BOOK REVIEWS
J. Chem. Inf. Comput. Sci., Vol. 37, No. 3, 1997 625
BOOK REVIEWS Parallel Computing in Computational Chemistry. Edited by Timothy G. Mattson. ACS Symposium Series 592. American Chemical Society: Washington, DC. 1995. viii + 222 pp. Hardcover: $69.95. 0-8412-3166-4. This book is based on “Parallel Computing in Computational Chemistry” symposium held at the 207th ACS National Meeting in San Diego, March 13-17, 1994. It covers papers that are a reflection of authors’ work. It can be useful to both novice and experienced parallel computational chemists. Novice readers can benefit from topics such as concepts and jargon of parallel chemistry, overview of parallel ab initio programs, program GAMESS, psudocodes, and various algorithms in parallel molecular dynamics. More experienced parallel computational chemists can enjoy such topics as object-oriented programming, Tool Command Language (TCL), Global Arrays (GA) package, parallel fast multipole approximation, the force decomposition algorithm, and the use of distributed shared memory in post-Hartree-Fock calculations. In the book one can learn about a variety of methods and hardware usage in computational chemistry from parallel Single InstructionMultiple Data (SIMD) machines to cost effective workstation clusters.
Zahra Behdadfar Kamarei Isfahan UniVersity of Technology, Iran CI960428A S0095-2338(96)00428-3
Electronic Conference on Trends in Organic Chemistry: ECTOC-1/CD. June 12-July 7, 1995. Edited by Henry S. Rzepa, Jonathan M. Goodman, Christopher Leach. The Royal Society of Chemistry: London. 1996. CD-ROM disk. L50 (plus VAT in the UK); U.S. $95.00. Were you on sabbatical in Lower Slobovia in the Summer of 1995 with no communication to the outer world? Did your Internet provider go bankrupt in Midsummer? If you missed the first ECTOC electronic conference for these or any other reasons, take heart: ECTOC-1 is now available on CD-ROM. This version can be run on your own machine at your leisure and provides most of the features provided in the online version (http://www/ch.ic.ac.uk/ectoc/). Admittedly, some of the hyperlinks lead to files available only on the server, but the conference is very usable in this medium. In fact, usage may be superior if your link to the Web is not very fast. Requirements include a loading of Netscape (v. 2.0 preferred, 1.1 usable) plus RasMol or a similar molecule viewer to rotate 3D molecular images. The conference is a collection of 75 papers on organic chemistry, grouped as six keynote papers (by such names as Albert Padwa and J. I. G. Cadogan), 35 synthetic, and 34 physical organic/mechanistic/ biological. Effective use is made of color not only for hyperlinks but also for highlighting and identifying fragments of molecules undergoing rearrangement reactions. Hyperlinked items include titles to text of papers, structures to preparations in the Experimental Section, and figures in text to the figures themselves. A molecule database with a hypertext glossary (described in a paper by editor Henry Rzepa) allows molecules to be searched by substructure and associated properties. Structures, 2D GIF images, can be viewed separately, saved, or printed. The E-mail discussion is documented, although following the threads is not always straightforward. In addition, many questions and much discussion was directed to the presenters and does not appear on the list. Usage statistics can be calculated for each paper. Efforts have been made to make this seem like a residential conference and not just a virtual conference. A list of all participants is provided, including those providing discussion. A matrix of photos of authors is provided. Clicking on a photo yields either an enlargement or recalls the text of the author’s paper. “Other conference activities” is unfortunately limited to the online version but included tours of the Natural History Museum, the San Francisco Exploratorium, and
London. Chemical Abstracts Service abstracted and indexed 66 of the 75 papers from this CD-ROM. This reviewer applauds the publisher and editors for extending the scope of electronic conferences to a wider audience, plus the advantages of a “living” archival tool. The CD-ROM publication of the 1996 conference, ECHET96, Trends in Heterocyclic Chemistry, is eagerly anticipated.
