DSS Book Reviews

Burkan supports his advice with specifics. He presents a detailed appendix on how to interview executives (including a sample list of questions) and a form for obtaining reference information. He is quite well versed in describing such managerial aspects of these systems as organizational structures that help in successful implementation and the organizational pitfalls encountered.

Burkan supports the use of consultants (after all, he is one) but also warns about the risks of consultancy. He is long on anecdotes on his personal experiences but short on describing actual systems. Burkan does present brief discussions of the functions he believes an EIS should perform (e.g., report retrieval, custom reports, exploration, variance analysis, what-if analysis, and modeling) as well as some EIS techniques (e.g., buttons, drill down, exception reporting, and color coding). However, he buries these descriptions in the middle of the book.

This book would have been helped if Burkan had given an expository treatment of what conventional EISs do and how they work. A case study or two (there are several in the literature either as research papers or Harvard Business School case studies) and descriptions of some commercial systems would have accomplished this purpose. The book also does not examine human interface issues (except briefly in a discussion of menus in the last chapter) or the rise of EIS packages on microcomputers.

Despite these shortcomings, I recommend Burkan's book to those considering creating an EIS and to those who have one already in place. It is well written and contains much sage advice. However, I still wish someone would write that definitive EIS book.

Reviewed in "Journal of Information Systems Management", Volume 9, No. 2, Spring 1992 by Paul Gray. Reprinted from Information Systems Management (New York: Auerbach Publications, 1992 Research Institute of America). Used with permission.

Leading Business Teams: How Teams Can Use Technology and Group Process Tools to Enhance Performance. R. Johansen, D. Sibbet, S. Benson, A. Martin, R. Mittman, and P. Saffo. Reading MA: Addison-Wesley Publishing Co., 1991, 216 pp.

It would be simplistic to say that Leading Business Teams is a follow-up to Robert Johansen's seminal book Groupware: Computer Support for Business Teams, which I reviewed in the Spring 1989 issue of INFORMATION SYSTEMS MANAGEMENT. This new paperback is much more. It summarizes the state of the art and provides a road map for IS managers who need to bring groupware into their organizations.

In Groupware, published in 1988, Johansen saw groupware (i.e., hardware and software for supporting business teams) as an emerging technology to which businesses should pay attention. The book defined groupware and presented 17 scenarios that described what were then actual and potential cutting-edge applications. In the new book, published three years later, Johansen and his team of coworkers present their conclusions about groupware and its implications after a year of hands-on work with the Groupware Users Project, which was sponsored by several companies and conducted jointly by The Institute for the Future and Graphic Guides.

Two new major themes have emerged. First, rather than just looking at all combinations of teams meeting at the same or different time and the same or different place, Johansen concludes that groupware will ultimately support teams meeting anytime, anyplace. Second, he joins forces with David Sibbet to look at the human aspects of teams and their development. In particular, they describe the use of group process tools to leverage the performance of task-oriented teams.

Groupware actually dates back to the late 1960s, when Doug Engelbart at SRI International built a conference room with a terminal at every seat. Such conference rooms are now being marketed commercially by IBM Corp. and others, and many such rooms exist at universities. Software for such electronic meeting rooms is an example of same-time, same-place groupware. Exhibit I illustrates the four combinations.

Johansen believes that we must move forward to the notion (first presented by Stanley Davis in Future Perfect) that support has to be provided for teams meeting anytime, anyplace. Such support allows team members whether they are in the same or different locations, on the road with a portable, or working at the client's location-to connect to other team members. Although technologically difficult today, the future dictates the need for anytime, anyplace support to reduce the cycle time and meet customer requirements.

Johansen and Sibbet realize that it's necessary to consider the team itself as well as the technology if a group is to be turned into a team. They adopt the Drexler/Sibbet Team Performance Model to provide a framework for thinking about how groupware should support teams as they move through different stages of performance. The model involves seven stages:

The first four stages are the creating stages; the last three are the sustaining stages.

Orientation, trust building, and renewal are stages when the team should work in same-place, same-time model because these stages require the richness of experience provided by face-to- face meetings. Electronic tools help but are subsidiary. During goal and role clarification and commitment, the team can be in different places but must coordinate in real time. Therefore, same-time groupware, such as videoconferencing and screen sharing, is appropriate. Once the team has entered the implementation and high-performance stages, different-time groupware (e.g., E-mall, computer conferences, and team rooms) comes into play.

