Cataloging Community Galvanized as U.S. National Libraries Move To Embrace RDA
By Michael Kelley Aug 2, 2011The roof beam of a new bibliographic framework has been raised high and likely will be nailed into place in the not too distant future.
The three U.S. national libraries --- the Library of Congress (LC), the National Agricultural Library (NAL) and the National Library of Medicine (NLM) --- announced on June 13 that they will adopt, with certain conditions, the Resource Description and Access (RDA) cataloging code. This comes despite lingering fiscal and technological concerns, mainly around the cost of staff retraining and difficulties fitting the new framework into an aging record carrier.
Full implementation will not occur before January 1, 2013, but at the end of June 2011, at the American Library Association (ALA) annual conference in New Orleans, it was announced that the Library of Congress will begin partial RDA cataloging across all subject areas in November, with RDA records expected to represent about five percent of cataloging output.
"We believe that the long-term benefits of adopting RDA will be worth the short-term anxieties and costs," the three principals from the three national institutions --- Deanna Marcum, Sheldon Kotzin, and Simon Liu --- said in a joint statement on June 13. "We must begin now. Indefinite delay in implementation simply means a delay in our effective relationships with the broader information community." Libraries are increasingly sharing data with other organizations and institutions (such as publishers who have been developing their own metadata set called ONIX), a transfer facilitated by an updated and more compatible cataloging framework that better accounts for non-print resources, among other changes.
The decision came after the U.S. RDA Test Coordinating Committee released the results of its national test of the RDA code, which took three years to conduct from the initial committee meeting on June 9, 2008, to the public release of the committee's 192-page report and recommendations on June 20, 2011.
Work on RDA itself had been underway for several years previously, but the test became necessary after the LC Working Group on the Future of Bibliographic Control recommended in January 2008 that further development of RDA be suspended until a business case for RDA could be articulated. Specifically, the group asked for more insight into the benefits to libraries and end users, as well as a cost analysis for retraining staff and re-engineering cataloging processes.
The test involved 26 institutions (including the three national libraries) and generated 10,570 bibliographic records, 12,800 authority records, and 8000 survey responses. Participants hailed from libraries of every stripe as well as archives, museums, book vendors, systems developers, library schools, and consortia.
A majority of test partner institutions anticipated some negative impact on local operations (for example, on bibliographic file maintenance), but a majority felt that the U.S. community should move to RDA.
Thirty-four percent of institutional test responders said it should be implemented, 28 percent said it should be implemented with changes, 24 percent were ambivalent, and 14 percent said it should not be implemented. Among record creators, 25 percent gave an unqualified yes, 45 percent said yes with changes, and 30 percent said no.
Longterm goal of "robust metadata infrastructure"
Despite support for implementation, the test revealed several problems with RDA, not the least of which was that "there is little discernible immediate benefit in implementing RDA alone."
The committee said that it "wrestled with articulating a business case for implementing RDA," and it concluded that "the adoption of RDA will result in inevitable and significant costs in training."
But the committee, like the U.S. national libraries that followed the its lead, concluded that immediate economic benefits should not be the sole determining factor when deciding how best to construct "a robust metadata infrastructure" that will help the librarian community "retain its significance as an information provider."
RDA, which was published in June 2010, supersedes the Anglo-American Cataloguing Rules, 2nd edition (AACR2). One of RDA's aims is to better collaborate with a growing and international community of metadata users (including those beyond the library silo, such as museums or archives) and also to permit integration with emerging database technologies as well as the Semantic Web. To this end, RDA is constructed around two conceptual models developed by the International Federation of Library Associations and Institutions (IFLA): FRBR (functional requirements for bibliographic data) and FRAD (functional requirements for authority data), as well as IFLA's International Cataloguing Principles (ICP).
Beacher Wiggins, the director for acquisitions and bibliographic access at the Library of Congress and a co-chair of the coordinating committee, said FRBR terminology will present a challenge.
