Shaping Up the Web: Subject Access and Other Technical Services Fitness Tips
Spring Meeting of the New England Technical Services Librarians (NETSL)
Friday, April 20, 2001
The College of the Holy Cross, Worcester, Massachusetts

Breakout sessions:
Pete Merrill-Oldham (Acme Bookbinding): Digitizing Books for Preservation
Karl Fattig (Bowdoin College) and John Harrison (Bates College): Vendor-Supplied Bibliographic Records
Betsy Like (NELINET): Search Engine Fitness
Amy Benson (NELINET): Maintaining URLs

Lunchtime speaker: Patricia Oyler, Professor, Graduate School of Library & Information Science, Simmons College:
IFLA: Technical Services at the Boston Meeting

Afternoon speaker: Karen G. Schneider, Director of Technology, Shenendehowa Public Library:
Subject Access Denied: Internet Filtering and the Future

Lois Mai Chan, School of Library and Information Science, University of Kentucky, Lexington:
Subject Organization of Web Resources

One of the topics generating the most discussion in librarianship in recent times is the profession's relationship to electronic resources, particularly those available via the World Wide Web. Within technical services, there are many experiments in progress for organizing and providing appropriate access to these materials. Their specific characteristics - including instability, volatility, rapid proliferation and lack of self-containment - imply differences in treatment from that given to traditional resources contained in tangible packages. At the same time, however, electronic materials still express human thought and creative effort, as tangible items do, with the result that a complete reinvention of bibliographic control methods is neither called for nor justifiable. In her address to NETSL's 2001 spring conference, Lois Mai Chan addressed a range of issues pertinent to subject analysis and classification of web resources, and demonstrated Knowledge Class, a current research effort designed to integrate the best of traditional and newer approaches.

Prof. Chan began by outlining the functions of subject access tools: to assist searchers in identifying paths to materials on a subject, to focus searches, to enable optimal recall and precision, and to do all of this in an efficient, economical manner. Schemes for supplying subject data for electronic resources have a number of requirements, if they are to be optimized for a broad range of users. They should be simple and logical; easy to apply and to maintain; intuitive, so that sophisticated training is not required for implementation; flexible, adaptable, and scalable. In addition, they should be appropriate to specific disciplines and subjects, and to their domains of implementation (e.g., museums, libraries, individual use). At a higher level of sophistication, subject data schemes should be interoperable across disciplines and in various knowledge discovery and access environments. This last requirement involves cross-linking and mapping across vocabularies and languages.

Turning to classification, Prof. Chan contrasted the traditional roles of classification data with potential expanded functions in an electronic environment. Established roles include classification as a shelf-location device (especially in the United States) and as an organizing tool in classified catalogs and bibliographies. Classification schemes are likely to become increasingly important, however, as providing access points for cataloging/metadata records, organization for subject browsing and navigation in portals to electronic resources, and switching languages between vocabularies in multilingual environments.

Prof. Chan next described two recent initiatives with emphases on organization of Web resources: the Library of Congress Bicentennial Conference on Bibliographic Control for the New Millennium, and the work of two subcommittees of ALA's Subject Analysis Committee. Discussions during Library of Congress conference, held in November 2000, highlighted a number of themes. These included developing a tiered approach to bibliographic control of network resources, in order to explicitly recognize their unequal nature while reserving the greatest concentration of professional effort for materials judged to be of greatest value and permanence. As part of this approach, it will be necessary to develop and integrate multiple standards and schemes for use with different types of resources. Bibliographic control techniques for networked resources, including schemes for subject analysis and classification, should also be scalable, extensible, and interoperable.

The Subject Analysis Committee (SAC) is a standing body of the Cataloging and Classification Section, Association for Library Collections and Technical Services (ALCTS), American Library Association. The Subcommittee on Metadata and Subject Analysis began its work in 1997. The group's initial deliberations focussed on the Dublin Core, but its focus later broadened to include metadata schemes in general. The issues of vocabulary, semantics and syntax; application of subject analysis schemes in metadata records; and recommendations for systems design were among the major topics addressed. The subcommittee held discussion forums and programs at ALA conferences, and produced two papers: Subject Data in the Metadata Record: Recommendations and Rationale and Comparison of Subject Treatment of Metadata Standards.

The related SAC Subcommittee on Metadata and Classification began its work in 1998 and presented its final report in 1999. The subcommittee studied portals, such as the Scout Report Signpost, Internet Public Library, and BUBL, which use existing classification schemes to organize resources. These sites were compared with regard to their implementation of multiple classification functions, including browsing, hierarchical movement, identification, retrieval, limiting, partitioning, and profiling. Dr. Chan showed examples of sites which incorporated hierarchical structures in a variety of ways. The major variables were hierarchical depth (or number of levels), and the types of categories employed - frequently either academic subject based or more populist. Advantages of incorporating hierarchical structures include improved subject browsing facility, the possibility of multilingual access, and the ability for a site to serve as a conceptual map of either a single area of an entire domain of knowledge.