Robert E. Buntrock Buntrock Associates, Inc. CI960437J S0095-2338(96)00437-4
Introduction to Theoretical Organic Chemistry and Molecular Modeling. William B. Smith. VCH: New York. 1996. 192 pp. $59.95. ISBN 1-56081-937-5. This book was obviously written by a professor with considerable experience teaching theoretical organic chemistry. The alternation of theory and example create an excellent flow throughout the book. The book discusses the areas of Hu¨ckel MO theory, PMO, aromaticity, applications of Hu¨ckel theory, pericyclic reactions, molecular mechanics, semiempirical methods, and ab initio methods. The theoretical coverage is at a level that gives familiarity with the assumptions involved but avoids tiresome derivations. The chapters do an excellent job of discussing the strengths and weaknesses of each method. The interesting examples of how each method can be applied to understand organic reactions are very valuable. Many chapters have bibliographies leading students to more advanced discussions. The chapters incorporate examples and exercises using common programs, such as PCMODEL, CAChe, MOPAC, and programs available from the J. Chem. Educ. While information on the source of each program is available in the footnotes, it would be more convenient for instructors to have the software sources listed separately and with more complete information to aid in acquiring them for a class. The appendix includes a short program in basic for computing eigenvalues and vectors. This commendable inclusion helps students have a more hands-on feel for basic MO theory. Most programs are now designed to insulate the users from many technical details to make the use easier, and thus they lose their didactic value for the methodology involved. I recommend this book as a text for advanced undergraduate or introductory graduate classes on theoretical organic chemistry. It is a concise introduction to theoretical organic chemistry that could be covered neatly in one semester.
Matthew Clark The Institute for Scientific Information CI970384L S0095-2338(97)00384-3
The Information Revolution: Impact on Science and Technology. Edited by J.-E. Dubois and N. Gershon. (Data and Knowledge in a Changing World) Springer: Berlin. 1996. 273 p. $115.00. ISBN: 3-540-60855-9. This book is part of the CODATA series, Data and Knowledge in a Changing World. The monograph is an edited compilation of essays on the pure and applied concepts of scientific data management, provision, and research. Six chapters, each with an average of five essays/chapters, fall into two very broad categories: (1) the state of the information society and (2) examples from the growing world of scientific information collaborative projects.
626 J. Chem. Inf. Comput. Sci., Vol. 37, No. 3, 1997 Large scale scientific data projects, both national and international in scope, are used to demonstrate the variety of information issues. Examples of the topics dealt with are as follows: the Internet and Web services, free or fair circulation of scientific data, distributed networks, security, and national legislation. Part 2 provides examples from all over the globe of large information systems in use by the different scientific communities. The essays are written by information scientists from the field and/or the specific data project. Wherever relevant, the Web addresses are provided to the project and all the project participants. The inclusion of these Web addresses is a very helpful aid for the readers, and all of the addresses were correct at the time of review. The monograph’s emphasis is largely on the current state of the technology with some discussion of the direction these different projects may take. It is important to note that the projects in place are by no means of equal strength. Different regions are still dealing with limited internet connection, while others have already provided universal desktop access. East Asia struggles to standardize software languages, hardware, alphabets, and Internet access, while the West pushes toward something ominously called “the federation”. The interesting aspect of the federation is that it aims towards “interoperability” at a local, collaborative, and community level. It is, in essence, the information science manifestation of how scientific progress is attained. It is that unique balance of workstation raw data, finessed into the larger data pool, which is being added to the community’s (biology, chemistry, etc.) greater knowledge. The monograph is exciting in many ways: some organized projects are in the last stages of fine-tuning, while others are just approaching a prototype. However, the issues all parties and countries face, regardless of the strides already made, are issues of standardization, price, fairness of access, flexibility for local environments, and strength, or robustness, for international data collaborations. One chapter discusses the fairness of scientists having to pay for data to which they contribute gratis. This section rings especially true for those working in academic libraries and laboratories; science faculty are only too familiar with the ever-rising prices of scientific journals and databases. The monograph does not assume the reader has extensive information or technological literacy. The early chapters provide elegant and comprehensive overviews of the Internet, Web, hypertext, multimedia, database retrieval, and management. In addition to presenting the issues, the authors also present their solutions to data management and retrieval issues. The problem with this monograph is the same problem that any state of the art analysis has; there are already advances and refinements to much of the software mentioned in the text. This is an insurmountable problem where adVances and state of the art are concerned; this kind of material may be better suited to theme issues of relevant journals. The price of this work, $115.00, is expensive enough that any potential buyer may want to review the book prior to acquisition in order to determine which chapters still apply and which ones are no longer quite so valid. The audience for this work is broad: science and technology libraries, scientific laboratories struggling with work station implementation, students of information science, data management teams, in short any individual involved in the provision, development, and management of scientific data.
