HCI Bibliography Home | HCI Conferences | BCSHCI Archive | Detailed Records | RefWorks | EndNote | Hide Abstracts
BCSHCI Tables of Contents: 85868788899192939495969798000102

Proceedings of the HCI'91 Conference on People and Computers VI

Fullname:Proceedings of the Sixth Conference of the British Computer Society Human Computer Interaction Specialist Group
Note:People and Computers VI: Usability Now
Editors:Dan Diaper; Nick Hammond
Location:Heriot-Watt University, Edinburgh
Dates:1991-Aug-20 to 1991-Aug-23
Publisher:Cambridge University Press
Standard No:ISBN 0-521-41694-9; hcibib: BCSHCI91
Papers:30
Pages:456
  1. Invited Papers
  2. HCI Frameworks
  3. Scenarios and Rationales in Design
  4. Groupware
  5. Graphical Interaction
  6. Evaluation
  7. UIMS and Workstations
  8. Cognitive Dimensions
  9. Applications
  10. Task Analysis
  11. Space and Location
Preface BIB 1-2
  Terry Mayes
Giving Referees Their Head BIB 3-7
  Dan Diaper

Invited Papers

Demonstrational Interfaces: A Step Beyond Direct Manipulation BIBAK 11-30
  Brad A. Myers
Direct manipulation interfaces, where objects on the screen can be pointed to and manipulated using a mouse and keyboard, are now almost universally accepted. However, some limitations of these interfaces are well known. These include the lack of programmability and the difficulty of providing abstract commands. Demonstrational interfaces can overcome these problems while still providing the benefits of direct manipulation. A "demonstrational interface" watches while the user executes conventional direct manipulation actions, but creates a more general abstraction from the specific example. For instance, the user might drag a file named v1.ps to the trash can, and then a file named v2.ps, and a demonstrational system might automatically create a macro to delete all files that end in .ps. This paper defines demonstrational interfaces, presents a number of examples, and then discusses some potential application areas.
Keywords: User interface styles, Demonstrational interfaces, Direct manipulation, Programming-by-example, Inferencing
Multimedia -- What is It and How Do We Exploit It? BIBAK 31-44
  James L. Alty
The importance of a user-centred approach to multimedia interface design is stressed. We must respond to what users wish to do rather than considering what users might be able to do. Much existing multimedia research has been disappointing partly because of a concentration on exploratory presentation approaches rather than on investigations of how media in combination can improve human-computer bandwidth. Multimedia terminology is examined and suggestions are made as to how to improve the terminology. Arguments are put forward to support more research into multimedia devices and for knowledge-based support for non-deterministic multimedia situations. Finally a set of questions is posed as a possible research agenda.
Keywords: Multimedia, Modalities, Terminology, User-centred design, Media combination, Channels, Modes, Styles

HCI Frameworks

History and Hysteresis in Theories and Frameworks for HCI BIBAK 47-55
  John M. Carroll
The contextualist critique of HCI theories and frameworks both complicates the analyst's job and enriches the analytical result by permitting, indeed requiring reference to situational details of user interaction. It is suggested that this critique be broadened to envision HCI as fundamentally historical, relativizing situational details to points in time, and taking remote direct causation (hysteresis) as the usual state of affairs. This further enrichment may have the same double-edged effect on analysis in the field.
Keywords: Frameworks, Methodology, Contextualism, History
Human Factors and Structured Software Development: The Importance of Software Structure BIBAK 57-72
  Gilbert Cockton
This paper reviews current Software Engineering practice and User Interface Management research on internal software structure. It argues that existing analytical categories in Software Engineering do not expose the structural properties emphasised by recent developments in User Interface Management. New analytical categories are introduced, and it is shown that there are direct connections between these categories and important HCI goals such as flexibility, consistency and task fit. HCI goals are thus relevant to the internal structure of software, and not just to the inputs, methods and techniques of the design and testing stages of software development methodologies.
Keywords: Structured methods, Software models and architectures, Architectural frameworks, Coupling and cohesion, User interface management, Software engineering
Users, Systems and Interfaces: A Unifying Framework for Interaction BIBAK 73-87
  Gregory D. Abowd; Russell Beale
We introduce a basic framework for the analysis of existing interactive systems which will also serve for the principled design of more usable systems. We present a simple yet effective model of an interactive system that extends previous interaction frameworks. Within our framework, the user, system and interface are all represented equally. We also present several notions of distance as qualitative measurements of the interactive features of a system based on specific tasks. These notions of distance can be formalised to give an understandable quantitative approach required for principled design and analysis.
Keywords: Framework, Analysis and design, Formal methods

