Annual Review of Information Science and Technology, Vol. 34

Citation

Cawkell, T. (2002), "Annual Review of Information Science and Technology, Vol. 34", Journal of Documentation, Vol. 58 No. 3, pp. 323-327. https://doi.org/10.1108/jd.2002.58.3.323.3

In the preface of this book it is stated that: “Volume 34 was delayed because of an unusual number of very late chapter cancellations requiring late additional invitations … and an unusual number of illnesses”. The 1999‐2000 cover date means that volume 35 will be the 2001 volume.

I am assuming that every reader of Journal of Documentation has heard of ARIST so that introductions are unnecessary. Having written a chapter for ARIST many years ago I know that Martha Williams is an exacting taskmaster – a quality which is evident in this volume. Professor Blaise Cronin, Dean of the Library School at Indiana University, will become editor commencing with Volume 36.

A set of ARIST volumes on your bookshelf is a useful reference source. The appearance and size of Volume 1, 1966, edited by Carlos Cuadra, is similar to Volume 34 being reviewed here. Carlos says in the preface: “This volume, the first in what is hoped to be a long and increasingly useful series, is the product of many hands and a great deal of faith”. His hopes have been fulfilled. He continues: “grateful acknowledgement must, first of all, be made to the National Science Foundation and in particular to Mrs Helen Brownson without whose vision, encouragement, and support this series could not have been born”. Incidentally, it was Helen Brownson who assisted Cyril Cleverdon financially with his pioneering research on indexing languages carried out at Cranfield. She was in the audience when he was talking about his ideas in the USA. Cleverdon’s work endures – his name appears several times in the index of the current volume.

It is also stated in the preface: “the time period covered varies from chapter to chapter, depending on whether the topic has been treated previously in ARIST and if so, on the length of the interval from the last treatment to the current one. Thus, reviews may cover a one year or a multi‐year period”. For instance, Chapter one contains references going back to the 1970s.

The chapter titles and authors in this volume are:

1. (1)

   “Cognitive information retrieval”, Peter Ingwersen (Royal School, Copenhagen);

2. (2)

   “Methodologies and methods for user behavioural research”, Petling Wang (University of Tennessee);

3. (3)

   “Informetrics”, Concepçion S. Wilson (University of New South Wales, Australia);

4. (4)

   “Literature dynamics: studies on growth, diffusion, and epidemics”, Albert N.Tabah (University of Montreal);

5. (5)

   “Measuring the Internet”, Robert E. Molyneux and Robert V. Williams (University of South Carolina);

6. (6)

   “Applications of machine learning in information retrieval”, Sally Jo Cunningham, H. Witten and James Littin (University of Waikato, New Zealand);

7. (7)

   “Text mining”, Walter J. Trybula (Sematech);

8. (8)

   “Using and reading scholarly literature”, Donald W. King and Carol Tenopir (Ann Arbor and University of Tennessee).

Chapters 1 and 2 are headed “Planning information systems and services”, Chapters 3 to 7 “Basic techniques and technologies” and 8 “Applications”. As usual, the references, an index of 46 pages, and a cumulative index to earlier volumes are comprehensive. Most chapters contain up to 15 pages of references.

Michael Caine quite often uses the phrase “most people don’t know that” after imparting a useful piece of information. Not having received any formal education in librarianship or information science, I must classify myself as one of those people. Such knowledge as I possess was picked up in transit, so to speak. ARIST is both a good reference source and a good way to fill in knowledge gaps by browsing. Having recommended this book as a bookshelf necessity, this review will continue with a brief discussion of selected parts of chapters which may help to fill those gaps or which seem to be of particular interest.

I know very little about cognitive science, as discussed in this volume by Ingwersen. However, I remember reading somewhere about the work of Jean Piaget who was interested in the way the mind processes new information; he discussed the progress of children adapting to the environment. Ingwersen says: “Various researchers are finding ways to observe and understand the relationship of work tasks to information need with the goal of designing more responsive IR systems that employ mechanisms such as feedback and query expansion”. He dismisses Cleverdon’s “crude Cranfield model that has hardly changed for decades, although the same tradition increasingly produces evidence that suggests including variables of a more interactive nature than simplistic relevance feedback”. That model may be crude but I have not yet come across much in the way of effective models to replace it.

