When the IFLA Study Group on the Functional Requirements for Bibliographic Records (FRBR) published their final report in 1998, they identified “data normally recorded in authority records” as warranting further study.1 Consequently, in April 1999 IFLA established a Working Group on Functional Requirements and Numbering of Authority Records (FRANAR). This group’s work—like that of the FRBR study group—has informed the development of our new cataloging code, Resource Description and Access (RDA), and of the broader IFLA Cataloguing Principles (ICP).² It has now been published as Functional Requirements for Authority Data (FRAD).

The original charge of the FRANAR working group included the investigation of an international standard authority data number (ISADN), but the group determined early on that this was no longer worth pursuing.³ Likewise, subject authority data—originally within the group’s purview—was hived off to a separate working group in 2005.⁴ Therefore FRAD restricts itself to modeling the data that has historically been recorded in name and title authority records, i.e., those that deal with what FRBR calls Group 1 and Group 2 entities (bibliographic resources and agents, respectively).

FRAD explicitly models data rather than records in recognition of the fact that authority and bibliographic records, while historically separate, are not necessarily so. For example, a catalog record for a work might conceivably include both work-specific authority data such as a heading and references, and work-level bibliographic data such as content summaries, subject descriptors, and classification.

Because FRAD was undertaken as an extension of FRBR, it uses the same entity-relationship model to analyze its subject matter—to identify entities (objects of interest), their attributes, and the relationships between them—and notes the user tasks (including the FRAD-specific contextualize) supported by each attribute and relationship. Like FRBR, which extracts from our descriptive cataloging tradition the four Group 1 entities of work, expression, manifestation, and item, FRAD introduces into its model a set of fundamental entities that are both familiar to us and slightly foreign, the latter primarily because we are accustomed to seeing them expressed in the two-dimensional context of a catalog card—whether physical or online—and its machine-readable carrier, the MARC record. Using these fundamental entities as guideposts, the structure of FRAD can be stated as follows: Each bibliographic entity—the Group 1, 2, and 3 entities inherited from the FRBR model—is known by one or more names and identifiers (the latter entity excluding—perhaps unnecessarily—record control numbers). These in turn serve as the basis for controlled access points (the headings and references of our authority records).

FRAD also includes two “back office” entities helpful in the interpretation of authority data—rules and agency—as well as the associated user task of justify. The rules entity in particular provides a context for interpreting the scope and values of other entities in the model. For example, under the 1949 ALA cataloging rules, a change in the name of a corporate body did not typically signal a new corporate body entity, while under later rules it did. Similarly, under the first edition of the Anglo-American Cataloguing Rules (AACR), a persona was not considered to be a person in its own right, while under the second edition of AACR it is. The agency entity, on the other hand, may be of more limited use in a world where data in authority records may be added or modified by a variety of interested agencies, and audit trails for such discrete actions are difficult to resurrect.

Most attributes are familiar, though some such as gender, affiliation, and field of activity are new, having been introduced from one or another of the sources from which FRAD drew its inspiration. As with FRBR, these sources are principally international specifications not widely used in the United States, such as Guidelines for Authority Records and References and UNIMARC Authorities.

Finally, FRAD, like FRBR, makes explicit the various types of relationships that often exist between entities. Many of these relationships are implicit in the data carried in MARC records but not in the associated content designation. While humans have little trouble teasing out such implicit relationships, machines are notoriously obtuse in this matter. FRAD’s explicitness in this regard will greatly aid machines in their efforts, if only by leading to additions and changes to the MARC21 authority format.

As can be seen from this review, in many ways the FRBR and FRAD models have been extrapolated from Things As They Are, which in most cases is also Things As They Were—a cataloger of 1970 would be able operate quite comfortably within these models. But they also carry in their structure the seeds of Things As They May Become. This is because entity-relationship models are very much at home in the online world, and over the long term, bibliographic data in such a world will be less and less constrained by the card catalog and MARC. It will be interesting to see how future catalogs take advantage of these models and once again are perceived to be saving the time of the reader.