XOBIS endeavors to restructure bibliographic and authority data in a consistent and unified manner using XML (1). Our objectives are to optimize integration and maximize the use of generalized solutions. The structure we have created provides a flexible framework for managing a broad array of information and reduces complexity without eliminating necessary granularity.We chose to broaden the scope of the "bibliographic" system in an attempt:
This also has the potential to facilitate improvement in the correlation of cataloging rules with the data encoding structure. XOBIS endeavors to address complex bibliographic problems, while providing a sturdy foundation to support future, integrated, advanced information retrieval and presentation systems. These goals are addressed in several ways as outlined next..
First, the delineation of fundamental XML elements creates homogeneous classes of information. Within these, crisp entry elements promote consistent identification of entities with a generalized provision for disambiguation when necessary. Secondly, rigorous separation of content from the information framework provides a mechanism for accommodating many types of change (e.g. adding new choice values without needing to revise the schema).Data types and their choice values are referenced generically in the schema to avoid prescribing values in the schema.This provides a degree of self-documentation and structural resilience. Thirdly, recognition of the importance and ubiquity of relationships is manifest in their unitary treatment as a special class of linking concepts.Relationship values are controlled in the data, not in the schema. Equivalence and variance relationships expressed within a record are treated identically to, but separately from, inter-record relationships. Overall, the schema is highly recursive in that there are multiple references to individual generically-defined elements, and in that choices are referenced as values occurring within the data. This affords a balance between richness of content accommodation and economy of expression. Each of these areas will be addressed in some detail.
The XOBIS schema can be interpreted very broadly, or very narrowly, to meet various needs.It permits many things, but prescribes few. While it could accommodate and potentially integrate entire dictionaries, directories, citation indexes, thesauri, catalogs, metadata sets, museum records, etc., it could just as easily be used for only a single type of data, such as a bibliography of books, with a minimal subset of data elements and no authority control of data content.
While XOBIS may appear at first to entail radical change, with few exceptions it only restructures existing data in an effort to achieve greater coherence, usability, and efficiency. We have tried to avoid arbitrary departure from, and adherence to, convention.That there would be a set of correlative cataloging rules is implicit in the tightly integrated framework, albeit with the assumption that major revisions in current practice would be required for XOBIS to move successfully beyond its initial exploratory stage.It helps to keep in mind that content and presentation are discrete in XML, and that indexing may bridge content structures.
Perhaps the most difficult part of arriving at the current schema was attempting to resolve long-standing problems extant in the current bibliographic apparatus.Problems with cataloging rules and MARC were presented to a Joint Meeting of MARBI/CC:DA in 2000 (2-5).Grappling with such problems consumes valuable staff time and resources, competing with qualitative and/or quantitative efforts to improve databases, and vexes developers of library-related information storage and retrieval systems. Inversely applying the 80/20 rule, we have concentrated on the 20% that is problematical, mindful that 80% is generally not in contention.Difficult decisions had to be made in reaching a balance between structural coherence, policy implications, and the need to support envisioned functionality. Resultant solutions in the XOBIS synthesis are interdependent.Selected illustrations of these are interspersed below.
XOBIS is viewed as only one component of a suite of schemas, each optimized for a particular functional area of information management. Such a superstructure of independent, yet coordinated, schemas would provide a more open approach to "integrated" library systems. The best solution for a particular function at any given time may vary, as do a library's needs and ability to acquire an entirely new system. The wisdom of bulk replacement of systems, with sometime mixed results, is questionable as the scope of library automation continues to grow.Emulating the "Web services model", which emphasizes discrete modules and inter-communication between them as a design goal, would enhance the sustainability of our information systems in the future.