"Catalogers will need to absorb the distinctions among work, expression, and manifestation. To help pave the way, the coordinating committee is recommending that institutions begin now using the FRBR terminology in every day work discussions and to begin exposing their staff to briefings that focus on FRBR," he said.
The elements in RDA have been registered as an element set at the Open Metadata Registry, which should facilitate its integration with the Semantic Web once it has been developed. It was announced Monday that the first group of RDA controlled vocabularies have been reviewed, approved, and their status in the registry changed to "published."
(A presentation by Barbara Tillett describes the history and purpose of RDA. Tillet is the chief of the Policy and Standards Division, Library of Congress and the LC representative to the Joint Steering Committee for Development of RDA as well as a member of the test coordinating committee.)
The hope is that an increase in machine-actionable, relationship information in bibliographic data will allow it to function in a wide range of technological environments and also help users "find, identify, select, and obtain the information they want." These are the user tasks from FRBR.
Further improvements in the works
However, the test revealed that RDA, as presently configured, does not, or only partially, meets these and other goals that were set forth in the strategic plan for RDA devised by the Joint Steering Committee for Development of RDA (JSC), and that these "serious flaws" must be addressed "before the Committee can wholeheartedly recommend community adoption of RDA."
Among the goals that were not met were that RDA be optimized for use as an online tool, that it be written in plain English, and that it be easy and efficient to use. Over the next 18 months, these and other problems, such as improving the functionality of the RDA Toolkit, which is an online resource for implementing RDA, are going to be addressed.
"Many responders found the Toolkit to be clunky and difficult to navigate," the coordinating committee's report said. "Attempting to navigate to particular rules in the text via the table of contents confused many users."
In a July 12 webinar, Troy Linker, the publisher for ALA Digital Reference, which is a co-publisher of the toolkit, addressed the committee's report and discussed recent enhancements to the toolkit. Among the costs of implementing RDA is the price of the toolkit, which runs from $195 for a single user annual subscription to a base price of $325 for an unlimited number of users with a fee added for each additional user profile.
Moving away from MARC
Even though most libraries will implement RDA in a MARC environment, the committee recommended that "credible progress towards a replacement for MARC" be demonstrated. In May, LC announced the first step in a transition away from MARC with its Bibliographic Framework Transition Initiative.
Wiggins, from the Library of Congress, said this was one of the most important conditions necessary to address if the implementation of RDA was to move forward.
Existing systems can import and store RDA-based MARC 21 records, but "respondents indicated that substantial local configuration changes would be needed for indexing and record displays for the public," the coordinating committee's report reads. "Many survey respondents expressed doubt that RDA changes would yield significant benefits without a change to the underlying MARC carrier. Most felt any benefits of RDA would be largely unrealized in a MARC environment."
The coordinating committee reached a similar conclusion, saying that MARC had reached the end of its design life.
Other data communities do not use MARC as a record format. "As a result library data doesn't work well with other data. This puts library data at a disadvantage when it comes to being discovered on the web," wrote Liz Miller, a cataloging librarian at New Mexico State University Library, in the April 2011 issue of Reference & Users Service Quarterly.
In March, the co-executive directors of the eXtensible Catalog (XC) Organization, David Lindahl and Jennifer Bowen, met with the test committee to discuss the eXtensible Catalog's partial implementation of RDA.
"XC software represents the first live implementation of a subset of RDA in an FRBR‐based, non‐MARC environment," Lindahl and Bowen reported. "Using RDA in a non-MARC environment is not something that is far off in the future - it is possible now."
In the meantime, the MARC 21 website lists all changes to MARC for use with RDA. The RDA Toolkit also maps between RDA and MARC.
Core concepts still misunderstood
Among the more prominent changes to AACR2 instructions, RDA does away with the General Material Designator (GMD) and introduces three new elements (now covered in MARC fields 336, 337, and 338) for Content, Media, and Carrier Types.
