Fabio Paternò

Software Architecture Modelling: Bridging Two Worlds Using Ergonomics and Software Properties

The process of designing and constructing user interfaces is critical for building systems that satisfy the customer’s needs, both current and future. This process includes the original design of the interface, the implementation of the system, and the modifications to the operational system. These modifications are endemic in interactive systems. Since the user interface can account for approximately 50 per cent of total life cycle costs (Myers 89), the software engineer has a vested interest in constructing a user interface that both satisfies the customer and is constructed using the best available tools and techniques. In addition, the increasing complexity and size of software systems require sound engineering principles and frameworks to formally structure the design process into multiple but consistent perspectives.

Specifying a Web Browser Interface Using Object-Z

A specification describes the functions that a system provides to its users without explanation of how those functions are implemented. A user interface specification is a description of a system from the perspective of its users (i.e., a description of the view of the system that is provided to users); such a specification is useful for developing user interface designs. A specification of a user interface indicates those functional aspects of the interactive system (data and operations) that are perceivable by the user and the logical organisation of information provided by the interface.

TLIM: A Systematic Method for the Design of Interactive Systems

The TLIM method is the result of some years of research in the area of model-based user interface design, development and evaluation with formal support which has been developed at the User-Centred Design Group of CNUCE.

Electronic Gridlock, Information Saturation and the Unpredictability of Information Retrieval over the World Wide Web

In 1980 there were approximately one hundred computers attached to the Internet. In 1990 there were one hundred thousand. In 1994, the number of systems connected to the Internet exceeded one million. A recent estimate placed the number of Internet users at just over twenty-five million (Schofield 94). Hundreds of sites in many different domains provide access to a vast range of information sources. The growth of these information sources and the development of mass-market browsers has encouraged the active participation of new groups of users (Berners-Lee et al. 92).

Towards an integrated proposal for Interactive Systems design based on TLIM and ICO

The importance of applying formal methods in the design and development process of Interactive Systems is increasingly recognised. However it is still an open issue the identification of systematic methods able to support designers and developers in specifying and demonstrating properties of user interfaces. TLIM and ICO are two formal methods which have been used for this purpose with interesting results. They address similar concepts but also have different features which allow us to consider useful their integrated use to obtain synergistic and complementary results. In this paper we show their application to some examples in order to discuss similarities and differences and we outline a proposal for their integrated use.

Introduction to the special issue on interface issues and designs for safety-critical interactive systems: when there is no room for user error

Software is increasingly being used to control safety-critical systems. Much research since Levesons fundamental article Software Safety: Why, What, and How (ACM Computing Surveys 18, 2 (1986), pp. 125163) has focused on ways to reduce or avoid software failures. However, the reliability of even the best-engineered software can be undermined by its user interface. Indeed, interface design for safety-critical interactive systems poses special challenges to the human-computer interaction community. This special issue addresses the challenge of analyzing, designing, and building reliable and usable safety-critical interactive systems. From a pragmatic point of view a safety-critical system is a system for which the cost of a failure is more important than the cost of developing the system. Safety-critical interactive systems add the human dimension to a software system by putting control into the hands of a human operator. Prominent examples of such control systems include nuclear power plants, railways systems, airplane cockpits, and military systems. Recent years have seen much effort put into the reengineering of the control system that is well represented in this special issueair traffic control. When compared to office automation systems, human-computer interac-tionfor safety-critical interactive systems is both familiar and different.For instance, the management of a functionality like undo, that can beseen as a usability issue in an office automation system, can become acritical functionality when the user interacts with a safety-critical system. The three articles in this special issue provide three snapshots for how human-computer interaction issues play out in the broader field of safety-critical interactive systems. In the first article, Is Paper Safer? The Role of Flight Strips in Air Traffic Control, Wendy Mackay provides adetailed ethnographic study on how air traffic controllers work. As in Mackays article, the case study entails en-route air traffic control. An important contribution of this article is a method for an integrated analysis of three important methods of this field: task performance, analysis of user deviation and consequent hazard, and cooperation among users. Each of the three articles deals with the analysis and design phases of safety-critical interactive systems. If changes are to be made to large, complex, safety-critical control systems, the changes must be made early in the development lifecycle, where redesign in response to identified problems is feasible.This special issue arose from a CHI98 Workshop organized by Palanque and Paterno´ (“Designing User Interfaces for Safety-Critical Systems&rdquo, SIGCHI Bulletin 30, 4). The three articles included in this special issue were selected from more than a score of papers received. The editors thank and acknowledge their debt to the many qualified external reviewers from several countries who have helped select and improve (through their comments) the contributions in this special issue.

Development of a WWW Browser Using TADEUS

Users expect advanced interactive applications to be easy-to-use and easy-to-learn. Today users expect to be able to sit down and use software without spending their time reading manuals. But such user interfaces are hard to design and implement. Different studies have shown that an average of 48% of the code of an application is devoted to the user interface, and that about 50% of the implementation time is devoted to implementing the user interface portion (Myers & Rosson 92). As user interfaces become easier to use, they become harder to create. User interface developers need tools which provide a rich support for the development of advanced user interfaces.

How to Model a Web (Without Getting Tangled in Nets)

This is not a chapter about assessing the “design” of the World Wide Web, or human issues in the interface of a browser. And it certainly does not put forward a case that the design of Web browsers could somehow be improved through the use of formal specification techniques. Rather, the aim of this chapter is simply to demonstrate how, in the early stages of designing an interactive system, it is possible to use an existing state-based specification language to capture some of the issues concerned with what the system might do. In this context, a WWW browser serves as a convenient example.

A Component-Based Approach Applied to a Netscape-Like Browser

A wide variety of toolkits13 have been developed to allow user interface designers to prototype and build graphical user interfaces (GUIs) rapidly. Unfortunately it is usually difficult to verify that GUIs built with such toolkits behave as intended (in the sense that they possess certain formally expressed properties); this is of concern for safety- and security-critical applications. In this chapter we describe an approach to structuring GUI-based applications that helps address this issue and illustrate its possible use in the validation of behavioural properties. The approach that we propose can be adapted for use with most existing toolkits.

Specifying History and Backtracking Mechanisms

Whereas most of the chapters in this book are specifying the same Netscape-like WWW browser from different perspectives, this chapter takes a single element of a browser, the history mechanism, and compares the formal specification of this aspect of four different hypertext browsers: Netscape Navigator, HyperCard, Microsoft Windows Help system and Think Reference.