Leading Business Teams covers the equipment (e.g., telephone, computer, conference room, projector, electronic mail, ISDN, and LANs) and the software as well as the facilitation and group process issues. The message is that both the technological and psychological issues must be addressed. One chapter emphasizes the importance of training and counters the argument that groupware is so simple that training is unnecessary. After all, someone has to know how to turn on the teleconferencing equipment and how to conduct the teleconference once the equipment is running.

The authors also discuss some of the pitfalls: the use of teams for a quick fix rather than long-term solutions, the assumption that groupware can overcome organizational inefficiencies, the potential for abuse of the groupware by a strong group leader, the built-in preconceptions in some of the software, count when organizations consider using groupware.

The book concludes with a series of scenarios about what to expect from groupware in 1995 and beyond. Among the scenarios are:

Leading Business Teams is an important, well-written book that is must reading for every IS manager whose department uses the team approach.

Reviewed in "Journal of Information Systems Management", Volume 9, No. 3, Summer 1992 by Paul Gray. Reprinted from Information Systems Management (New York: Auerbach Publications, 1992 Research Institute of America). Used with permission.

Executive Information Systems: Emergence, Development, Impact. H.J. Watson, R.K. Rainer, and G. Houdeshel (eds.) New York: John Wiley & Sons 1992 350 PP.

Finding the definitive book on executive information systems (EISs) is somewhat akin to seeking the Holy Grail. In past columns I reviewed three books on the subject (J.F. Rockart and D.W. De Long in Fall 1988, A. Paller and R. Lasker in Fall 1990, and W.C. Burkan in Spring 1992). Although each had specific merits, none passed the Grail test. Executive Information Systems: Emergence, Development, Impact, edited by Hugh Watson, R. Kelly Rainer, and George Houdeshel, is the latest entry.

A collection of 26 articles and working papers, the book reflects the ongoing exploration of ElSs at the University of Georgia under Watson as well as the development at Lockheed-Georgia of one of the seminal ElSs by Houdeshel and Watson. The papers range from Mintzberg's classic study, which showed that senior managers rarely spend more than 30 minutes on a problem (and even more rarely obtain their information from computers), to the initial definition of EIS by Rockart and Treacy in the Harvard Business Review. More recent papers presented at the Decision Support Systems 1991 meeting are also included.

As in any such collection, some papers are stronger than others, and the book contains a fair amount of overlap because many papers start with a definition of EIS and work forward. Unlike a textbook, the information is not presented in a straight line. You have to dig a little and do some synthesis on your own to put it all together. Nonetheless, there is a lot of meat here. From case studies (e.g., Lockheed, North American Rockwell, Marine Midland Bank, Hertz) to people considerations, critical success factors, requirements analysis, and then implementation issues, the book is comprehensive.

The book opens with two overview pieces (Rockart and Treacy and the Lockheed case study) followed by an examination of the nature of executive work. It then goes on to considerations of developing an EIS (e.g., requirements, software selection, user considerations, and available packages) before concluding with case studies that illustrate the impact of EIS. The major missing piece is an integrative view of the future, although there are a few hints. One survey finds a trend to EISs being initiated by IS departments rather than coming from executive requests. David Friend talks about the coming impact of groupware on EIS development as systems reach further down into the organization.

Executive Information Systems comes much closer to passing the Grail test than any of its predecessors. If you have to choose among the available EIS books, I strongly recommend this one for your bookshelf.

Reviewed in "Journal of Information Systems Management", Volume 10, No. 1, Winter 1993 by Paul Gray. Reprinted from Information Systems Management (New York: Auerbach Publications, 1993 Research Institute of America). Used with permission.

Technology for Teams: Enhancing Productivity in Networked Organizations, S. Opper and H. Fersko-Weiss, New York: Van Nostrand Reinhold, 1992, 181 pp.

Susanna Opper and Henry Fersko-Weiss come from the world of computer magazines and newspapers. In Technology for Teams they report on the status of groupware as of late 1991. The authors use groupware to refer to information systems specifically designed to help groups work together electronically. The groups can be at the same place at the same time or they can be dispersed, working in different time zones and at scattered sites. Networks facilitate group work but are not groupware by themselves.

The authors divide groupware according to its applications: administration (e.g., group calendars and meeting schedulers), information management (e.g., group editing or project tracking), communication management (e.g., computer conferences), and real-time meeting facilitation (e.g., group decision support systems). The book discusses the benefits and costs of groupware, the process of getting started with groupware, the human dynamics of groups, and the effect on the organization.

