Svenonius describes the ideals that catalogers have reached for and their struggles to adapt to new technologies and circumstances without losing the intellectual legacy (40 million documents already cataloged) or the functionality embodied in previous generations of catalog. The viewpoints and insights (even idiosyncrasies) of the librarians who have shaped the profession and who have taken important steps along the path to "universal bibliographic control" provide a refreshing human context. Sir Anthony Panizzi, around 1850, explained the need for a catalog for the books in the British Library to bring together like items and differentiate among similar ones. Svenonius sees this as the defining objective of a system for organizing information. A century later, the need to distinguish between information and its physical embodiment had become clear. For example, in 1955 S. R. Ranganathan used the distinction between a work as expressed thought and a document as embodied thought.
Catalogs, meanwhile, have migrated from book form to cards and from cards to online records; local inventories have been merged into a global resource; practices developed for cataloging books have been adapted for use with maps, music, pictures, and other non-book materials. Each transition entailed adaptation but not complete replacement of practice. For example, some uses of cross-references in book catalogs were more effectively handled by added entries in card catalogs. Many inconsistencies in today's library catalogs can be traced to the cumulative effect of these transitions and the cost of retrospective changes to records.
Some aspects of the conceptual framework may be familiar from Functional Requirements for Bibliographic Records, the report of an International Federation of Library Associations (IFLA) study group for which Svenonius was a consultant. Her starting point is a set of objectives: finding, collocating, choosing, acquiring, and navigating (among items in a bibliographic universe). She argues that cataloging rules and bibliographic systems should be evaluated against these objectives, which "can be looked on as a statement of what users have a right to expect from systems for organizing information." This approach parallels that of use cases by software engineers. Her assessment is that many library catalogs, although subscribing to the objectives in principle, "have become deranged over time, in large part due to retrospective conversion and shared bibliographic records." Here, and elsewhere, Svenonius presents shortcomings in practices or systems; she does not suggest specific solutions. It is the objectives rather than existing online catalogs that can guide builders of digital library systems.
In addition to the objectives, Svenonius introduces principles of description that have been developed by the cataloging community over the years and against which recommendations for change in cataloging practice are assessed. These too deserve consideration in other contexts. Again, Svenonius points out inconsistencies and notes that blind adherence to one principle can lead to practices that prove short-sighted. This does not detract from the usefulness of these principles as a checklist to consider when designing a metadata scheme or formulating guidelines. First in the list of principles is that of user convenience, with a subprinciple of common usage. The principle of representation indicates that descriptions for an information entity should be based on the way it represents itself. The principle of sufficiency and necessity argues for assessment of choices for metadata elements and guidelines for values against stated objectives. More general principles include the principle of sufficient reason and the principle of parsimony. The careful reader will notice that Svenonius introduces her own principle: avoid using one device to serve multiple functions (e.g., to disambiguate and to order). As she explains, although favored by the principle of parsimony, this "introduces a lack of flexibility that can become an obstacle as technology changes." As I read this sentence, I understood the root cause of several problems I have encountered integrating bibliographic records into American Memory. Articulated insights like this led to my excitement on reading the book.
Given objectives and general principles, Svenonius then delves into discussion of "languages" to describe works, documents, and subjects. In each area, she introduces additional principles and highlights problems that have not been satisfactorily resolved. She describes struggles to find standard ways to deal with part-whole relationships and multiple versions that may be equivalent in some contexts but not in others. She declares, "for document types, as for general-format types, it is not possible to construct a classification that is both natural and whose categories are mutually exclusive." Anyone who has puzzled over similar challenges should find illumination and comfort, if not immediate solutions. Throughout, and in an afterword, she suggests areas where computer systems have the potential to alleviate problems. As Professor (now Emerita) of Library Information Science at the University of California, Los Angeles, her research focus has been on the design and evaluation of cataloging systems in the automated environment. In the last 30 years, since she wrote her Ph.D. dissertation on "The Effect of Indexing Specificity on Retrieval Performance," Svenonius has written what are now standard texts on cataloging and classification, and played a key role in many of the professional bodies and processes that have sought to harmonize and improve cataloging practices within a sound theoretical framework. The experience shows. Some might argue that she has the luxury of an ivory tower, but she has kept the windows and doors open. Here she acknowledges that some practices have not kept up with the challenges of electronic information and questions whether certain practices are helpful in an online environment.
Svenonius does not hesitate to use specialized terminology, but acronyms are kept to a minimum and explanations and examples make the meaning clear. This volume in the MIT Press series on Digital Libraries and Electronic Publishing is a scholarly work, with many footnotes and references. Do not ignore the footnotes; some examples and insights are hidden away there. This book about cataloging certainly deserves to be read outside the community whose principles and traditions it describes and illuminates.