Moderator: Linda Bien, Slide Library, Concordia University
Sponsors: Visual Resource Division and Reference Section
The first session of the 1999 Vancouver conference opened with announcements of moderator and speaker changes by Margaret Webster, including that she would stand in for Linda Bien as moderator and that Howard Besser and Maria Odall were unable to attend the conference. Maria’s paper would be read by Heidi Hass.
Webster prefaced her introduction of the session by recounting the high and low points of the online horizon of the past year. The VISION project was the high point of the year, with 32 institutions contributing records to a testbed database sponsored by the Visual Resources Association, the Research Libraries Group and the Getty Information Institute. The low point was the announcement that the Getty Information Institute would be dissolved and the vocabularies moved to the Getty Research Institute. Margaret stressed that the work of the GII was of immense importance to the visual resources community and that the membership must let the Getty know of the tremendous value of its projects in moving to a more collaborative environment. With all of the changes occurring on the digital horizon, subject access to image collections as remained elusive. Subject analysis has been avoided due to local concerns about interpretive language as well as just the time consuming nature of the activity. With new technologies such as powerful search engines, our collections are opened up to a wider audience and it becomes necessary to provide information about subject matter. Best of all, the experience of the Vision project shows that this information can be shared among visual resources professionals.
The first speaker on the panel was Ben Kessler, Librarian of the Slide and Photography Collection at Princeton University. Kessler began with general observations and a statement that he did not employ subject cataloging, as such, at Princeton. Much of the information that can be identified as subject analysis, is to be found in other parts of his image cataloging, especially the title field. His assertion is that the job of the visual resources professional is not to recreate thesauri or analyze subject matter. We should top into the work of others whose job it is to create these tools. For many collections, cataloging is worked on a triage basis and little time can be allocated to determining subject content. Tools such as the Union List of Artist Names, The Art and Architecture Thesaurus and the Thesaurus of Geographic Names (all products of the late Getty Information Institute) provide specific information useful for describing images. The terminology of subject indexing, however, is nebulous and often, subject determination is not always exact. Kessler stresses that this is not a dismissal of subject indexing, he is merely stating that it carries a high overhead of cost and time and needs an efficient method. He suggests that we try to learn from what others have achieved, such as the Time-Life Photo Archive, Corbis, the interactive museum kiosks and the AMICO consortium. The extent of subject access of each of these varies from extensive to very basic, but they are models from which we can learn. Other efforts such as ICONCLASS, the Index to Christian Art, the Marburger Index and the Witt Index provide wonderful information and have become much more user friendly with changes in technology. These changes in technology have made it much easier for all of us to pool our efforts and work together in a project such as VISION. In the past ten years, it has become clear to the visual resources community that the only feasible way to define protocols for cataloging images in a shared environment, we must define them ourselves. From this decision was born the VRA Core Categories which saw image cataloging in terms of two types of information: information about works of art and the visual document of the work of art. VISION was a one time experiment to show us what shared cataloging would look like. It was gratifying, but pointed out that there is still a long way to go. It also proved that just having a term in the subject field was not going to be enough.
Elizabeth Bellas, Manager of Information Processing at CORBIS, was the second speaker on the panel and spoke on the way the subject access was addressed for the commercial market. CORBIS is a commercial collection consisting of over 25 million images, 1.4 million of which are available on-line. Corbis’ customers range from large, main-stream publishers to individuals and the corporations goals for indexing are to provide access to the images and provide information about the image. Bellas described the four different types of information that are used to describe images in the collection: factual, depicted information, contextual information, and conceptual information. Corbis uses a controlled vocabulary and a hierarchical database that validates information against the controlled vocabulary. Since it is a commercial archives, common language is preferred, although some specialist language is sometime used. At the present time, the vocabulary uses American English as a standard language, but there is an interest in making the vocabulary multi-lingual. Personal names are not held in the vocabulary, but are contained in a separate authority list. As Corbis grows and acquires new collections, additions must be made to the vocabulary, often trying to map terms from the new collections into the Corbis vocabulary. The vocabulary often holds only the most narrow term for each concept because their early research showed that people most often search for specific items and not for abstract concepts. This may be changing as more people use the archive because there seems to be an increase in abstract queries and it may be time for a new study to be undertaken. Corbis is also looking at the difference in the way different consumers use the database. Some may be only be browsing the collection while others are searching for something specific. This has led to the development of a new search engine for 1500 images called Picture-It, which allows the viewer to browse certain categories. Their initial findings are that people browse more than they search, but this may change over time.
Heidi Hass read a paper prepared by Maria Odall entitled, Varietas delecata: the pleasures and pitfalls of multiple subject thesauri. Odall began with recounting an discussion with Linda Bien at the 1996 Miami Beach conference. She recalled telling Linda Bien that if she could find good, detailed records of Morgan Library miniatures in visual resources collections, she would be happy to derive, clone, or copy those records for use in the Morgan Library. Finally in 1998, she was doing just that as a participant in the VISION project. As she worked with visual resources catalogers, she realized that each collection was an individual, and that there was a stunning need for authority control.
Museum of Fine Arts, Houston