Technology for Teams is written at a popular level, in a cheerleading and sometimes a visionary style. It is a fine place to get started in reading about groupware; however, based on my own experience, groupware is much tougher to implement than is implied here.

Reviewed in "Journal of Information Systems Management", Volume 10, No. 2, Spring 1993 by Paul Gray. Reprinted from Information Systems Management (New York: Auerbach Publications, 1993 Research Institute of America). Used with permission.

Group Support Systems: New Perspectives, L.M. Jessup and J.S. Valacich, eds. New York: Macmillan, 1993, 365 pp.

Group Support Systems: New Perspectives is a collection of chapters edited by Leonard Jessup and Joseph Valacich, two bright young professors trained at the University of Arizona, one of the key places where group support systems have been developed. Having contributed one of the chapters to this book, be warned that I may not be completely objective.

The book examines the design and research issues associated with creating computer-supported meeting facilities. The core model is a group of people meeting in a single room (sometimes called a decision room); participants have access to computers during the meeting and one or more public screens are set up that display computer-generated information to everyone present. These facilities and their associated software were developed mostly in business schools, but they are now reaching the marketplace, with several vendors, including IBM Corp., providing software and hardware.

The book focuses on the research and development aspects of group support systems, although chapter 2 describes existing facilities and software. Of the remaining chapters, some are practical, dealing with laboratory and field research findings, the use of facilitation to make computer-supported groups more effective, space design, and interface design. Others are conceptual, discussing the implications of group dynamics, research methods, communication theory, and organizational design with respect to these systems. The book concludes with six scenarios that describe future uses, including moving to anytime, anywhere situations.

Although written by academics, only a few chapters are a tough read. Despite the research orientation, the book is useful to practitioners because it focuses on the multitude of technical, operational, and human issues involved in implementing group support systems.

Reviewed in "Journal of Information Systems Management", Volume 10, No. 2, Spring 1993 by Paul Gray. Reprinted from Information Systems Management (New York: Auerbach Publications, 1993 Research Institute of America). Used with permission.

Computer Augmented Teamwork, R.P. Bostrom, R.T. Watson, and S.T. Kinney, (eds.) New York: Van Nostrand Reinhold, 1992, 384 pp.

Two books with edited collections of chapters about decision support systems appeared almost simultaneously in late 1992: Computer Augmented Teamwork and Group Support Systems: New Perspectives. I contributed a chapter to both books, so be warned, I have some stake in each volume.

The term "group support system" (or group decision support system or computer-supported collaborative work, to name just a few of the variants) refers to combinations of hardware and software designed to help teams. Two early books on the subject were Groupware: Computer Support for Business Teams by Robert Johansen and Leading Business Teams by Johansen and five co-authors. These early volumes focused on the range of software available and the human interactions involved. The newer books are the first to describe the full range of efforts under way worldwide in the group support area.

Computer Augmented Teamwork is the result of an idea initially generated by Gerald R. Wagner. Believing that a book that describes representative facilities for group support would be of value to people contemplating building one for their own use, he commissioned several people to write descriptions of their facilities. Because of the exigencies of work with his own firm, he turned the project over to three people at the University of Georgia (Robert Bostrom, Richard Watson, and Susan Kinney) who served as editors and carried the project to completion. At one level, Computer Augmented Teamwork is a catalog description of 16 facilities with relatively standard information about each (e.g., their history, physical design, hardware, software, infrastructure, uses, and research activities). At another level the book demonstrates the diversity of what has been accomplished. The systems described involve commercial firms and universities. They are located in the US, the UK, Canada, and Singapore. Some systems primarily support communications among team members, whereas others are oriented to completing tasks. In most cases, people are in the same room at the same time, but sometimes, the participants are dispersed in space and time. The same-time, same-place facilities divide further into those that allow direct input to the computer by the participants and those that do not.

