Gabrielle Michalek
Head of Archives/Digital Library Initiatives
Carnegie Mellon University Libraries

Conversation with Reid Smith (24-Jan-2007)

Question: What are you doing with video today?

• We take in collections associated with important people and deal with whatever comes with each collection. (See Carnegie Mellon University Libraries Digital Collections.) If resources are included, we try to preserve audio and video.
• We apply the following processes to all received materials: organize, describe, and preserve.
• We would like to be able to make an argument for more resources to capture videos.
• Because Computer Science is important to CMU, we go after papers of Newell, Simon, Traub, Kanade, etc.:
  • Allen Newell Collection
  • Herbert Simon Collection
  • Joseph F. Traub Collection
  • Envisioning Robotics - from the Takeo Kanade collection
• We started capturing video in 1992. At that time, the standards for digital video preservation were not in place. Hence, for the Senator John Heinz III Archives, we performed analog remastering of the original videos to Betacam SP. We provide service copies in VHS.

Question: Are you already capturing AI videos of historical interest?

• There is quite a bit of video in the Simon collection. There are also Media Services Collections.

Question: What metadata do you include with the audio/video materials and what software do you use to create it?

• Background: Gabrielle Michalek. Digital Archives: Into the Light. Computer History Museum (August 1, 2002).
• Three types of metadata: descriptive, structural, administrative.
• 1995 - Developed Helios System to digitize, create metadata, OCR, and provide access to a collection. HelioScan was used to generate metadata.
• 2000 - Migrated to DIVA (Digital Information Versatile Archive).

• Encoded Archival Description (EAD) - for archival collections
• Dublin Core - for books, journals, photographs, etc.
• Metadata Encoding and Transmission Standard (METS)
• Today, we create the metadata in an MS Word file, using a tool on top of MS Word that is configured to be EAD compliant.

• The Library R&D department has created some EAD templates which are used as an outline for creating new finding aids. A template is copied into <oXygen/>, then the archive-specific information is filled in, and the new EAD file is saved. The new finding aid is fully EAD-compliant and works with our stylesheets to convert the EAD file into HTML. Details and templates are available at: README for CMU EAD templates. Contact: Chris Kellen, Head of Research & Development, Carnegie Mellon University Libraries.
• (RGS: Other organizations also use XMetal and XMLSpy.)

Question: Names of video services; Prices.

• The Dartmouth Digital Library Group is a good contact for this kind of question. See the Dartmouth Digital Publishing Standards.

Question: Are you serving pages from a database? If so, what database, what code on the server side?

• The Library R&D department built the system (Oracle, jsp).
• Latest release of DIVA: March, 2007. New, key features: EAD compliance, better relevance ranking, persistent URLs. Contact: Chris Kellen.
• "DIVA allows students and researchers to search, browse, view and print digital images of books, technical reports and archival documents. With specifications developed by Carnegie Mellon librarians and archivists, DIVA provides conventional access to library and archival materials, and adds powerful new functions for searching and retrieving documents, supporting multimedia, and customizing the structure and presentation of collections. At present, DIVA is used for more than one million pages of archival materials (Carnegie Mellon’s H. John Heinz III Archives, Allen Newell Collection, and Herbert Simon Collection; and, in cooperation with The Carnegie Museum, the Diplodocus and Douglass Archive)."

Question: If CMU were to act as a regional archive for AAAI videos, what would be the preferred form of delivery from AAAI?

• CMU does not have resources to support such an activity.
• If AAAI is interested in CMU playing a role, the first few questions are:
  1. What will it cost and who will provide the resources?
  2. What are the expectations of donors regarding masters, originals, streaming on the Web, cataloging, providing a copying service, reformatting, cataloging, etc.?
  3. What roles will the various organizations play (e.g., classification, addition of metadata, indexing).?
• While the standards for preservation of video are still evolving, we can make some good choices around formats and metadata now.

Question: What kinds of consumers (and devices) are you trying to serve?

• We try to find a least common denominator in terms of access bandwidth. We are especially concerned about people who do not have fast connections.
• As an example, Verizon is not putting all its eggs in one basket. See: Tony DiMaso, Vice President for Strategy and Development at Verizon Communications. (Verizon Perspectives on) The Emerging Telecommunications Environment and Implications for the User Experience. Digital Libraries Colloquium Services, 12-Oct-2005.
  • PowerPoint Slides
  • Online Presentation Parameters (from Windows Media Player)
    • Title: dmaso_10_12_05; Length: 01:26:02; Bit rate: 338 Kbps; Media type: Video; Video size: 320 x 240; Aspect Ratio: 4:3; Audio Codec: Windows Media V9 Standard 48 kbps, 44 kHz, 16 bit, Mono (CBR); Video Codec: Windows Media Video V9, 320 x 240, 29.96 fps, 302 kbps (CBR); Location: mms://media2.andrew.emd.edu/001/dimaso_10_12_05.wmv.
  • Key Points
    • Digital convergence is coming (Internet, Mobile, Broadband): increasing bandwidth and increasingly capable devices.
    • Verizon is moving from being a connectivity provide to being solutions & platforms provider, primarily integrating content from others (SPOA?).

Question: How have you approached copyright permission?

• Carnegie Mellon University Libraries Projects - listing of all projects
• CMU has a lot of experience with this topic, involving much discussion with University counsel.
• Relevant Projects
  • Copyright Permissions
  • Taking a Stand on current legislation
• Obtaining copyright permission takes a lot of legwork—to capture and maintain it.
• Start with materials that are in the public domain or where you own the rights and the materials do not violate privacy, propriety, etc. (i.e., "we aren't going to get in trouble over this").
• Remain within fair use guidelines.
• Put the info you capture into the administrative metadata.
• Respond quickly to "cease-and-desist" letters.
• For material that cannot be covered by the above, do not publish it. One must go to the physical archive to access it.

Follow-Up Meeting with Bruce Buchanan and Jon Glick, June 15, 2007

• Gabrielle has sent the list of over 400 videotapes, movies & audio tapes that she has been able to collect at CMU. She would welcome help from AAAI subject-matter experts in prioritizing those items with respect to the importance of their contents.
• Chris Kellen, Head of R&D, Library Information Technology, CMU has XML templates for meta-data that the library uses, which he has subsequently sent us.
• University libraries are already cooperating with the Library of Congress on the EAD standard for descriptive meta-data, so asking them all to use that standard is not imposing a hardship.
• We discussed research in a social networking model for low-cost identification, preservation and access of AI-related video materials. The wiki is a start and can be expanded as a funded research project.
• CMU will be an active collaborator in this project and already has digitized some of its wealth of AI material.

Contributions?