of abstracts or annotated representations. The project envisioned will not be simply a collection of digitized archival materials, but rather an online reference book, with abstracts, annotations, and explanatory notes that makes use of the audio portion of selected reading and panel session video archives.
At the onset, the project team selected a conference known to be rich in content as its logical starting point to develop the project's "ptotype". The project team, however, started with a very vague understanding of what prototype development was supposed to do. It was rather overwhelming to deal with a large portion of the video collection all at once. Starting with one conference seemed reasonable. The project team thus simply called it "pototype development”".[9] In the process of developing the first prototype, the project team gained a new understanding about prototype development and its role. The project team now sees project development as a useful framework which guides each development stage and within which design and organizational issues are considered and tested. In other words, Prototype development will involve testing and creation of new organizational and design features by considering the project elements, such as audio clips, abstracts, annotations, and explanatory notes, and their relationships to each other.
The project team learned some practical lessons while designing the first prototype. The initial and fundamental lesson was that a digital project requires appropriate facilities and technology on site. At the outset, the project team assumed that developing a prototype using one sample conference would be a relatively easy task. It became immediately obvious, however, that it did not have the appropriate equipment on site to produce even one entire sample conference. A minimum of 1.5 Gb computer disk space was estimated for processing and storing the entire conference sample.[10] Despite the initial readjustment, the project team managed to design the first prototype with support from the National Research Council Canada, Institute for Biodiagnostics.[11] The prototype consists of five web pages, including the main project page, a panel session page, a reading session page, a Writer's session page, and a conference page. These pages are hyperlinked to each other and constitute the basic building block of the project. [12]
With the experience of developing the first prototype, the project team learned to be realistic about each developmental step. For example, its initial objective of making one sample digitized conference was revised, and instead, the project team first established a much smaller unit by digitizing a sample tape from one reading and one panel session. When the first prototype was established, the project team reassessed the next realistic project development step and now plans to continue using contents of much smaller scale-- probably, half a dozen or so authors’ archives--to test the envisioned electronic reference book. The project team foresees the following future prototype development described below. In the next stage, the project team will use the current prototype as a template to process other selected reading and panel sessions. With the template, the team can further edit accompanying abstracts and explanatory notes, and organize indexes for audio files. After establishing this basic ground, the project team will be ready to incorporate additional annotations, explanatory notes and appropriate hyperlinks to mold the prototype into the envisioned reference book. The project team will be able to revisit any changes to the core organization (if there are any) and experiment with additional interface features at this time.
Three web pages--Writer's, reading, and panel session pages--constitute the core building block of the project. A Writer's page will be created for each author, containing a concise biographical note, hyperlinks to reading and panel session pages, a selected bibliography and selected Internet resources. This page is designed to include basic background information on the author. The reading session page includes an abstract, some graphic images providing the visual flavour of the session, and indexed audio clips. The panel session page is organized similarly.
RealAudio files can be indexed by starting and ending time of each segment. For the reading page, the project team chronologically segmented each item read and remarks made by the reader. Each reading item was indexed by title. After analyzing the content, the panel session audio clips were also segmented and each segment annotated, using free form sequential and topical highlighting. Many web projects that utilize RealAudio audio clips organize them sequentially as in the first prototype for the UND Writer's Conference project. For example, the National Public Radio site is a typical news site organizing radio clips sequentially by date of broadcasting and by program within the same date.[13] Another typical sequentially organized audio project is the 1992 Presidential Debate at the Michigan State University.[14] The Presidential Debate and the NPR sites are essentially archival in nature, that is, they faithfully reproduce the material of the actual events: the debates or, in the case of the latter, the newscasts themselves, and other NPR programs. In contrast, the UND Writer's Conference project is not strictly a collection of archival materials, but rather an online reference book. The book will contain contextual items to highlight and support audio clips. The project team currently envisions thematic pages where annotated audio clips will be organized by themes selected. For example, if the theme is "feminism",” the project team will gather all relevant comments made on "feminism" by participating authors over the years. A RealAudio metadata file (.ram) containing a series of audio clips with comments on "feminism" can be created and set up to play the collected comments one after the other. An appropriate annotation to the audio clip can be added to guide the user. This is an example of the kind of interesting feature the project can include by utilizing the UND Writer's Conference materials and the RealAudio indexing capability.