The book begins with an integrative chapter by Robert Johansen that describes the needs of teams working in each cell of the now standard two-by-two matrix, in which there are four types of group interactions: same place, same time; same place, different time; different place, same time; and different place, different time. Johansen then indicates the directions in which he sees computer-augmented work going over the next five to seven years. The book's second chapter is a tutorial on typical electronically augmented face-to-face meetings. Chapter 3 describes IBM's experiences in computer conferencing as observed by Lynda Applegate of Harvard Business School. These opening overviews are followed by descriptions of four communications-oriented team support systems. The technology here ranges from simple computer conferences to artificial intelligence-based support to multimedia support. The remaining chapters describe various single-room facilities. In reading these descriptions, it is interesting to see the variety of views of how groups work that are embedded in the software. Some software is quite hierarchical, with participants restricted to interacting with the computer only when asked to and only giving input in specific form. Other software is quite free-form, allowing anyone to take control at any time. Computer Augmented Teamwork holds great immediate interest for practitioners; in addition, the book contains invaluable appendix material that lists 37 products and the places where information can be obtained about them. I recommend readers augment it with Group Support Systems: New Perspectives (New York: Macmillan Publishing Co.), which from the point of view of the researcher, serves as the fundamental compendium of issues in the field. NOTE: Each of the books referred to above is listed on our web page and includes a review.

Reviewed in "Journal of Information Systems Management", Volume 10, No. 3, Spring 1987 by Paul Gray. Reprinted from Information Systems Management (New York: Auerbach Publications, 1993 Research Institute of America. Used with permission.

Readings in Groupware and Computer-Supported Cooperative Work: Assisting Human-Human Collaboration, R.M. Baecker (ed.) San Mateo, CA Morgan Kaufmann Publishers, 1993, 882 pp.

As Ronald Baecker, editor of this volume, points out in his preface, computer-supported cooperative work (CSCW) represents a shift in emphasis in computer science from human-machine communications and problem solving to human interaction mediated by information systems. With 882 pages printed on 81/2 x 11 paper, this collection of reprints of 73 articles appears formidable; it is certainly not the kind of book that you sit down and read at one sitting. Yet it is an indispensable guide to the present state of CSCW and groupware. The underlying assumption in CSCW is that people will increasingly work in teams and that, through appropriate software, hardware, and communications, the spatial and temporal location of people will become irrelevant. Although relevant commercial products are just starting to be offered by major vendors, it is clear that most IS managers will shortly have to make decisions about CSCW and groupware for their organizations.

The book contains 14 chapters divided into four parts: introduction, behavioral foundations and enabling technologies, asynchronous groupware, and synchronous groupware. Introductions are provided by the editor at the beginning of each chapter to put the issue and the papers in context. From the point of view of the IS manager, this volume puts all of the current developments in one place. The papers describe many of the hardware and software systems that are currently in development in Japan and Canada as well as in the US.

Although the literature is mostly in academic journals, the selections made here are highly readable and accessible to the non-specialist. A few papers (e.g., "How to Run a Meeting," reprinted from Harvard Business Review and "Networks" from Scientific American) provide general background. Others deal with business issues, including one paper describing how Boeing made a business case for an electronic meeting system.

Among the technologies discussed are hypertext, electronic mail, structured messages (i.e., the "electronic lens"), desktop conferencing, decision rooms, and media spaces. Of these technologies, commercial examples exist for all but media spaces. Media spaces refers to the idea of people being linked together through audio, data, and video media. Artificial reality (described briefly in one article) is the extreme form of media space. More realistic media spaces, which are still leading edge, are those being developed at XEROX PARC, at the University of Toronto, and at NTT and NEC in Japan. The University of Toronto system, called CAVECAT, allows people to remain in their offices and see everyone else in the meeting live on a split-screen monitor. Each conference can have different attendees and can be set up quickly. Thus, spur-of-the moment meetings can be assembled. Media space systems are still experimental and are turning up new and unexpected phenomena, such as the need to arrange the system so that users maintain eye contact.

Reviewed in "Journal of Information Systems Management", Volume 10, No. 4, Fall 1993 by Paul Gray. Reprinted from Information Systems Management (New York: Auerbach Publications,1993 Research Institute of America). Used with permission.

Executive Information Systems: Management Handbook. Jill Bird, Oxford, UK: NCC Blackwell Ltd., 1991, 141 pp.

Executive information systems (EISs) are designed for the most senior of end users, the managers who run organizations. Several books on the subject have appeared, most notably Executive Information Systems: Emergence, Development, Impact by H.J. Watson, R.K. Rainer, and G. Houdeshel (New York: John Wiley & Sons, 1992). Executive Information Systems: Management Handbook by Jill Bird offers a British view. The book is simple and easy to read. Bird points out that the justification of most EIS projects results from acts of faith or from the conviction that these systems will help the firm's most expensive employees. She discusses Implementation issues and points out that failures, when they occur, have little to do with the technology but are related to organizational. and people issues.

