Ch. 0: Preface
Search is an integral part of peoples' online lives; people turn to search engines for help with a wide range of needs and desires, from satisfying idle curiousity to finding life-saving health remedies, from learning about medieval art history to finding video game solutions and pop music lyrics. Web search engines are now the second most frequently used online computer application, after email. Not long ago, most software applications did not contain a search module. Today, search is fully integrated into operating systems and is viewed as an essential part of most information systems.
Many books on information retrieval describe the algorithms behind search engines and information retrieval systems. By contrast, this book focuses on the human users of search systems and the tool they use to interact with them: the search user interface. Because of their global reach, search user interfaces must be understandable by and appealing to a wide variety of people of all ages, cultures and backgrounds, and for an enormous variety of information needs.
The truly world-wide reach of the Web has brought with it a new realization among computer scientists and lay persons alike of the enormous importance of usability and user interface design. In the last ten years, much has become understood about what works in search interfaces from a usability perspective, and what does not. Researchers and practitioners have developed a wide range of innovative interface ideas, but only the most broadly acceptable make their way into major web search engines. This book attempts to summarize these developments, presenting the state of the art of search interface design, both in academic research and in deployment in commercial systems.
This is a fast-changing field, and any attempt to summarize the state-of-the-art will no doubt soon be proven obsolete. Nonetheless, certain principles and techniques seem to hold steady over the years, and there is much that is now known about search interfaces that should stand for at least the near future.
0.1: Book Overview
This book outlines the human side of the information seeking process, and focuses on the aspects of this process that can best be supported by the user interface. It describes the methods behind user interface design generally, and search interface design in particular, with an emphasis on how best to evaluate search interfaces. It discusses research results and current practices surrounding user interfaces for query specification, display of retrieval results, grouping retrieval results, navigation of information collections, query reformulation, search personalization, and the broader tasks of sensemaking and text analysis. Much of the discussion pertains to Web search engines, but the book also covers the special considerations surrounding search of other information collections. The chapters are elaborated on in the following paragraphs.
The Design of Search Interfaces: (Chapter 1) introduces the ideas and practices surrounding user interface design generally, and search interface design in particular. It opens with an analysis of why Web search interfaces appear standardized and relatively simple compared to other interfaces, and places modern search interfaces into a historical context. The remainder of the chapter is a summary of interface design guidelines as applied specifically to search interfaces. This chapter is intended to be useful for those people who do not have time to read the entire book but want to understand best practices and problems to avoid in the design of search user interfaces.
The Evaluation of Search Interfaces: (Chapter 2) is a companion to the chapter on Design, as user-centered design requires tight-coupling with evaluation. The chapter summarizes the key methods for evaluating search interfaces: informal studies, formal studies, field studies, longitudinal studies, and large-scale log-based studies (also known as bucket testing). This is followed by advice about best practices and special considerations to keep in mind when evaluating search interfaces.
Models of the Information Seeking Process: (Chapter 3) summarizes the theoretical models that have been proposed about how people seek information. These models are the foundation upon which much of search interface design is based. These include the standard model, the cognitive model, the dynamic (berry-picking) model, information seeking as a strategic process (including cost-structure analysis and foraging theory), orienteering and incremental strategies, and the theory of sensemaking. This is followed by a discussion of information needs and query intent, including attempts to create taxonomies of searcher's information needs and intents, by manually and automatically analyzing queries and online behavior.
Query Specification: (Chapter 4) is the first of a set of three chapters that describe interfaces to support the interlocked information seeking cycle of query specification, viewing of retrieval results, and query reformulation. This chapter summarizes both research and the state of current practice in search interface design for query specification, including textual queries, natural language questions, query specification forms, dynamic feedback, queries using Boolean and other operators, faceted queries, and command-based queries.
Presentation of Search Results: (Chapter 5) is the second of three chapters on interface support for the standard information seeking cycle. This chapter summarizes research as well as the state of current practice for displaying search results pages. Topics include document surrogates, properties of results listings, summaries or extracts as used in search results, and user response to search results ordering.
Query Reformulation: (Chapter 6) is the third of three chapters on interface support for the standard information seeking cycle. This chapter discusses the need for and frequency of query reformulation, followed by interface ideas that support reformulation from both research and the current state of practice. Specifically, these are spelling suggestions and corrections, automated suggestions for query refinement and expansion, suggesting popular destinations, relevance feedback, and suggesting related articles.