Scenarios and Rationales in Design

Signature Tasks and Paradigm Tasks: New Wrinkles on the Scenarios Methodology BIBAK 91-101
  Richard M. Young; Philip J. Barnard
Scenarios are increasingly being used in HCI to explore alternative designs or assess user models. We seek to strengthen the use of scenarios within modelling methodologies by clarifying what scenarios are good for and what makes a good scenario. The first clarification concerns scenarios that are "privileged" in certain ways with respect to the modelling technique used to analyse them. A signature task is one deliberately chosen to match the capabilities of the technique. A paradigm task is one which has been thoroughly analysed and understood in terms of the technique. Perhaps surprisingly, signature tasks and paradigm tasks are often not the same. The second clarification is that although scenarios represent a particular concrete instance of human-computer interaction, some form of contrast is generally involved -- whether explicitly stated or merely implied. Good scenarios are characterised by the presence of a meaningful contrast that captures an issue and focuses the analysis.
Keywords: Scenarios, Methodology, User models
The Use of Scenarios by User Interface Designers BIBAK 103-115
  Lesley Clarke
There are many factors which drive software design. This paper focuses on the use of scenarios by designers and examines the effect this has on the design process. Scenarios are examples of user interactions with systems which are used to structure and communicate information about how a design might be used in the real world. The use of scenarios was examined in a study of designers in a commercial setting engaged in the design of an air traffic control system. The findings are used to make recommendations about how design should be done and also about the tools required to support design.
Keywords: Scenarios, Design process, Commercial design, Empirical study
Communicating Human Factors Expertise Through Design Rationales and Scenarios BIBAK 117-130
  Tom Carey; Diane McKerlie; Walter Bubie; James Wilson
This paper discusses the preliminary results and ongoing work of a collaborative research project which investigates methods to communicate the Human-Computer Interaction expertise of a human factors consulting group in a large corporate setting. The objective of the project is to extend the contribution of the human factors group. Their existing direct consultation on user-interface design will be combined with new methods which allow the results of those consultations to be widely accessible. The current experiments focus on design rationales and scenarios for user interfaces as the methods for communicating the expertise.
   In work to date, we have adapted previous research on design rationales for our new context of use. This has produced an enhanced format for recording and presenting design decisions and the reasoning process behind them. We have produced a prototype presentation system, seeded with design rationales from a large project with extensive involvement by the human factors group. This prototype is currently undergoing iterative test and refinement.
   We have also developed a framework for access to this information, by product engineers working on user interfaces, using complementary tables of contents within a hypertext space. These paradigms, or ways of looking at a user interface design problem, are themselves an important component of HCI expertise.
Keywords: Design rationale, User interfaces, Scenarios, Reasoning process

Groupware

Supporting Prediction in Complex Dynamic Systems BIBAK 133-144
  V. C. Miles; C. W. Johnson; J. C. McCarthy; M. D. Harrison
The classical perspective in the study of human-computer interaction has focussed upon single users operating single systems. Interactive dialogues in such applications are often assumed to be sequential and deterministic. These assumptions support operator predictions about the effects of their commands. Unfortunately, there are an increasing number of applications for which such assumptions are no longer appropriate. This paper examines approaches which support predictability in systems, such as groupware and process control, where execution may be neither sequential nor deterministic.
Keywords: Predictability, Process control, Groupware
Applying Temporal Logic to Support the Specification and Prototyping of Concurrent Multi-User Interfaces BIBAK 145-156
  C. W. Johnson
First order logic provides a concise and precise means of specifying interface requirements for complex systems. Designs can be constructed in terms of high level abstractions which avoid the representation of unstructured bitmaps and 'raw' devices during the early stages of the development process. Unfortunately, logic specifications provide the non-formalist with little idea of what it would be like to interact with potential implementations. This limitation can be avoided by employing executable subsets of first order logic to rapidly derive prototypes from high level specifications of interactive systems. The following pages show how this formalism must be extended in order to support the design of concurrent multi-user systems, such as joint-editors, conferencing or distributed control applications. The introduction of a temporal ordering into logic specifications provides a means of analysing concurrency and contention between multiple users of shared resources. PRELOG, a tool for Presenting and REndering LOGic specifications has been enhanced to support this investigation.
Keywords: Concurrent interaction, Groupware, Formal methods, Temporal logic
Coordination and Control for Collaborative Workstation Design BIBAK 157-167
  Mark O. Pendergast; Margaret M. Beranek
This paper presents current work on software to support collaborative workstation design. Prevailing problems with groupware systems which prevent their widespread use for software engineering tasks such as systems design, software engineering, and computer-aided design are discussed and the methods for alleviating these problems through the use of collaborative workstation design are described. The collaborative workstation design model introduces strategies for dealing with synchronous work problems such as data and view integrity, view coordination, and change arbitration as well as asynchronous work problems such as version control and presentation of design modifications.
Keywords: Computer supported cooperative work, Groupware

