The Cognitive Dimensions of Notations framework (CDs) provides a powerful vocabulary for discussing the usability of programming languages, tools and environments. Although originally proposed as a discussion tool for designers, they have recently been used to design questionnaires intended for system users evaluating the usability of the programming tools they use (Kadoda et al., 1999) We extend their approach: rather than develop questionnaires tailored to specific systems, we propose a generalised questionnaire in which the definitions of the CDs themselves are offered to the users, and respondents can choose for themselves the features of the system that they wish to criticise. This questionnaire has been completed by an extremely diverse range of users, showing that a generalised CDs questionnaire indeed a suitable tool for user evaluation. Not surprisingly, some problems emerged as well. Introduction The Cognitive Dimensions of Notations framework (CDs) provides a powerful vocab...

This paper discusses some observations on the differences between what the designers of theorem proving assistants (TPAs) think about the systems they designed and what the users of those TPAs actually find. A questionnaire based on the cognitive dimensions framework was sent to designers and users of a sample of TPAs. The aim of this work is to be able to identify specific areas that designers of TPAs need to devote extra attention to. It was observed that the cognitive dimensions of closeness of mapping, visibility and juxtaposability and perceptual cues are of particular significance.

The Cognitive Dimensions framework has inspired research both more and less varied than expected. In this paper we revisit the original aims and briefly describe some subsequent research, to consider whether the original aims were too austere in rejecting knowledge-based dimensions; whether the dimensions can be shown to have real-world relevance; and whether their definitions can be improved, either piecemeal or by refactoring the entire set. We mention some issues that remain unexplored, and conclude by describing two different ventures into defining clear procedures for real-life application, operating in very different milieux but both accepting that the framework should be developed from its original formulation. 2 1.

The field of HCI is gradually achieving sound theoretical descriptions of the activities, context and cognition of computer system users. How do these research results get applied by the people who design new user interfaces? Although we have theoretical descriptions of the activities of system users, we have fewer descriptions of the design activities of user interface designers. There are certainly theory-based design methods prescribing the things that designers ought to do. Almost all of these have been developed from the perspective of descriptions of the user, rather than from consideration of the needs of designers – the current vogue for “user-centred design ” clearly expresses this emphasis on the user. According to this perspective, if user interface designers are to apply research into user needs, those designers must be able (and interested) to interpret and apply theoretical results. Alternatively, there are also many popular approaches to user interface design that have minimal theoretical grounding. Such design methodologies generally attempt to present designers with a checklist (Nielsen & Molich 1990) or a procedural list of design activities (Wharton, Rieman et. al. 1994) that will generate a good design. The reduction of design to a checklist or a predefined procedure is widely proposed in other areas of software development,

Information Sciences and Technology iii Intelligent agents and cognitive models are useful for a number of purposes. Unfortunately, limited theory-based tool and language support for the creation of intelligent agents has made it difficult for modelers to create, debug, and reuse agent software. This dissertation explores how to make it easier to create intelligent agents, and especially cognitive models, by taking advantage of established software engineering principles. The benefits of applying software engineering principles to intelligent agent development is demonstrated with the creation of a high-level language and development environment that embodies these principles, and with an evaluation of this language and environment, in use, by students and cognitive modelers. TABLE OF CONTENTS