Introduction:

With the continuing evolution and convergence of previously disparate technologies around electronic commerce, the World Wide Web is increasingly pervasive in the corporate value chain. Most business processes, from procurement and inbound logistics to marketing and after-sales service, create and use information (Porter 1998, 167f.). They are connected via networked information systems, which have become the basic infrastructure for global transaction-oriented applications. Every transaction process occurring in such an electronic marketplace goes hand in hand with the access, absorption, arrangement, and selling of information in heterogeneous ways (Zakon 1999).

Despite these technological and organizational changes, the customers’ information needs provide a uniform purpose for Web information systems. Accordingly, developers of such systems have to analyze the requirements of the newly empowered and technologically savvy customers in order to exploit the potential of online trading of information, services, and physical goods.

This book presents a methodology for analyzing and developing Web information systems that considers the structural changes in electronic markets outlined above. Chapter 1 defines the field and introduces the term “ergodic literature” for textual material whose access and utilization requires non-trivial efforts by the reader. It identifies navigational and textual features of interactive Web applications, a subset of ergodic literature, and compares these features with those of printed media. An analysis of the multifaceted term “interactivity” then sets the stage for Chapter 2, which investigates the life-cycle economies and diffusion characteristics of traditional and Web-based applications. The subsequent sections delineate the evolution of electronic markets in general and the World Wide Web in particular from both Darwinian and methodological perspectives. To investigate this process, an evolutionary framework based on system adaptivity  and the underlying communication patterns is introduced. It classifies Web information systems and analyzes their ability to support the various phases of electronic business transactions. Only advanced system architectures can fully leverage the potential of the World Wide Web to deliver additional customer value. The framework specifies the following four categories: Static Web information systems that provide basic information of limited value for the average customer (Chapter 3), interactive systems that enable explicit customer feedback and transaction processing (Chapter 4), adaptive systems that instantly and automatically generate the hypertext structure according to embedded user models (Chapter 5), and agent-mediated architectures that allow individual negotiations regarding product and non-product attributes (Chapter 6).

Chapter 3 summarizes the characteristics of static Web information systems. These collections of rudimentary hypertext documents are still common for many small and medium-sized companies. More in line with current business practices of larger organizations, Chapter 4 focuses on gathering customer feedback and its immediate analysis within an iterative cycle of design, implementation, usage, and analysis. The emergent attributes of modern organizations suggest such an evolutionary approach to Web development. Analytic activities are no longer captured within the early stages of a system’s life-cycle but represent a continuing task of system maintenance. Analysis, operation, and maintenance activities become parallel but highly interrelated processes. Cyclical planning is rendered obsolete, since the results of the ongoing analysis are continuously fed into the maintenance phase. Section 4.2 outlines the current state of Web engineering with special regard to design methodologies, conceptual modeling approaches, and related visualization techniques. It introduces a symbolic modeling language for the construction of both reference and customized models during the development process of commercial Web information systems (Section 4.3). This graphical notation helps analysts to visualize individual and aggregated access patterns of online customers derived from log file data of corporate Web servers. It enhances the limited, statistically oriented representations of commercially available Web-tracking software with a map-like overview similar to customer tracking in traditional retailing outlets (Section 4.4).

Numerous tools support the structured design of Web information systems, ranging from rudimentary layout products for individual documents to sophisticated Web site management solutions that facilitate conceptual authoring-in-the-large. Structured analysis, by contrast, has largely been neglected by both theory and practice but belongs to the most relevant questions in formulating business strategies for electronic commerce. This comes as a surprise, since analysis is less constrained by the technical limitations of existing architectures. Section 4.5 intends to fill this gap by suggesting a methodology for the automated analysis of Web information systems. The chapter summarizes empirical results covering several business sectors, namely information technology, travel and tourism, retail banking, and environmentally oriented non-profit organizations. Classifications, comparative sector assessments, longitudinal studies, and exploratory textual analyses help practitioners chart the industry evolution and compare the performance of their own Web information system with those of competing organizations.