Graphical Interaction

Iconographer as a Visual Programming System BIBAK 171-185
  Stephen W. Draper; Kevin W. Waite
Iconographer, a tool for exploring alternative iconic representations of objects, is briefly introduced. Its own user interface, which itself is largely though not completely visual (pictorial), is described in detail: it comprises five separate interactive representations. These visualisations are then analysed in an attempt to understand their nature and limitations. This analysis is then supported by showing how Iconographer can be directly extended to an isomorphic, yet apparently quite different, task: a subset of the class of data processing programs addressed by the JSD (Jackson System Development) method.
Keywords: Icons, Visual programming, Toolkit, Visualisation, Data processing, JSD
User Input to Iconographer BIBAK 187-198
  Kevin W. Waite; Stephen W. Draper
Iconographer is a toolkit that allows designers to explore alternative iconic representations of collections of application entities in a highly interactive manner. This paper describes the development of a complementary facility for exploring alternative regimes for interacting with these generated iconic representations. This mechanism provides a flexible means for selecting and modifying entities and their attributes' values using a strict direct manipulation style of interaction. The paper considers generalised editors and virtual input devices as a means of implementing this input mechanism.
Keywords: Icons, Interaction, Input, Rapid prototyping, Direct manipulation
Wet and Sticky: Supporting Interaction with Wet Paint BIBAK 199-208
  Tunde Cockshott; David England
All current paint systems are based on the same conceptual model. This model does not actually model real paint as an artist would understand it, rather it is more akin to a potato-cut printing technique. This paper describes an interactive model of real paint, "Wet and Sticky", and its successor, "Wet and Runny". Both models provide the artist with a realistic, simulation of real, wet paint applied to a canvas.
Keywords: Paint modelling and systems, Novel interaction, Parallelism

Evaluation

The Use of Focus Groups as an Evaluation Technique in HCI BIBAK 211-224
  Patrick J. O'Donnell; Geoff Scobie; Isobel Baxter
This study examines the focus group as an evaluation technique. Many of the methods used for user requirements capture and for evaluation of usability suffer from being individual based. This leads to serious concerns over the validity of techniques. It is noticeable that in the plethora of definitions of usability and in the variety of techniques (both 'objective' and subjective) pressed on the designer for employment during user requirements capture, prototype evaluation and field evaluation, the issues of reliability and validity are not often in focus.
   This paper addresses the question of construct validity as it affects specifically the focus group in HCI. A central heating control interface from Honeywell Control Systems was subjected to prototyped based evaluation using a range of different techniques including a focus group. 20 subjects operated a predesignated task scenario on an interface and had their performance videotaped. Indexes of performance were calculated, time for sub task completion and error rate. Subjective assessments by subjects of the performance were also taken as were a range of other measures. Subjects then took part in a moderated focus group. The discussion was content analysed and measures of satisfaction/dissatisfaction constructed.
   Measures derived from the task scenario were correlated with the focus group based measures. The scenario based measures predicted the focus group indexes especially the number of criticisms uttered. However some dimensions of subject evaluation were not predicted by the scenario measures. This implies that the construct validity of the focus group does not overlap completely with that of other evaluation techniques.
Keywords: Focus groups, Evaluation, Validity
Training within an Interactive Multimedia Environment BIBAK 225-236
  Arja Vainio-Larsson
In this paper, we report on a case study to evaluate the introduction, organization and planning of education utilizing new technology in order to obtain an understanding of the different needs for users and enterprises in multimedia learning environments. Two different interactive video training packages were tested in this study. Evaluation methods comprised a combination of seminars, questionnaires, video recordings and interviews. In total 125 users participated in the study. The results show that: a mouse-based interaction technique and a 'point-and-click' dialogue certainly simplify the users' interaction with a system, but they also make it easier for users to make mistakes. If implemented as a point-and-click dialogue, direct manipulation tends to become semantically overloaded confusing the users' interaction with the system. A major drawback of the system tested was its insufficient support for navigation and browsing. The mere possibility of obtaining feedback from a system and not from another person was highly appreciated by the users and positively influenced their willingness to try different parts of the system. Finally we discuss the design of a learner-driven system that provide both a teacher-directed and a learner-directed interaction.
Keywords: Multimedia, Hypermedia, Evaluation, Learning
Guessability, Learnability, and Experienced User Performance BIBAK 237-245
  Patrick W. Jordan; Stephen W. Draper; Kirsteen K. MacFarlane; Shirley-Anne McNulty
An experimental study investigated three distinct components of usability, that account for how a user's performance with a system changes with learning: guessability, learnability, and experienced user performance (EUP). Two small experiments, involving the performance of simple editing tasks on a word processor, were used to illustrate these components. Further possible components of usability are discussed with a view to obtaining a comprehensive definition.
Keywords: Guessability, Learnability, Experienced user performance (EUP), Usability