The book includes a list of the major systems available in Europe, most of which are also being marketed in the US. Like all such lists that appear in a book, they are dated by the time the book is published. In this case, the list is basically the state of the mainframe art of 1989. She misses completely the PC-based systems that have been coming on the market in the last several years. Although the book has several merits, including its conciseness, its focus on the critical issues, and the presentation of case studies of user experiences, it is recommended only to those who want a very quick introduction to the field.

Reviewed in "Journal of Information Systems Management", Volume 10, No. 4, Fall 1993 by Paul Gray. Reprinted from Information Systems Management (New York: Auerbach Publications, 1993 Research Institute of America). Used with permission.

Data Warehousing

Using the Data Warehouse W.H. Inmon and R.D. Hackathorn. New York: John Wiley & Sons, 1994, 285 pp.

The Data Warehouse Toolkit: Practical Techniques for Building Dimensional Data Warehouses. R. Kimball. New York: John Wiley & Sons, 1996, 388 pp.

Building the Data Warehouse. W.H. Inmon. New York: John Wiley & Sons, 1993, 298 pp.

The year 1996 may be remembered as the year of the data warehouse. It was a year when computer trade journals rarely went to press without at least one story, white paper, or new product announcement on the subject. Data warehouses are an integral part of a set of recent and interrelated developments in data management, executive information systems (EISs), and decision support systems (DSSs). These developments include online analytical processing (OLAP), multidimensional data bases, data mining, and document management. Although such names as data warehousing and data mining are unfortunate in the white-collar worlds of computing and executive decision making, the developments are important for IS managers.

Two niche software technologies that were both in trouble alone, executive information systems and data warehousing, are converging in the marketplace. For about 10 years, ElSs had cornered the market on elegant front ends suitable for executive use. Most of them, however, used proprietary data bases that required an army of people for updating. Furthermore, although the range of information provided often drew initial raves, it quickly became stale as executives absorbed what was in the data bases and hungered for more. From an EIS vendor's point of view, the market saturated. The software was expensive and the nice front ends that distinguished ElSs could now be produced at much lower cost with various visual programming languages.

Data warehouses also had problems. Whereas online transaction processing (OLTP) systems were designed to keep records of the current state of a business, data warehouses contained cleaned-up, historical time-series data. Because they kept information about long time periods, data warehouses required large amounts of storage, reaching toward the terabyte range. Although all interesting invention, data warehouses lacked a killer application that justified the large investment they required.

Somehow, around 1993, the data warehouse and the EIS people found one another, with the data warehouses obtaining their needed application and the EIS people receiving a new breath of life from expanding beyond the pretty screen. They were helped along by a seminal article co-authored by an early leader in relational data bases, E.F. Codd (E.F. Codd, S.B. Codd, and C. T. Salley,."Beyond Decision Support," Computerworld (July 23, 1993)) which asserted that data should be used for online analytic processing and that multidimensional data bases would replace relational data bases for this purpose. Data warehouses became a growth industry, with most Fortune 500 firms committing to creating one.

Data warehouses are still a developing field. A major controversy concerning them comes from the competition between so-called ROLAP and OLAP. The R and M before OLAP refer to the way the multidimensional data is stored and indicate, respectively, relational and multidimensional data bases. Whereas Codd called for the use of multidimensional data bases, the major data warehouse providers' stock in trade was inherently two-dimensional relational data bases used for transaction processing. These firms proceeded to work on ways of simulating multidimensionality with their relational data bases. One relational firm, Oracle, played both sides of the controversy by buying a major multidimensional data base provider. At this point, the issue of MOLAP versus OLAP is not resolved.

Building the Data Warehouse, W.H. Inmon. New York: John Wiley & Sons, 1993, 298 pp.

W.H. Inmon, author of Building the Data Warehouse is one of the early practitioners of data warehousing and a principal with a major data warehouse software firm. The book jacket and other sources bestow to him the imposing title of the father of data warehousing. Fortunately, Building the Data Warehouse lives up to the hype.

Inmon begins by nominally describing what he calls the evolution of decision support systems. Actually, he describes the differences in data requirements between operational data and derived DSS data. Inmon argues that in systems based on operational data, the classic systems development life cycle applies, with the first step being requirements gathering. In the data warehouse world, the life cycle is reversed. A simple data warehouse is built and then over time, as people understand what the data can and cannot do for them and the warehouse evolves, the requirements become understood.

In other words, the life cycle of the data warehouse is data-driven rather than requirements-driven. Inmon notes that CASE tools are designed for requirements-driven analysis and hence do not apply here. Data warehouses are designed and populated a step at a time and as such are evolutionary. The total costs involved, the resources required, and the disruption to the IS environment all necessitate that the warehouse be built in an iterative, orderly manner.