Veronica Calderhead Rutgers UniVersity CI9703866 S0095-2338(97)00386-7
Reviews in Computational Chemistry. Vol. 8. Edited by Kenny B. Lipkowitz and Donald B. Boyd. VCH Publishers, Inc., 220 East 23rd Street, New York, N.Y. 10010. xxi + 324 pp., June 1996. List Price $110.00. ISBN 1-56081-929-4 (Hard Copy), ISSN 1069-3599. This series brings together respected experts in the field of computeraided molecular research. Computational chemistry is increasingly used in conjunction with organic, inorganic, medicinal, biological, physical,
BOOK REVIEWS and analytical chemistry. This volume examines various aspects of computations in treating fullerenes and carbon aggregates, pseudopotential calculations of transition metal compounds, core potential approaches to the chemistry of the heavier elements, relativistic effects in chemistry, and the ab initio computation of NMR chemical shielding. This volume, the eighth, of Reviews in Computational Chemistry, represents the editors’ ongoing effort to provide tutorials and reviews for both the novice and the experienced computational chemists. The five chapters are written for newcomers learning about molecular modeling techniques as well as for seasoned professionals who need to acquire expertise in areas outside their own. All the chapters in the volume have a quantum mechanical theme. In Chapter 1, the authors show how ubiquitous semiempirical molecular orbital techniques need to be adjusted to correctly determine the three-dimensional geometries, energies, and properties of fullerenes and carbon aggregates. Chapters 2 and 3 elucidate the so-called effective core potential or pseudopotential methods that have proved invaluble for handling transition metals and other heavy metals. Quantum theory for describing relativistic effects, particularly important to heavy metals, is presented in Chapter 4. In Chapter 5, the author reviews NMR chemical shifts and explains the methodology with examples of heterocycles, buckminsterfullerenes, proteins, and other large molecules. The volume contains an excellent author and subject index. Information about the Reviews in Computational Chemistry is now available on the World Wide Web (http://www.chem.iupui.edu/∼boyd/ rcc.html).
Venkat K. Raman Chemical Abstracts SerVice CI970387Y S0095-2338(97)00387-9
Genetic Algorithms in Molecular Modeling. Edited by James Devillers. Academic Press: London, 1996. xi + 327 pp. $74.95. ISBN 0-12-213810-4. This is the first book in the new series: Principles of QSAR and Drug Design, edited by J. Devilers. The series is a welcome addition to scattered literature on QSAR and drug design in over a dozen journals, and if judged by this first volume, the introduction of the series is timely. QSAR, the quantitative structure-activity relationship, has grown considerably in the last 20 years, not only by the volume of researches devoted to this discipline but also the diversity of methodologies applied to QSAR. For example, the relatively recent methodologies include the partial least squares method, the cell automata, the neural networks, orthogonalized multiple regression analysis, and genetic algorithms, to which this book is devoted. The book consists of a dozen chapters written by leading researchers in the field, starting with introductory chapters on genetic algorithms in computer-aided molecular design (34 pp by J. Devillers), an overview of genetic methods (32 pp by B. T. Luke), and genetic algorithms in feature selection (20 pp by R. Leardi). The remaining eight chapters are devoted to different applications of the genetic algorithm. D. Rogers (22 pp) illustrates nonlinear modeling with splines and makes a comparison between GFA (genetic function approximation) and PLS (partial least squares). He started with a quote of Ernest Rutherford: “If your experiment needs statistics, you ought to have done a better experiment”, which only reminds us about the bias and misunderstanding of statistics at the turn of this century. It would be nice to know what would be the reply of Stainslaw Ulam (the father of the “Monte Carlo” method) to such criticism, but the quote of E. Rutherford is not quite out of place if suitably modified: “If your experiment needs better statistics, you ought to have used better descriptors”. W. J. Dunn and D. Rogers (22 pp) continue with introducing PLS and combining the advantages of PLS (extraction of latent variables approximately along the axes of greatest variations, optimal correlation) with the model generating ability of genetic algorithms to create modified genetic PLS. A. J. Hopfinger and H. C. Patel consider two