Metadata for Information Management and Retrieval

Citation

Trickey, K.V. (2005), "Metadata for Information Management and Retrieval", New Library World, Vol. 106 No. 3/4, pp. 193-194. https://doi.org/10.1108/03074800510587390

This a usefully small book (186 pages) on a mushrooming topic that is threatening to tame intranets and ultimately the internet. In producing this review I think it appropriate to identify myself as an old school cataloguer so certain of the comments I make about this volume reflect my personal biases and will be acknowledged as such rather than being criticism of this presentation of the new order. For example I wanted a huge amount more “nuts and bolts” type detail that simply is not appropriate in a book of this size and purpose.

The book is divided into ten chapters that include an introduction, describing and expressing metadata, standards and data models. The next five chapters review the five purposes of metadata, this is followed by a chapter on managing metadata and the final chapter looks to the future.

The text is well structured, the preface describes the two audiences for the book “information professionals … and managers … ” (p. xi), I think it will also become a standard student text as it is of appropriate size and complexity to provide the level of “basics” that most professional courses require in this area. The preface also outlines the structure of the book and each chapter opens with an overview and closes with a summary. The book also has a useful index. The two other elements that I would find helpful are a glossary and a list of abbreviations used are sadly missing – I would have liked a neat definition of the term “granularity” which is banded about so much and infrequently defined by the aficionados of matters metadata. When a book refers to so many web‐based resources a CD or website containing the bibliography would facilitate access to the material mentioned. We are getting lazier (and less accurate) when it comes to information retrieval and the ability to “click on” the web address would make life so much easier.

As I have stated my bias, I will savage our author for the cardinal sin of lumping MARC in with metadata in general. The precision of the MARC format adds a spurious credibility to metadata systems that other formats, due to their failure to adequately deal with the “carrier and content” debate simply do not achieve. (MARC is a carrier that holds data that has been structured using “rules” like AACR2, Dewey Decimal Classification, Library of Congress Classification, ISBN, ISSN, etc. Most metadata advocates initially assumed that their format alone would provide sufficient organising capacity to allow for precise data retrieval.)

The core of the book deal with the five functions of metadata, these are resource description, information retrieval, management of information, rights management etc and interoperability. The author draws on a range of examples from different metadata systems to illustrate his argument. This provides for me one of the problems with the text. As we shift from format to format for points to be illustrated, I felt my sense of security slipping away. I would have preferred the use of a dominant format, like the ubiquitous Dublin Core to illustrate the main argument and occasional sorties into differing formats to make specific points that the basic model did not deal with.

As this is a brief tome, significant areas such as coherence of data and quality control get a mention, but of necessity a brief one so the rhetoric of interoperability wins the day over the grinding practicalities that are required to actually make it happen.

David Haynes is a persuasive advocate for the cause of metadata, and has provided a useful short volume that allows those new to this area to gain a useful understanding of the main features of this electronic landscape. The shelf life of this title will be short, as the topic is developing so rapidly so I look forward to subsequent editions of what could become the standard introduction to this topic.

In closing I feel I need to stay true to my antediluvian cataloguing roots and request that MARC21 should be presented as MARC 21.