The prototype development with successive developmental stages will continue to address general web layout design and interactive features, possibly using CGI scripts, JAVA applications, a search engine, and other technical mechanisms for the research tool interface. For example, software that allows on screen user annotation would, as pointed out by Aigrain, significantly increase the value of the research tool and allow users flexible browsing according to their research needs, regardless of the pre-arranged organization.[15] In addition, metadata indices will be HTML-encoded in order to ensure wide accessibility of the research tool.[16]
In order to proceed to the next developmental stage, the project team needs to have appropriate equipment and an on-site facility. With the project facility on site, the project team will address quality control methods applied to the audio digitized files using a sound editor. Also at that time, the project team will design some tools that facilitate content development and editing process using simple computer programs.
Recent increased availability of multimedia development tools and systems in the market will make it more feasible for small to medium academic libraries to start their own digital library projects utilizing unique and valuable humanities collections held at their libraries. This paper summarized some lessons learned by developing the first prototype for the UND Writer's Conference Project and outlined possible future steps with content-based user interface in mind.
References:
[1] Most notable digital library projects are usually large in
scale. The American Memory Project at the Library of
Congress, and the NSF/ARPA/NASA Digital Library
Initiative are examples from publicly funded
organizations; and the digital archival system set up at
Simon and Schuster's Higher Education Group, is one
from a privately-funded organization. A short list of
notable digital library projects is available from the web
site:
[2] Aigrain, Philippe, "Image and Sound Digital Libraries Need More Than Storage and Networked Access" p.113. In the paper, Aigrain covered key theoretical and
technical issues in developing image and sound digital projects. He pointed out that "Content-based user
interface for interactive listening and viewing are maybe the most important components of digital libraries of
time-based media".
[3] I would like to thank Prof. McKenzie for sharing with
me the detailed history of the Writer's Conference at the
University of North Dakota and also for providing
timely insights and his knowledge of the intellectual
contents of the archival collection. He has been the
organizer of a number of past UND Writer's Conference
series and also a member of the national writer's conference organization, Writer's Conference and
Festivals. Prof. McKenzie is a collaborator and an
active participant in the UND Writer's Conference
Project.
[4] The UND Writer's Conference was inaugurated in
1970 and the video collection started since its 8th
Conference in 1977. The list of writers and poets who
attended past conferences is available from the author.
[5] This analysis of the panel sessions emerged in the
discussion with Prof. McKenzie.
[6] See no.3. p. 112-117.
[7] Progressive Network, Inc. develops both sound and
video multimedia systems and players. For more details,
see
[8] Many primary sources in humanities are held as
archival collections in libraries and unless they are
deemed to be highly valuable and to have some local or
national significance, many are not even catalogued, nor
listed on their on-line catalogues.
[9] The Writer’'s Conference video archival collection
consists of a series of panel and reading sessions,
covering on average 15 hour-long videotapes for each
conference. The whole video archival collection covers
over 20 conferences. The video archival collection
contains close to 200 reading and 100 panel sessions.
[10] An audio file initially digitized with the processing
speed of 11.025 kHz for a one-hour videotape requires
approximately 100 Mb disk space. When the same file
is compressed to a RealAudio file (.ra), the size reduces
to 7 Mb. Required disk space for a RealAudio server
and digital processing would accumulate substantially.
[11] The project team especially acknowledges Mr. Walter
Roberson, System Administrator at National Research
Council Canada, Institute for Biodiagnostics for his
technical support and for making the Institute’'s Unix
facilities available for the project.
[12] Creating these basic building elements also involve
establishing basic file and directory organization on the
Unix machine and a naming scheme for each file type.
[13] See http://www.npr.org.
[14] See http://web.msu.edu/debate/debate.html.
[15] For example, Paul Lansky developed software called "marksnd”". See the Princeton University Department of
Music web site:
[16] For an example, the Dublin Core Element set has been
developed over the last couple of years; it consists of 15
metadata elements. For more detail information is
available from the Dublin Core home page:
Acknowledgements: The author is grateful to a number
of people who helped her finalize the paper. Prof.
McKenzie, Mrs. Betty Gard, Mrs. Sandy Slater, Mr. Bob
Garret, and Mr. Walter Roberson, all of whom provided
helpful editing comments on her paper. The author
especially acknowledges Mr. Roberson's extensive
technical support on the first prototype development.