Péter Jacsó is something of a one‐man campaign in the USA against poor quality databases, and I have heard him speak on the subject and own his previous volume (Jacsó, 1992), which this, to some extent, replaces. The latest volume, Content Evaluation of Textual CD‐ROM and Web Databases, specifically does not include the software mentioned in the title of his previous work, but otherwise takes much the same approach in that it demonstrates quality issues with reference to numerous examples of good and bad practice.
The book is carefully researched – it has to be as it covers a huge amount in considerable depth – and demonstrates the author’s passion for, and knowledge of, the topic. The many examples show that he knows his raw material, too. Obviously, so many illustrations can make the book sensitive to the passage of time but the author is at pains to inform us that all searches were undertaken by him between July 1999 and November 2000, with some being run again in July 2001 at the page proof stage. The book was published in late 2001, so readers cannot expect more than this! But this is a book demonstrating methodology and evaluation criteria, which means that, even if a particular example has changed, or a database has vanished, the issue being explained remains.
Aside from two introductory chapters which define terms, explain the various iterations of a database (from producer to online vendor), the various database media, and the evaluation criteria, there are ten chapters dealing with subject scope, dimensions (time, depth, currency), source coverage, record content, accuracy, format and content consistency, the subject indexing, abstracts, and cost considerations. There is also a bibliography. In all there are some 324 figures demonstrating anything from an unlikely 18 records in the Agricola database apparently written in 1078, to the exemplary Getty name authority file! If I have one criticism of the text it is that nearly all of the examples are taken from only 13 databases, one of which must certainly have been taken down by July 2001. Readers can only guess whether the examples are typical of all 15,000 databases that the author suggests are available for our use.
Presenting the raw statistics of such a book – “never mind the quality, feel the width” – does little to help readers assess its value and is curiously at odds with the book itself, which urges database users to ignore the width and test the quality; the trouble for this reviewer is that its breadth of coverage makes it difficult to describe in any other way. What Jacsó does so well in this volume is to demonstrate the need and the means to evaluate databases: to compare one with another and each with its publisher’s hyperbole. Indeed, Jacsó says somewhere that the “word ‘comprehensive’ pops up quite often in scope statements by file producers, but the meaning of the word is obviously interpreted differently”. There is a whole chapter on subject scope and another on database dimensions. And, reading the book, it becomes abundantly clear that these things are important: coverage of a journal title is not necessarily comprehensive in one database (see, for example, the coverage of Program for some years in Information Science Abstracts); in another some countries may dominate the coverage (e.g. USA, the UK, Canada and Australia accounted for over half the entries in Ulrich’s). If the user is unaware of these anomalies, search results cannot be interpreted correctly. But it is easy enough to test almost any aspect of a database; although it is more difficult to impress the need on users. The author found it very easy to test the accuracy of databases on one system: 12 databases contained 66 instances of his name as “JASCO” – that is 66 records that might not have been located; 66 articles that may never be read.
The book, despite its size and the density of subject‐matter, is not a difficult read; Jacsó has a very easy style that should keep most readers interested. The chapter divisions also make it accessible, so that readers wanting to test a particular aspect of one or more databases can almost use the book as a handbook of tests. I have mentioned comparative testing twice and it crops up many times in the 250 pages. It is important – the coverage of Library Literature, Library and Information Science Abstracts and Information Science Abstracts varies considerably, for example, and not just in the number of records – but readers are warned to be circumspect in making such simplistic judgements. For example, in comparing the size of Ulrich’s and The Serials Directory (EBSCO), it must be remembered that the latter retains records for “ceased titles”, while in the former “successor” records “absorb information from a former title”.
It would need a very long review indeed to do justice to the content of this book, so in this short review I have tried to dip into it for examples that show how it addresses database evaluation, and in so doing to illustrate the author’s dedication to, and knowledge of, this subject. The book shows us how important it is to know the databases you use and it demonstrates how – given access and time – it is relatively easy to gain that knowledge. Content Evaluation works at two levels: it is both a reference tool for database evaluation and it is a treatise on information quality. It is a book that will impart wisdom to both database searchers and database producers, if only they are both willing to learn from it.