Chapter 5 introduces adaptive Web information systems that promise a sustainable competitive advantage in an environment where information redundancy becomes increasingly evident. Such a competitive advantage can only be achieved through customizing products, services, and communications. Key elements of customization are the separation of product and process life-cycles as described in Section 5.1. The resulting “classic loop of adaptation”, delineated in Section 5.2, is founded on the continuing gathering of information regarding the customers’ current needs and preferences. Sections 5.3 and 5.4 compare explicit and implicit acquisition methods and describe ways to combine the various information sources and transform the resulting body of knowledge into consistent user models. On the basis of such embedded user models, adaptive Web information systems can automatically generate the content of documents, the primary navigational system comprising links between and within these documents, and supplemental navigational systems such as index pages, trails, guided tours, or interactive site maps. Many prototypes that incorporate adaptive components are developed without a clear model of the components’ functionality. Therefore, Section 5.5 provides a conceptual guideline for the development process by classifying the various mechanisms into content-level, link-level, and meta-level adaptation.

In the fourth and last stage of Web evolution, which is portrayed in Chapter 6, efforts to automate and optimize electronic business-to-consumer transactions gradually transfer certain tasks from adaptive Web information systems to agent-mediated architectures. Many deployed applications use content-based information agents and collaborative filtering systems for general information retrieval or specific product recommendations (Sections 6.1 and 6.2). By contrast, most transaction agents are only available as prototypical implementations up until now (Section 6.3). Transaction agents allow multidimensional negotiations regarding a variety of product and non-product attributes. Once the necessary infrastructure is in place (Section 6.4), their customizability and remarkable flexibility promise to change the inherent characteristics of doing business electronically.


Table of 
Contents:


1 Introduction

 

1.1 Formal Definitions of Text

1.1.1 The Metaphysics of Written Text

1.1.2 Technological Implications for the Discursive Space

1.2 Ergodic Literature

1.3 Hypertext and Hypermedia

1.3.1 History

1.3.2 Adaptability versus Adaptivity

1.4 Web Information Systems

1.4.1 Theories of Communication

1.4.2 Media Characteristics of the World Wide Web

1.4.3 Evolving Web Genres

1.4.4 Structure and Navigation

1.4.4.1 The Myth about Linearity

1.4.4.2 The Labyrinth Metaphor

1.4.5 Textual Characteristics

1.4.5.1 Differences between Printed and Electronic Media

1.4.5.2 Mapping Textons to Scriptons

1.4.6 Interactivity: An Ambiguous Term

1.4.6.1 Media Permeability

1.4.6.2 Communication Patterns

1.4.6.3 Discourse Structure

 

2 The Evolution of Electronic Markets

 

2.1 Darwinian versus Technological Evolution

2.1.1 Alternative Life Cycle Economies

2.1.1.1 Prototyping

2.1.1.2 Organizational Emergence

2.1.1.3 Types of Maintenance

2.1.1.4 Stable System Drag

2.1.2 Electronic Business Ecosystems

2.1.2.1 Autopoietic Self-Reference

2.1.2.2 Chronological Development and Demarcation

2.1.2.3 Critical Mass of Content and Adopters

2.2 A Brief History of Web Adaptivity

2.2.1 Stage Models for Describing Web Evolution

2.2.2 Diffusion of Web Technology

2.2.2.1 Temporal Pattern of the Diffusion Process

2.2.2.2 Rate of Adoption

2.3 Electronic Market Transactions

2.3.1 Evaluating Transactional Infrastructures

2.3.2 Redefining the Customer Delivered Value

2.3.3 Extending the Evaluation Model

2.3.4 Electronic Customer Relations

2.3.4.1 Customizing Corporate Communication

2.3.4.2 Gathering and Representing Customer Information

 

3 A Static World

 

4 The Emergence of Interactivity

 

