From James D. Anderson and José Pérez-Carballo “Information Retrieval Design”
● purpose of this book : 1
The primary purpose of this book is to lay out and describe components and attributes of information retrieval (IR) databases and the range and variety of options for each of them, together with the implications of choices. The objective is to encourage designers to weigh options for each component or attribute in relation to the needs and preferences of users, so that the best possible design will result.
● impetus for design : 2
The inspiration or impetus for a new IR database is generally the perception of a group of potential users who have information needs or desires that are not well met, or a collection of documents containing valuable messages that should be made more accessible.
● precursors to design : 3
Before high quality design can begin, the designer must know a great deal about the potential users of the IR database — what kinds of questions they will want to bring to the database and their level of experience with and preferences for IR database searching. Designs for children, for example, will generally be quite different than designs for subject specialists. Also the operational work/action and cultural domains of users’ information seeking and use are very important. The domain of an adult seeking novels or movies for entertainment, or messages for spiritual enlightenment, will be very different compared to the literary or film critic’s domain or the professional theologian’s domain, which in turn will differ from the domain of the sociologist studying religious behavior. Cultural domains may also impact information needs and desires. The situational domains of high-income urban African American lesbians are likely to differ from those of low-income rural subsistence farmers in a barely developing country.
● collections of documents as impetus for design : 4
If the impetus for design is a collection of documents, the designer will want to know quite a bit about the nature of these documents and their content (their messages!), and also about the kinds of users that would find them useful, and the operational or cultural domains of these users.
● user needs assessment : 5
Thus, the assessment of users and their needs, desires, information-searching preferences and the work/action/cultural domains of their information seeking behavior is the first step for any IR design project. This important preliminary step falls outside the scope of this book. For help in designing and implementing a users’ needs assessment, a good place to start is with Information tasks: toward a user-centered approach to information systems, by Bryce L. Allen (1996). The importance of domain analysis is stressed by Birger Hj�rland in his Information seeking and subject representation: an activity-theoretical approach to information science (1997).
● attributes and components of IR databases : 6
Twenty key attributes or components have been identified as most important in the design of IR databases. Each will be treated in a separate chapter in an order that seems to make sense with respect to the impact of choices on other options.
● list of attributes and components of IR databases : 7
Here is the list of these twenty attributes and components:
1. Subject scope and domain: What topics correspond to the kinds of questions that users will want to ask of the IR database? What is the operational or cultural domain to be targeted by the IR database? What is the activity or cultural domain of the intended users?
2. Documentary scope: What kinds of messages, texts, and documents will provide answers or responses to user needs and desires? What features of messages, texts, and documents should be searchable?
3. Documentary domain: Where and how can these documents be found?
4. Display media: What media can most effectively display the IR database to users? Which media can provide the most effective searching experiences?
5. Documentary units: What are the most effective message units to convey answers and responses to users?
6. Indexable matter: What portions of documentary units should be analyzed for indexing and retrieval?
7. Analysis methods: How can human intellectual methods and automatic computer-based methods be used most effectively for analyzing the content, meaning, and applications of messages, texts, and documents?
8. Exhaustivity of indexing: How detailed should indexing be? Will users be more interested in maximum retrieval of useful messages, or in minimum retrieval of uninteresting or useless messages?
9. Specificity of indexing terms: How closely should indexing vocabulary match the topics and features of messages and the questions of users? How large should the indexing vocabulary be?
10. Displayed and non-displayed indexes: What kinds of searching opportunities should be provided? Will users want to browse or inspect displayed arrays of topical and feature headings, or will electronic computer matching techniques be sufficient?
11. Syntax for index headings and search statements: How should terms be combined in headings for effective browsing and visual inspection? How should terms be combined and manipulated for effective searches by means of electronic computer matching techniques?
12. Vocabulary management: Should users be expected to find the best terminology for their questions on their own? Or should they be assisted with suggestions for synonymous or equivalent terms, clarification for terms with multiple meanings, and suggestions for related terms?
13. Surrogation (representation) of documentary units: How can messages, texts, and documents be best described for users?
14. Locators for documentary units: How will message/text/document descriptions be linked to the actual, full documents that they represent?
15. Surrogate displays: How should message/text/document descriptions be displayed in various situations?
16. Arrangement of displayed indexes: If the IR database provides for browsing or visual inspection of topical or feature heading arrays, how should these arrays be arranged?
17. Size of displayed indexes: Are there constraints that will limit the size of a displayed index? How can such an index be designed to fit the available space?
18. Search interface: How can self-evident and easy-to-use displays be designed to facilitate effective searching of the IR database?
19. Record structure: How should message/text/documentary units be represented within the IR database? What data should be recorded, and how should it be organized within the database record?
20. Full-text display: How can full-text documents be effectively displayed for examination?
● interaction among design options : 8
As Bates (2002) has reminded us, “each design element or layer in an information system interacts with every other design layer in a synergistic, neutral, or conflicting manner. This cascade of interactions culminates in the interface, where all the prior interactions have either worked to produce effective information retrieval or to produce a hodgepodge of system elements working at cross purposes” (p. 381).
● sequence of design decisions : 9
The interaction and influence of design decisions has been the determining factor in suggesting an effective sequence for considering design options in this book. First comes decisions regarding the subject scope and domain, based directly on user needs assessments. Documentary scope and domain should be determined in manner that meets the goals of the subject scope and domain. IR database display media and policies for indexing (size of documentary units analyzed, appropriate indexable matter, analysis and indexing methods, exhaustivity of indexing, and specificity of terminology) should all reflect the nature of the subject and documentary scope and domain (and the needs, desires, and habits) of users. Similarly, searching and browsing options and the display of documentary surrogates (electronic searching versus displayed index examination, syntax for combining terms in displayed indexes or search statements, vocabulary assistance and management, the design and display of surrogates, and the use of locators) need to be in synch with earlier decisions. The design of browsable indexes, both alphabetical and classified, and the integration of vocabulary assistance are crucial elements of the search interface. Finally, the design of the documentary unit record format must support all previous design choices, and for full-text IR databases, the display and manipulation of retrieved full-text documents should further assist users in finding exactly what they want.
It is hoped that the ordering of design decisions, and the discussion of each option and its impact on IR effectiveness, will help achieve the goal voiced by Bates (2002): “For effective information retrieval to occur, all layers of a system must be designed to work together …” (p. 398).