Conçepcion Wilson provides a table listing the top 20 journals which between 1990 and 1999 contained at least 11 documents relating to informetrics. The Journal of Documentation appears in this table which, as might be expected, is led by Scientometrics. Wilson devotes quite a lot of space to methods of evaluation using citations. Surprisingly, she says very little about the evaluation of departments or individuals by this method – an application which is widely used. For example, following a study by the UK Advisory Board for the Research Councils conducted back in 1987 using citation evaluations, it was concluded that UK research had slipped versus other countries between 1973 and 1982. There had been an appreciable drop in the number of citations received per UK published paper. However, the evaluation of individuals for tenure or promotion assessment purposes remains controversial. Eugene Garfield has written a number of cautionary articles about such operations, one of which is cited at the end of this review (Garfield, 1987).

Albert Tabah concludes his chapter with the remark: “some of the principal questions are: do publication counts adequately measure scientific progress? Are the quantity and quality of publications related? Can information science contribute to a better understanding of scientific progress? While these questions are not formulated in this way at the outset, the evidence presented in this review should allow one to answer all three in the affirmative”.

Molyneux and Williams do not think much of the literature in which the characteristics of the Internet are discussed. They say “the literature of Internet measurement is disbursed, fragmentary, fugitive, and rarely scholarly … . The bounds of any literature, including scholarly materials, are difficult to define with any confidence. In fact the indexing of commercial sources is often amateurish and feckless”. However, there is much of interest in the chapter itself. In discussing the literature about Internet search engine performance, the authors say “none has received the kind of attention given to the Lawrence and Giles studies of late 1997 to late 1998”. Molyneux and Williams cite their article published in Science (Lawrence and Giles, 1998) but not the one published in Nature (Lawrence and Giles, 1999). These articles are similar and well worth reading. As at December 2001, the Nature article has been cited 132 times and the Science article 114 times.

Various search engines are discussed in the course of this chapter but there is no mention of one of the best – Google. I find this rather surprising. Quite a few articles have been written discussing its speed, size and other attributes. I agree with the comments of Derek Rumsey from the University of Iowa who praises it (Rumsey, 2000).

Much of the discussion about machine learning in the chapter by Sally Jo Cunningham et al. covers rather simple stuff in terms of current ideas about artificial intelligence. There is nothing about intelligent agents for searching the Web – one of the most difficult and interesting aspects of this subject. I do not know of any agent which has yet been offered commercially and successfully used. The idea of having an agent conduct searches for you has been around for some years. For example, Patti Maes from MIT founded a company called Agents Inc. in 1996 which subsequently became Firefly. This company acquired an appreciable amount of venture capital and at one time was discussing cooperation with Microsoft. As far as I know no agent was subsequently marketed.

Text mining, discussed by Trybula, is “the process of analysing databases and developing methods for presenting results to the user”. A major problem is to develop effective methods for reducing large numbers of retrieved document collections to manageable proportions. The most popular method of doing this is by clustering, which depends on a measure of similarity between documents. The presentation problem is to “distinguish characteristics of the information being presented and to transform that information into knowledge”. The writer concludes that “a major part of the text mining effort is driven by computer scientists, many of whom do not have knowledge of the 30‐plus years of information retrieval work”. This pertinent comment seems applicable to quite large parts of information technology.

King and Tenopir take up a point which has been widely discussed by other authors: “the economic nature of the current journal system is one of its major weaknesses – particularly the general lack of understanding of why journal prices have skyrocketed beyond what might be explained by inflation and increase in journal sizes. The demand for and use of scholarly journal information are important economic aspects of the scholarly journal system”. Tenopir and King’s chapter is packed with information. Their more recent book about electronic journals, reviewed in the September 2001 issue of this journal, is probably a better choice for those interested in the subject rather than the section “Factors related to electronic journals” in this chapter.

Like its predecessors, the book contains an interesting selection of authoritative reviews. Any library which attempts to cover the field of information science and technology should include it upon its shelves.