4.1 Overview

4.1.1 Development Process

4.1.2 Database Connectivity

4.1.3 Presentational versus Semantic Markup Languages

4.1.4 Traversal Functions

4.2 Design: The Current State of Web Engineering

4.2.1 Processing and Visualizing Complex Information

4.2.2 Cooperative Web Development

4.2.3 Review of Existing Design Methodologies

4.2.4 The Extended World Wide Web Design Technique

4.2.4.1 Information Object Types

4.2.4.2 Diagram Structuring

4.2.4.3 Navigation Design

4.2.5 Application of eW3DT (Electronic Shopping Mall)

4.3 Implementation: Industry-Specific Reference Models

4.3.1 Definition and Overview

4.3.2 Deducing Customized and Technical Models

4.3.3 Removing Communication Barriers

4.4 Usage: Visualizing Topology and Access Patterns

4.4.1 Categories of Perceptualization

4.4.2 Observing and Visualizing Human Behavior

4.4.3 Units of Analysis

4.4.4 Tool Support


4.5 Analysis: Gathering, Extracting, and Processing of Multidimensional Web Data

4.5.1 Overview and Analytical Objectives

4.5.2 Manual versus Automated Evaluation

4.5.2.1 Previous Web Evaluation Projects

4.5.2.2 Web Classification Frameworks

4.5.2.3 Integrative Analysis of Content and Form

4.5.3 Methodology

4.5.3.1 Sampling Process

4.5.3.2 Mirroring and Archiving Web Data

4.5.3.3 Structural Parameters

4.5.3.4 Textual Parameters

4.5.3.5 Combining and Preprocessing the Extracted Parameters

4.5.4 Validating Manual Evaluations of the LSE/Novell Survey

4.5.4.1 Introduction to Neural Networks

4.5.4.2 Test Design

4.5.4.3 Results

4.5.4.4 Discussion

4.5.5 Snapshot Analysis (Classification)

4.5.5.1 Partitioning Clustering

4.5.5.2 Results

4.5.5.3 Discussion

4.5.5.4 Hierarchical Clustering

4.5.6 Cross-sectional Analysis

4.5.6.1 Structure and Navigation

4.5.6.2 Interface Representation

4.5.6.3 Content and Strategy

4.5.7 Longitudinal Analysis

4.5.8 Exploratory Textual Analysis

4.5.8.1 Types of Content

4.5.8.2 Common Methods for Describing and Analyzing Textual Data

4.5.8.3 Word List Generation

4.5.8.4 Lemmatization

4.5.8.5 Keyword Analysis

4.5.8.6 Concordance Analysis and Collocation

4.5.8.7 Analysis of Contingency Tables

 

5 Adaptive Solutions

 

5.1 Separating Product and Process Life Cycles

5.2 Classic Loop of Adaptation

5.3 Collecting Information About Users

5.3.1 Available Online Sources

5.3.2 Explicit Acquisition Methods

5.3.3 Implicit Acquisition Methods

5.4 User Modeling

5.4.1 Characteristics of User Models

5.4.2 Typology of the Extra-Linguistic Context

5.4.3 Building and Initializing User Models

5.5 Adaptive Web Presentation Techniques

5.5.1 Annotations

5.5.1.1 Historical Background

5.5.1.2 Automated Document Annotation

5.5.2 Content-Level Adaptation

5.5.2.1 Deictic Expressions

5.5.2.2 Expressing Shared Meaning

5.5.2.3 Techniques for Generating and Arranging Scriptons

5.5.3 Link-Level Adaptation

5.5.4 Meta-Level Adaptation

5.5.4.1 Site Maps

5.5.4.2 View Types

5.5.4.3 View Transformations

5.5.4.4 Implications for Hypertext Navigation

 

6 Agent-mediated Architectures

 

6.1 Content-based Information Agents

6.2 Collaborative Filtering

6.3 Transaction Agents

6.4 Developing Digital Agents

6.4.1 Agent Communication Models

6.4.2 Traversal Functions of Agent-Mediated Architectures

6.4.3 Modeling Multilateral Interactions

 

7 Conclusion and Outlook

 

8 References