The coordinating committee surveyed 163 library staff and patrons to assess the RDA record's understandability and its ability to meet user bibliographic information needs. Most of the survey respondents spoke favorably of the RDA record, which, among other changes, uses fewer abbreviations, eliminates Latin terminology and drops the "rule of three," which restricted the number of authors that could be included in the 245 and 700 fields. However, 65 percent criticized the dropping of GMD.
"RDA Content Type, Media Type, and Carrier Type are for the most part nonsensical and not helpful. Both the labels and the valid values are at best simply 'noise' in the record that users will skip over, and at worst confusing for user to identify format type," one tester wrote.
The coordinating committee wrote that a focus on such cosmetic differences between AACR2 and RDA in a MARC environment has led to a general misunderstanding about RDA and what it is intended to accomplish. In reference to the above comment, LC's Tillett said: "This is a clear misunderstanding that such data are intended as codes that could be displayed to end users as icons or through other terms that a local library preferred."
The coordinating committee recommended that training focus on FRBR concepts and "the ideal bibliographic description should be regarded as a set of reusable relationship information packets, rather than a monolithic set of individual and indivisible records."
By 2013, library schools should ensure that all of their students are familiar with FRBR concepts and terminology, the International Cataloguing Principles, and the value and potentials of linked data on the web, according to the report.
Vendors prepare to adapt, but stay AACR2 ready
The policy committee of the Program for Cooperative Cataloging (PCC), the international cataloging collaborative for which the Library of Congress serves as secretariat, backed the decision of the U.S. national libraries. It has already formed three task groups on RDA, and the coordinating committee recommended that the "issues associated with authority record creation be given first priority by the PCC" as this area generated the most controversy during the test.
A key concern for many librarians is whether any retrospective conversion will be necessary to adopt RDA.
"The consensus from the pundits and the early adopters seems to be that, whilst there are differences between AACR2 and RDA that will introduce inconsistencies, there is little justification for libraries to undertake retrospective conversion individually," wrote Terry Willan, a business analyst for Capita in the June 2011 issue of Catalogue & Index.
Vendors will need to adapt their systems to be ready for RDA, and a number of the most prominent vendors recently were separately interviewed on the RDA Toolkit blog about their readiness to accommodate RDA. Nearly all said they were current with RDA-related changes from MARC 21 Updates 9, 10, and 11.
Also, every vendor said that they had no plans to discontinue support for AACR2.
"Even if libraries adopt RDA for new records, it's likely that they will continue to require the same support for the large body of existing records encoded in AACR2," said John Larson, URM Product Manager for Ex Libris, in a typical response. "With that in mind, our approach is to design all systems such that RDA records can coexist with AACR2 records." Ex Libris provides LC's Voyager Integrated Library System as well as those of the other two national libraries.
LC's Wiggins recommended that public libraries stay in touch with their ILS vendors about changes that the vendors may be making to accommodate RDA.
Libraries urged to train toward possible 2013 implementation
However, the JSC will no longer update AACR2. "So continuing to use these rules does not remain a viable long term option," the committee report states.
"OCLC reported that updating its Bibliographic Formats and Standards, 4th edition, would entail a process comparable to its last full revision; that process required portions of time from about 30 staff members for nearly two years," the coordinating committee reported.
The coordinating committee urged members of the library community "to take advantage of the next 18 months and begin training their catalogers and working with their public services staff and system vendors to assure everything is in place for a potential January 2013 implementation."
The Coordinating Committee hosted a one day formal "train the trainer/tester" session before ALA Midwinter Meeting on January 15. This training session is available on the LC website. LC will also be mounting a new website in the next couple of weeks to track RDA planning efforts, along with other webinars for background, exercises, examples and more.
"[We] will be doing training for the Georgia Public Libraries Cataloging Summit next week." Tillett said. "Our focus, naturally is on public libraries. We hope to make those presentations available not only on their site but also on LC's forthcoming Web site about RDA planning for implementation. Our LC site should have a URL in the next couple of weeks," she said.
For those feeling a bit overwhelmed by the weight of the topic, a Twitter stream provides a little levity.