Supporting the Search Process: (Chapter 7) is a capstone to the previous three chapters, describing interfaces that encompass and augment the full standard process of information seeking. Topics include interfaces to support finding starting points for search, using history and re-finding, and to support the sensemaking process that often accompanies but is broader than search.
Integrating Navigation and Search: (Chapter 8) discusses interfaces to support the integration of browsing of information structures with directed search, primarily in information collections (as opposed to for the Web as a whole). Topics include using categories to sort, filter and group search results, organizing results by table-of-contents-like views, faceted navigation and automatically-derived clusters for organizing search results. The chapter concludes with a discussion of the tradeoffs of using categories versus clusters for search results organization.
Personalization in Search: (Chapter 9) explores the emerging area of using information about individual users to influence search results ordering, to create automated alert services, and to tailor information recommendations to users. There is intense interest in research and industry surrounding information personalization, although most attempts to personalize information are still short of their mark.
Visualization for Search Interfaces: (Chapter 10) is the first of two closely related chapters on the use of information visualization in search and text analysis. This first chapter provides a brief introduction to the main principles and techniques used in visualization of abstract information (as opposed to scientific visualization which renders real-world objects in visual form). It also discusses why visualization of nominal data -- of which text is composed -- is difficult to do effectively. It then describes some of the many attempts to use information visualization to improve query specification and display of retrieval results, as well as to give overviews of information collections. Unfortunately, in most cases, usability studies incorporating these visualizations find that they in the best case do not improve peoples' performance, and in the worst case they slow people down or cause them to make errors. That said, in many cases study participants find visualizations to be appealing, at least at first exposure if not for extended use.
Visualization for Text Analysis: (Chapter 11) describes information visualization for text analysis, which seems to be a more successful application area for visualization of textual information. Although primarily of interest for analysts and specialists, as opposed to for everyday search use, these techniques are often creative in design and captivating to view.
Emerging Trends in Search Interfaces: (Chapter 12) closes the book with a discussion of areas of search that are still relatively new but promise to be of increasing importance in the coming years. Topics include mobile search, multimedia search, social search, and a hybrid of command-based and natural language search.
There are a number of topics related to search which this book does not cover. These include interfaces for database systems, Search Engine Optimization (SEO), the role of advertising in search, spam detection and elimination, and ranking algorithms. This book assumes that the reader is familiar with the technical basics behind search engines and information retrieval, including crawling, indexing, ranking, Boolean queries, and PageRank. (For those who are not, see the Related Books section for suggested readings.)
This book is an update of and expansion to a chapter written in 1998 for the book Modern Information Retrieval, Baeza-Yates and Ribeiro-Neto (Eds.), Addison Wesley, 1999. At that time, little was known definitively about which ideas result in usable search interfaces. In the intervening ten years, much has been learned. Thus, one goal of this book is to back up every statement with verification from the literature. This can be challenging when much of the knowledge is locked up in industry, but fortunately, a number of recent papers have appeared that share insights from the major web search engines. (An exception is made for Chapter 1, which is intended to be a summary of best practices encapsulated into one chapter; unsupported statements made there are verified in later chapters.) Although a large proportion of the references are necessarily drawn from research from the last few years, readers are also exposed to early foundations and ideas.
0.2: Using This Book
This book has two intended primary audiences. The first is academic researchers, graduate students, and those teaching graduate level courses in information retrieval, user interfaces, and other information management-related topics. The second intended audience is practitioners who design and build search interfaces. Although the book makes heavy use of academic references, an attempt has been made to keep the language and concepts approachable. Instructors may want to view this book as having two main parts, with Chapters 1 - 7 covering search interface fundamentals, and Chapters 8 - 12 covering advanced topics.
The contents of this book are available online at: http://searchuserinterfaces.com. Updates to the subject matter presented in this book will appear at the Web site.
0.3: Related Books
For a nice introduction to the mathematical foundations and algorithms for search, geared primarily towards undergraduates, see Introduction to Information Retrieval, by Manning, Rhagavan, and Schutze, Cambridge University Press, 2008. For a more advanced research-oriented book on a wider range of topics related to search, see Modern Information Retrieval, 2nd Edition, by Baeza-Yates and Ribeiro-Neto (Eds.), Addison Wesley, 2009 (to appear).