The data warehouse environment is characterized as a subject-oriented, integrated, nonvolatile, time-variant collection of data created to support management decision making. The sources of the data are the organization's transaction systems as well as external information. These ideas define the architecture. Whereas transaction systems are oriented toward the current status of each customer transaction and hence are online systems that change continually as events occur, data warehouses are loaded periodically and use time as one of their dimensions. That does not mean that the data is fixed.

Typically, a warehouse system is loaded periodically and keeps data for 5 to 10 years. However, the granularity of the data gets larger as time passes, because not as much detail is needed about older events. The data warehouse is also characterized by the use of metadata; that is, data about the data being stored. Metadata includes information about the structure of the data as seen by the programmer and by the DSS analyst, the sources of the data, what transformation and aggregations have been made, the relation between the data in the warehouse and the organization's data model, and the history of what information has been extracted.

The book devotes a chapter to data granularity (i.e., the level of detail), because Inmon believes it is the major design issue once the architecture is defined. Granularity determines the volume of data stored and affects the types of query that can be answered. In general, the volume of data is traded off against the level of detail in a query. The author concludes by recommending that for large data warehouses, data be kept at two levels of granularity: a high level for recent data and a lower level for older data. Exactly where in historical time the transition should be made is still an art that depends on the type of business involved.

Data warehouses also differ from conventional data bases in that they are usually denormalized-that is, the same data may appear several times. Denormalization allows for the combination of data into larger tables and reduces the number of input/outputs that have to be made, thereby speeding system operation.

The chapters in Building the Data Warehouse that deal with data base aspects are much better than the ones that deal with DSSs or ElSs. Although the idea of metadata is discussed in several places throughout the book, readers have to integrate the metadata idea for themselves. Despite these shortcomings the book is first rate. It is easy to read and filled with simple drawings that explain the concepts in the text.

Reviewed in "Journal of Information Systems Management", Volume 14, No. 1, Winter 1997 by Paul Gray. Reprinted from Information Systems Management (New York: Auerbach Publications, 1997 Research Institute of America). Used with permission.

Using the Data Warehouse, W.H. Inmon and R.D. Hackathorn. New York: John Wiley & Sons, 1994, 285pp.

Unlike Inmon's Building the Data Warehouse, which focuses on the data base designer, Inmon and Hackathorn's Using the Data Warehouse is directed toward the user and the data base administrator. Hackathorn is a former professor who now consults in data base issues.

In this book, Inmon and Hackathorn assume that the data warehouse exists and therefore focus on managerial and user issues. They also introduce the concept of the operational data store or ODS, which is an extension of the data warehouse to operational systems. The idea of the data store is to create a central data bank, similar to the data warehouse, which integrates operational information obtained from a variety of applications, many of which may be on legacy systems. The ODS is designed for operational decisions (e.g., granting credit and reordering) not tactical or strategic decisions.

Like the warehouse, the ODS contains information that is subject-oriented and integrated. However, the ODS differs from the data warehouse because it contains:

Because an ODS contains only current data, it is much smaller than a data warehouse. The authors assert that as a separate storage system, an ODS should never be combined with a data warehouse.

They also recommend that the data warehouse be built first. The contents of the ODS serve as one of the inputs to the warehouse. Data warehousing brings with it several managerial issues, many of which apply to all data bases. These include storage requirements, budgets, growth (which can be explosive), changing technologies, reports and reporting, and the development process. Data warehousing also provides managers with the ability to reduce the information processing done inside operational systems and hence speeds up response time for end users.

An extremely readable and useful chapter that deals with applications provides 10 examples of data warehousing. These include a steel manufacturing company that generates 40,000 data points from each batch produced, a cell phone company that measures the effectiveness of advertising campaigns for new subscribers, and a retailer that uses an ODS to perform online inventory management. Although the applications are said to provide competitive advantage, the brief examples are generic and do not identify specific companies or indicate the cost savings achieved.

The book concludes with chapters on administering the data warehouse environment, migrating to the architected environment, and connecting to the data warehouse. The administration chapter covers many topics, most of them too briefly, and mainly 'provides the reader with is a listing of the problems that must be dealt with. The migration chapter stresses the importance of creating the warehouse before the ODS. The final chapter on connecting to the warehouse examines the problems created by the multiplicity of hardware and software that exists in legacy systems. It then discusses, in some detail, the options of using a common interface, gateway, or protocol to overcome these problems. Of course, none of these options solves all the difficulties by itself.