An important characteristic of classification in the Web environment is the separation of indexing data from the material indexed. Traditionally, classification numbers are embedded in sources: call numbers are affixed to spines, or printed in surrogate records, for example. In the Web environment, classification data can be separate from the resources involved, attached by links. This allows for greater flexibility in the use of classification, since the links only have to be changed or moved if the scheme changes. Maintenance of resources is minimized, since reclassification does not affect the resource itself. Further, the classification scheme can be adjusted for different user groups, with multiple schemes or implementations in play simultaneously. A good example of this is Yahoo's regional editions, where, for example, the religion page is different for Singapore, Ireland, and Hong Kong.

Organizers of Web resources currently develop new schemes based on perceived needs of their users. The familiar schemes used in the library mainstream represent a top-down approach to classification, applying schemes which cover the universe of knowledge or an entire discipline. An alternative, bottom-up approach is seen at many sites, where specific resources are organized first into clusters or microcosms, converging at different levels into a macrocosm represented by the overall site. There is, at present, no definite answer to the question of which approach is more likely to be effective overall. The top-down method may be best for established fields, with bottom-up method best for personalized organization of resources, or emerging areas in flux.

Finally, Dr. Chan demonstrated the Knowledge Class project, developed with Xia Lin, formerly on the faculty of the School of Library and Information Science, University of Kentucky, and now at the College of Information Science and Technology, Drexel University. Knowledge Class is described at its site as "a customized Web organizer and retriever tailored to individual users' interests and needs. ... Knowledge Class has been designed to help individual users create and maintain personalized maps of promising resources meeting their needs and interests. For use as a resource management tool, Knowledge Class organizes related topics within a specific domain in a logical structure. As a tool for resources discovery and retrieval, it combines the capabilities of subject guides and search engines to automatically construct search queries in a way to include not only the terms the user has selected but also other related terms and broader terms in the hierarchy." The project employs a powerful combination of explicitly mapped domains of knowledge with both "hard links" to quality resources in those domains, and "soft links" to initiate searches in a variety of search engines, such as Northern Light or Google. The current version of Knowledge Class is available at http://faculty.cis.drexel..edu/kclass/. A new version of Knowledge Class with additional features is currently in progress. The team is seeking participants to create individual knowledge classes, and the application may be downloaded and customized for local use. For more information or to participate, contact Lois Mai Chan at loischan+NELAsecure+uky.edu or Xia Lin at xlin+NELAsecure+drexel.edu.

Related article: Lin, Xia, and Lois Mai Chan. (1999). Personalized knowledge organization and access for the Web. Library & Information Science Research 21(2):153-172.

Also see Dr. Chan's presentation for the Library's Congress' Bicentennial Conference on Bibliographic Control for the New Millennium: Exploiting LCSH, LCC, and DDC to Retrieve Networked Resources.

Patricia Oyler, Graduate School of Library & Information Science, Simmons College:
IFLA: Technical Services at the Boston Meeting

Prof. Patricia G. Oyler, of the Graduate School of Library and Information Science, Simmons College, spoke to NETSL attendees during the lunch period about the 67th Council and General Conference of IFLA, the International Federation of Library Associations and Institutions, to be held August 17-25, 2001, at the Hynes Convention Center in Boston, Mass. The talk was intended both to whet the audience's appetite to attend, and to inspire service as an IFLA volunteer. This was the first in a projected series of brief presentations at NETSL's Spring Meetings, to be focussed on the work of other library and information science-related organizations.

Prof. Oyler shared her enthusiasm for IFLA and told the audience the story of how she began attending. She has always loved activities with an international focus, including traveling abroad, and studies everything from an international perspective. Her doctoral dissertation, for example, was on the subject of interlibrary cooperation in Sweden. This interest eventually led her to investigate IFLA, first attending the 1982 conference in Montreal. She outlined the meetings and programs to be presented in Boston which will be of particular interest to technical services librarians, and described the social opportunities available, including receptions in various venues and the chance to meet colleagues from 128 countries. One hundred librarians from developing countries will be hosted; Simmons is hosting Vietnamese library school graduates.

Karen G. Schneider, Shenendehowa Public Library (Clifton Park, NY):
Subject Access Denied: Internet Filtering and the Future

Karen Schneider's afternoon talk centered on two topics: understanding filtering, with particular emphasis on CIPA (the Children's Internet Protection Act), and understanding the assaults on our current professional environment. Using humor and sharp commentary, she bound the two topics together to provide insights into the political forces which threaten coherent access to Web resources, as discussed by Lois Mai Chan.