For details on Web search algorithms, see Mining the Web: Analysis of Hypertext and Semi Structured Data by Chakrabarti, Morgan Kaufmann, 2002, and for details on link-based algorthms as well as general web search algorithms, see Google's PageRank and Beyond: The Science of Search Engine Rankings by Langville and Meyer, Princeton University Press, 2006. For implementing search engines, see Managing Gigabytes by Witten, Moffat, and Bell, Morgan Kauffman, 1999, and Lucene in Action by Gospodnetic and Hatcher, Manning Publications, 2004.
To date there is no other academic book that focuses on search user interfaces. The most closely related is Information Seeking in Electronic Environments by Marchionini, Cambridge University Press, 1995, which focuses on the search process rather than on interfaces for search. Finding Out About: A Cognitive Perspective on Search Engine Technology and the WWW by Rik Belew, Cambridge University Press, 2008, is a new edition of a book first published in 2000, and describes basic algorithms as well as a discussion of some cognitive properties of search.
Books written by and for practioners include Information Architecture for the World Wide Web, 3rd edition by Morville and Rosenfeld, O'Reilly Media, 2006, which describes design of information architecture, including two chapters on search, and Designing Web Navigation, Optimizing the User Experience, by Kalback and Gufstafson, O'Reilly Media, 2007, which discusses navigation design for web sites.
0.4: A Note on Terminology
The words user, searcher, and information seeker are used interchangibly in this book to indicate a hypothetical or an actual person using a search system. Some authors object to the word user, both because they feel it reduces a person to what they are doing with a computer, and because of its association with recreational drug use. Its use is, however, standard in the field and is convenient to write with, and so this book is a user user. By contrast, the word participant is used to refer to a person who voluntarily participates in a usability study.
This book also unapologetically uses the pronoun they and the possessive their to refer to the third person singular as a way to avoid making explicit (and unnecessarily distracting) gender distinctions.
The author has been employed by, consulted for, and/or received research gifts or grants from the following institutions whose ideas, products, or projects are mentioned in this book: AltaVista, DeepDyve, Google, IBM, Microsoft, Powerset, (Xerox) PARC, University of California Berkeley, Yahoo, and Zvents. No compensation has been received or is expected in exchange for mentioning these organizations' products or ideas in this book.
I wish to thank the following people for commenting on this manuscript. Bob Glushko made extensive comments on an early draft which led to my writing a single chapter that can be read in isolation (the Design chapter). I am grateful to Ben Shneiderman, who was generous with an extensive conversation about the visualization chapters. Daniel Russell provided invaluable feedback on the Design, Query Specification, and Results Presentation chapters, and Dan Rose and Anne Aula provided detailed comments on the Evaluation chapter. I am also grateful to Omar Alonso, Anne Aula, Stephen Few, Greg Linden, Gary Marchionini, Avi Rappoport, and Jamie Teevan for comments on other chapters.
I would like to thank the hundreds of former masters students who have taken my courses in User Interface Design, Information Visualization, and Information Organization and Retrieval. The more than 50 projects I oversaw in the User Interface Design course were invaluable for deepening my understanding of how the interface design process unfolds, and the pitfalls as well as the successful paths towards good design. I also thank the former masters students, PhD students, and postdocs who worked with me on search interface research projects: Anna Divoli, Ame Elliott, Jennifer English, Melody Ivory, Kevin Li, Preslav Nakov, Ariel Schwartz, Rashmi Sinha, Emilia Stoica, Kirsten Swearingen, Michael Wooldridge, and Ka-Ping Yee.
I would like to thank the University of California, Berkeley, for granting me a sabbatical which allowed me to finish the writing of this book, Lauren Cowles at Cambridge University Press for acting as my editor and liaison to the publisher, and Ricardo Baeza-Yates and Berthier Ribiero-Neto, whose request for a chapter for the revision of Modern Information Retrieval led me to write an entire book.
Closer to home, I thank my brother Ed for persistently encouraging me to write a book, my sister Dor for inspiring me by finishing her first novel this year, my parents for surrounding me with books and a love of words and science, and Emmi for frequent play breaks. Finally, I thank Carl for being there throughout, in the best way possible.