Using the Data Warehouse nicely complements Building the Data Warehouse. The two books should be read together.

Reviewed in "Journal of Information Systems Management", Volume 14, No. 1, Winter 1997 by Paul Gray. Reprinted from Information Systems Management (New York: Auerbach Publications, 1997 Research Institute of America). Used with permission.

Building a Data Warehouse for Decision Support. V. Poe, Upper Saddle River, N.J.,: Prentice Hall PTR, 1996, 210 pp.

Perhaps the easiest-to-read but most cursory book is Vidette Poe's Building a Data Warehouse for Decision Support. The book is stronger on data warehousing concepts than it is on decision support systems. It is relatively weak on real-world examples, containing only four short case studies, and does not define several important current concepts such as data marts i.e., small, relatively cheap data warehouses for the departmental level), Codd's work on OLAP, data mining, or the so-called big button (the book's reference to how executives point and click on DSS and EIS menu items).

Poe's work assumes a relational data base approach and, with minor exceptions, ignores the multidimensional approach. It is written in a prescriptive tone and oriented primarily for marketing applications. The appendix listing products is skimpy.

Having said all this, there is still value to Poe's book. One way of looking at the book is as an extended checklist of things that need to be done. Although the work is not comprehensive, it is a good place for people with no experience in building a data warehouse to get started. It contains several important thoughts, some of which follow:

The book also contains several particularly useful features. Chapter 2 provides a set of definitions and short discussions of most of the major terms used in data warehousing. The diagrams in general are helpful, illustrating such ideas as alternative infrastructure required and arrangements of data. The chapter also contains a set of questions to ask executives about critical success factors.

The chapter on designing relational data bases for the data warehouse illuminates the various arrangements used. In particular, it clearly describes the idea of star schema (including fact tables and dimension tables), which were invented to make it possible to handle multidimensional data with relational data bases. The book contains useful but short discussions of the life cycle for data warehouses, which it attributes to DSSs, and the use of pi 'lots for prototyping. A chapter on successful data access contributed by Laura L. Reeves unfortunately falls short of its goals to cover the decision support topics basically ignored by Poe. It fails because it is relatively short and assumes that everything from predefined reports to ac hoc analysis to hard copy reports are DSSs. The final chapter on training, support, and rollout is much too brief.

Reviewed in "Journal of Information Systems Management", Volume 14, No. 1, Winter 1997 by Paul Gray. Reprinted from Information Systems Management (New York: Auerbach Publications, 1997 Research Institute of America). Used with permission.

The Data Warehouse Toolkit, R. Kimball, New York: John Wiley & Sons, 1996, 388pp.

Ralph Kimball's The Data Warehouse Toolkit comes with a CD-ROM containing a sample data warehouse and data bases to go with it. Kimball, founder of Red Brick Systems, a major data warehouse vendor, was one of the designers of the Xerox Star.

The book takes an unusual and interesting approach. Following a brief overview, Kimball devotes 10 chapters to 10 case studies, each dealing with a different industry and providing much more information than do the examples of Inmon and Hackathorn's Using the Data Warehouse.

Kimball's examples are designed to illustrate, in practical terms, different aspects of the (relational) multidimensional modeling needed for a warehouse. The chapters outline the business process to be modeled, the granularity of the data to be kept, the dimensions to be used, and the data items (i.e., facts) to populate the tables.

The cases start with a grocery chain and end with frequent flyers and other aspects of transporting people and goods. For each case, an estimate is given of the size of the data warehouse that needs to be provided. Thus the grocery chain with two years of daily data on 300 stores each carrying 30,000 products requires 657 million records in a base fact table of 21 gigabytes. Other warehouse sizes in the examples range from 0.1 to 50 gigabytes. In later chapters, Kimball discusses even larger situations such as telephone call tracking, which requires 3.5 terabytes. He does point out that warehouses this large are beyond the upper bound of the mid-1990's comfort zone of 1 terabyte.

Following the examples, Kimball discusses the logical design, degree of aggregation and effects of aggregation on performance, management of the warehouse, and user applications. In management, he separates what he calls the back room and the front room. The back room refers to the tasks handled by the data base administrator, whereas the front room is for business-oriented people. Back-room functions deal with creating and maintaining the technical aspects of the data. Building application templates, training, and keeping the network running are front-room responsibilities.