Ms. Schneider began by engaging the audience with Ranganathan's Five Laws of Library Science, asking people to call out each law in order. In this way, the audience restated for itself some of our basic principles, a core of practice under attack in certain quarters. She then summarized a set of pivotal censorship cases in the history of librarianship, from the in-house purges of novels in the 19th century, through the FBI's Library Awareness Program in the 1980s, to today's challenges, including CIPA and its Tweedledee brother, COPA (the Child Online Protection Act). We are witnessing dramatically changing perspectives by sectors of the general public regarding libraries and librarians. Where before, libraries were thought of as local safe harbors, places for finding books (or "one-trick ponies"), with predictable content, many regard libraries now as potentially dangerous places where anything goes. Libraries have become a battlefield for Beltway politics, to a substantial extent because of the "e-rate," the federal program for subsidized school and library networking. After the passage of this legislation, libraries were suddenly not everyone's best friend in Washington.

Ms. Schneider reviewed the basics of CIPA. Schools and libraries that receive federal funding, via the e-rate program or the Library Services and Technology Act (LSTA), will be required to certify they are using "technology protection measures" to block or filter Internet access to visual depictions that are obscene, include child pornography, or are "harmful to minors." The law applies to adult access, including library staff at their workstations, as well as to access by children. Libraries receiving the funding in question must adopt Internet policies focused on minors, hold a public hearing on their policies, and certify that they have implementated these policies and the use of "technology protection measures." CIPA will have a signficant impact on the many library consortia that are heavily dependent on e-rate funding, and on libraries in poorer districts, which receive discounts of up to 90%.

So, since nobody is advocating for child abuse or the spread of child pornography in libraries, what would happen if web filtering software was broadly implemented - more broadly than it is already - for purposes of CIPA compliance? Ms. Schneider has done considerable research into the details of how filters are constructed and how they operate, beginning with her work on the Internet Filter Assessment Project. A typical scenario involves college students, working part-time for low wages, skimming thousands of websites. They are instructed to decide which sites contain "pornography," for example, and then place the addresses in an encrypted database. The process begins, in other words, with individuals, working nonprofessionally, making subjective judgments with no direct accountability to the companies' customers. This is compounded by the incoherence of the taxonomies involved. Examples of actual filter categories include Job Search, Cult/New Age, Homosexuality, Drugs, Travel, Abortion, Email, Racism, Religion, Sports, Tasteless, Chat, Vehicles, Nudity/Sexuality, and "Other." In consequence, it is not unusual to discover examples of sites, such as the New Jersey TV Guide, blocked for no evident reason.

At this point, Ms. Schneider began making explicit the connections between the drive for mandated filtering and the assaults on professional librarianship, and technical services in particular. Where librarianship has developed detailed practices for providing specific services to real communities, CIPA and filtering software impose a "shoebox fit for community values," the very opposite of MyLibrary or other customizable user interfaces. Filters not only demonstrate categorical incoherence, they "organize the Internet by denial," imposing invisible taxonomies of resources that can't be found. The imposition of filters outsources site selection to nonprofessional third parties, and also provides "weeding for the fin de siècle," as if items had been cleaned out of the collection and removed from the catalog, with no notes remaining to record what was removed. Where the traditional technical services are built on the bases of certified expertise, and highly articulated standards, applied onsite with the intended outcome of high quality access, filtering ensures none of these.

Ms. Schneider referred to Andrew Abbott's The System of Professions in her concluding discussion of the "jurisdiction of librarianship." According to Abbott, every profession aims for a heartland of work over which it has complete, legally established control. CIPA and mandated use of filters threaten to take away that control, at both micro and macro levels, by intruding on our professional jurisdiction and essential values. The relationship between librarian and patron is changed, as patrons who wish unfiltered access must declare their intentions, violating the fundamental principle of confidentiality. The types of material that are freely accessible in libraries may be qualified, through outsourcing to profit-motivated parties, as they have not been. In addition, there is no political guarantee that this could not be applied retrospectively other materials in a library's collection. CIPA inserts lawmakers into traditional library-patron negotiations regarding acceptable behavior (offensive user behavior did not suddenly appear with the World Wide Web). CIPA threatens to widen the digital divide among libraries, and to dissolve consortial infrastructures by dividing library members against each other. Finally, CIPA significantly erodes the First Amendment, since the material to be blocked includes far more than that which is illegal.

Reported by David Miller
Curry College
NETSL Writer/Editor