COGNITIVE SYSTEMS, Volume 4, double-issue 3&4
Contents and Abstracts

to journal index...

Chr. Stary & G.J. Dalenoort
Introduction pp. 257-260

G.J. Dalenoort,
On the analysis of tasks for autonomous systems pp.261-273

P.H. de Vries and K. van Slochteren,
Text-processing tasks in autonomous networks pp.275-287

P. Markopoulos and S. Gikas,
Formal specification of a task model and implications for interface design pp.289-310

Ch. Stary,
Barriers to knowledge engineering for adaptive user interfaces pp.311-333

A. de Haan,
Computational modelling and conversational human-computer interaction pp.335-352

G. de Haan,
How to cook ETAG and related dishes: uses of a notational language for user-interface design pp.353-379

P. Slavik,
Grammars and rewriting systems as models for graphical user interfaces pp.381-399

J. Sefranek and M. Kravcik,
A model of collaborative writing pp.401-428

Balint, L.
On the perspectives of HCI design: Open questions and potential trends pp.429-448

G.J. Dalenoort
Dept. of Psychology, University of Groningen, The Netherlands
On the analysis of tasks for autonomous systems

The concepts we normally use for the description of information- and knowledge systems are specially conceived for constructed systems, in which information is localized, and in which there usually is a clear distinction between data and procedures. In distributed or collective information systems the knowledge and skills are not represented in a localisable manner. What are the consequences for the analysis of tasks ? Many of our conceptions must be adapted before we can hope to obtain an understanding of such systems: autonomous learning, control, a uniform representation for data and procedures.

P.H. de Vries and K.R. van Slochteren
Department of Psychology, University of Groningen, The Netherlands
Text-processing tasks in autonomous networks

A general representation of cognitive processes is proposed and contrasted with functionalist and connectionist approaches. Its fundamental feature is that it is based on principles of self-organization. Within the representation, several aspects of a model of text-processing and man-machine interaction will be discussed. In the model the autonomous development of a system is emphasized, in particular the transition from a stage in which information-processing is done without the help of machines to one in which it is. The representation of sequence in an autonomous network, which is an important topic for models of text-processing, is discussed.

P. Markopoulos and S. Gikas
Queen Mary and Westfield College, University of London
Formal specification of a task model and implications for interface design

This paper discusses the use of models of task knowledge in the design of user interfaces. For the majority of cognitive models it is not clear how they may contribute constructively to design. Even methods specifically geared to support the prescriptive use of task models do not address the central question of how the task model should relate to the constructed system. This relation can be characterised as the task conformance of an interactive system. This paper presents a scheme for the formal specification of some aspects of user task knowledge based on a theory of users' knowledge of their tasks. The formality helps clarify some questions relating to the nature of task conformance and of the implications for the task based design of interactive systems.

Ch. Stary
University of Linz, Department for Business Computing, Communications Engineering
Barriers to knowledge engineering for adaptive user interfaces

Knowledge engineering has become crucial for the enhancement of user interfaces with adaptivity towards end user tasks and cognitive models. In this paper several limitations of knowledge–engineering techniques for the development of task– and user–oriented user interfaces are discussed: the reliability barrier, the specification barrier and the acquisition barrier. The reliability barrier concerns the reliability of cognitive models that are used as representations of mental models to become part of adaptive user interfaces. The mapping and use of mental models to components of applications are investigated. The specification barrier concerns structural and epistemological deficiencies of knowledge engineering techniques when features of the user interface have to be represented at an implementation–independent layer. The acquisition barrier addresses the process of mutual understanding between users, developers and tools in the course of (artificial and natural) knowledge acquisition. In the paper examples for these barriers are given and ways out of some dilemmas are sketched.

Ab de Haan
Nijmegen Institute for Cognition and Information (NICI) Nijmegen, The Netherlands
Computational modelling and conversational human-computer interaction

Reciprocal understanding can be a vehicle for the coordination of conversations. Partners in a conversation with a similar functional architecture may rely on projective understanding. Theoretical understanding may however be necessary for partners with a dissimilar functional architecture, in which case contracts or common grounds may coordinate the conversation. These contracts can both be viewed from an evaluation and from a design perspective and even make it possible to simulate to a certain extent projective understanding on the basis of theoretical understanding.

Geert de Haan
Department of Computer Science, Vrije Universiteit, Amsterdam, The Netherlands (now at Origin, Eindhoven)
How to cook ETAG and related dishes: uses of a notational language for user-interface design

Extended Task-Action Grammar (ETAG) is a formal modelling notation to describe user interfaces for design purposes, in terms of the knowledge a competent user has to have to use the interface to perform tasks. This paper introduces the most important ideas underlying ETAG and briefly describes its place in the historical context of formal modelling in HCI. Next, the structure, contents and creation of ETAG models is extensively discussed. We briefly discuss some of the uses of the notation, and the structure of ETAG-based user interface design as a method to design user interfaces. Finally, we discuss the relation between Software Engineering and HCI design, and draw several conclusions about ETAG, ETAG-based design and the particular approach they exemplify.

Pavel Slavik
Department of Computer Science, Czech Technical University, Czech Republic
Grammars and rewriting systems as models for graphical user interfaces

This paper deals with the problem of formal description of graphical user interfaces. The formalisms used nowadays for the description of user interfaces in general, were mainly developed in times when user interfaces were keyboard-oriented and the communication between user and system had an alphanumeric character. Since graphical user interfaces have become a sort of standard in the HCI field, it is necessary to formalize both the graphical input and the graphical output in order to create an integrated model of a graphical user interface. Some aspects of graphical communication are discussed together with an overview of formalisms used.

Jan Sefranek, Milos Kravcik
Faculty of Mathematics and Physics, Comenius University, Bratislava, Slovakia
A model of collaborative writing

The paper presents a cognitive model of writing. The model is based on a concept of a heterogeneous multi-layered semantic network. The network forms the knowledge base part of the model. The other basic component of the model is a text base. We discuss the operations on the network. Reading and writing, the central interests of our model, are considered as complex processes interconnecting the knowledge base and the text base. A special concern is the aspect of collaborative writing. The basic features of the cognitive model are preserved in a conceptual model (its position is an intermediate one between the cognitive model and a specification of implementation). Its main role is to provide some data structures (a data model) and a control structure appropriate for the design of the system. Our aim is to preserve the basic features of our cognitive model in terms of both the data structures and the control structure. The document concept plays the role of the basic data structure. We use it in a very general sense. The control structures of the conceptual model are supposed to be components of an environment called MONSTER.

Lajos Balint
Hungarian Academy of Sciences, Budapest
On the perspectives of HCI design: Open questions and potential trends

The design of human-computer interfaces is investigated in view of their applications in complex man-machine systems. After a general introduction outlining the human-computer interfacing problem, and a brief analysis of the key development and application difficulties as well as constraints associated to human-computer interaction and interface construction, a short overview of specific aspects of designing human-computer interfaces is given. Next, a somewhat more detailed elaboration of these aspects follows, covering a bunch of new directions and evolving methodologies of how to achieve well working interface solutions for complex man-machine systems. It is emphasized that some problems arising in the construction of the interfaces can be handled by exploiting the potential advantages provided by promising new ways of abstract formal modeling, direct decomposition of the specifications, modular design, human-centredness, adaptivity, application orientedness and restricted interaction schema.

to journal index...

ESSCS european society for the study of cognitive systems