Kimball concludes with a view of the future, which he sees as coming less from hardware (although parallel processing appears promising) than from software, which he views as primitive at best. In part, this primitiveness results from the crude way in which star joins are handled. Improvements can also come from using better indexing schemes for both the fact and the dimension tables. Because of differing priorities, many of the software improvements will come as the mindset moves away from OLTP. In OLTP, for example, the need for online retrieval makes compression undesirable, but in OLAP applications compression can be justified economically.

Appendices summarize the design principles given in the text (there are 11 pages of them) and provide a system checklist, a glossary, and a user's manual for the software on the CD-ROM. As part of the CD-ROM, Kimball provides sample data bases for the case examples. By using the data base and the data models available on the CD-ROM, the reader is able to follow the discussion by creating reports and seeing the effects of the dimensionality. The data models are shown in visual form to make comprehension easier and the reports are shown spreadsheet form. The software can be used by readers to create small data warehouses for their own use.

Reviewed in "Journal of Information Systems Management", Volume 14, No. 1, Winter 1997 by Paul Gray. Reprinted from Information Systems Management (New York: Auerbach Publications, 1997 Research Institute of America). Used with permission.

Groupware: Technology and Applications, David Coleman and Raman Khanna (eds.) Upper Saddle River, N.J. Prentice Hall PTR, 1995 576 pp.

The software and hardware that supports teams is the subject of Groupware: Technology and Applications edited by David Coleman and Raman Khanna. Coleman is the founder of the annual groupware conference, a combination trade show and technical conference similar to Comdex. Khanna is responsible for distributed computing at Stanford University. The focus here is on what groupware is and how it can be used in organizations. The editors have brought together in one volume chapters by academics, consultants, sees groupware as enabled by such technologies as client/server, networked and distributed applications, multimedia and as enabling enterpriseware and virtual enterprises. He accepts both Bob Johansen's two-by-two matrix classification of groupware based on place and time (i.e., same place versus different place and same time versus different time) and Esther Dyson's classification by locus of control (i.e., user, work, process). He sees the current interest in groupware coming from a convergence of technical, economic, social, and organizational factors. Coleman also discusses factors favoring and opposing groupware, the internal company justifications for groupware, and the organizational changes it implies.

The first part of the book is devoted to groupware systems and includes chapters on electronic mail, workflow, group calendering and scheduling, document management, and electronic meeting systems. Of these, the calendering and scheduling chapter was weakest (this may be a personal bias because I don't like the concept of having my time scheduled for me), the electronic meeting systems chapter was the strongest. The document management chapter adds to discussion. An additional chapter on project management as groupware wouId have been welcome.

The second part of the book, 120 pages on vendor strategies, is mostly hype for the particular vendors: Lotus, Microsoft, IBM, Digital, and Novell. The following sentence (with the name deleted to protect the guilty) is typical: "Award winning [name deleted] business solution software provides a development environment in which you can integrate personal productivity, groupware, and business applications for departmental and line-of business success." Because these firms make continual changes in their strategies, the material is ephemeral at best. These changes are duly and more accurately reported in the trade press.

The final part of the book looks at how groupware is being implemented. The first chapter in the section, by Marvin Manheim and co-authors, deals with the relationship of groupware applications to business objectives. Manheim, a practice-oriented senior professor at Northwestern with long experience in such nuts and bolts areas as logistics management, gives a particularly good example of how groupware works in supply chain integration, particularly the integration of logistics and manufacturing. The chapter is filled with other examples. This chapter and the previously mentioned chapter on electronic meeting systems provide, by themselves, a reason for owning the book.

Other implementation chapters deal with the problems and successes in bringing groupware (mostly mail and networking) into the World Bank, the Bank of Montreal (workflow, mail, conferencing, and document management), and the Marriott Corp. (electronic meeting systems), as well as with the use of groupware by consulting firms both for their clients and for themselves. These case studies provide insights into how groupware affects large organizations and stress the people problems involved.

The book concludes with two conceptual chapters on strategies for transferring the technology and implementing the cultural changes that groupware brings. Be warned that this book is full of lengthy, boring biographies of each of the many contributors. They may satisfy the ego needs of the authors, but knowing that someone worked for a minor software house 15 years ago adds little to our understanding of the subject. The biographies could well have been omitted without any loss to the message.

Reviewed in "Journal of Information Systems Management", Volume 14, No. 2, Spring 1997 by Paul Gray. Reprinted from Information Systems Management (New York: Auerbach Publications, 1997 Research Institute of America). Used with permission.