UIMS and Workstations

The Active Medium: A Conceptual and Practical Architecture for Direct Manipulation BIBAK 249-264
  Roger Took
This paper presents a precise but general architecture (UMA) which attempts to resolve two critical but conflicting qualities of graphical user interfaces: directness and separation. This is achieved by placing central emphasis on the medium of interaction, and making this active through a dedicated user agent. The active medium allows surface interaction -- application-independent manipulation of medium objects by the user. A major strength of UMA is that it is both a conceptual and an implementation architecture, and therefore is both intuitive to the user and the application designer, and effective in rationalising the separate construction and execution of the user interface and the application.
Keywords: Interactive graphics, Interactive architectures, UIMS, Window managers, Formal models
The Development of a Visual Style for a BT X Window System Toolkit BIBAK 265-279
  Karen Mahony; Andrew Gower
This paper describes the work done in developing the visual style for a BT X Windows system graphical user interface toolkit. The approach to this design was innovative because it involved a team which had training in both graphical information and industrial design, in addition to more conventional human factors/HCI experience. This had a considerable impact on working methods, several of which were drawn from traditional design practice. The two principle design aims were to improve usability and convey an appropriate product identity and corporate image. These were achieved by appropriate and consistent representation and rendering of interactors in a range of states.
Keywords: Graphical user-interface toolkits, Graphic design, Industrial design, Design methods, Visual style, Corporate identity
A Predictive Reference Model for Use in a Speech Driven Word Processing System BIBAK 281-293
  Jill Hewitt; James Monaghan; Christine Cheepen
This paper explores the techniques necessary to build a domain specific predictive reference model which can be used to improve the performance of a speech recognition interface to a word processor. It draws on the expertise of Speech and Language Technology Group members in task analysis, linguistic analysis and data structure design. The results of initial evaluations and analysis are presented and the implications for the model design are discussed.
Keywords: Speech recognition, Tasks, UIMS, Text structure, Wordprocessor

Cognitive Dimensions

Describing Information Artifacts with Cognitive Dimensions and Structure Maps BIBAK 297-315
  T. R. G. Green
(Green, 1989) coined 'cognitive dimensions' to describe the cognitively important aspects of an information artifact for a given style of interaction, such as exploratory design, and to show how computer environments could provide tools tailored to complement notations. Cognitive dimensions provide a useful and much-needed vocabulary for discussing designs, but the problem is to define them clearly. This paper introduces 'structure maps' in which the information dependencies of many diverse notations and artifacts can be analysed in ER terms, and unexpected similarities can be brought to light. Structure maps go some way to defining certain of the cognitive dimensions: other dimensions still need to be defined by reference to psychological models.
Keywords: Cognitive dimensions, Information displays, Design representations, Entity-relationship model, Notations, Structure maps
Visibility: A Dimensional Analysis BIBAK 317-329
  David J. Gilmore
This paper presents an analysis of visibility -- a common HCI guideline which is often only loosely described. Although it might appear that visibility should be thought of as a cognitive dimension of notations (Green, 1989), my analysis separates it into three dimensions, two of which are static properties of the notation, while the third is dynamic. The two static dimensions are accessibility and salience, both of which can be examined through the use of structure maps (Green, 1991), whilst the third dimension is congruence which changes according to the user, their task and their strategy.
   Accessibility refers to the ease with which information structures can be accessed (psychologically), and this is assessed directly from the structure map and the number of different routes to certain information. Salience refers to the relative accessibility of an information structure in a display -- relative to the accessibility of other information structures in the display. Congruence reflects the extent to which the salient structure is relevant to the use being made of the display at any moment.
   An information search task is used to provide experimental evidence for this analysis, revealing that the effects of accessibility and salience on search performance and search strategy are separable. Accessibility affects speed of performance, but not strategy, whereas salience has an effect on strategy choice, and not necessarily on speed. The results also suggest that it is important to consider a display congruence, which is the match between the information structure required by the users strategy and the salience of that structure.
   The paper concludes with a brief analysis of the visibility of various programming languages, revealing that the dimensions of visibility can be assessed.
Keywords: Cognitive dimensions, Visibility, Information displays, Programming, Notations, Programming languages
Cognitive Dimensions of Design Rationale BIBAK 331-344
  Simon Shum
A design rationale (DR) is a representation of the reasoning behind the design of a system. One of the primary goals in developing a usable DR environment is minimising the cognitive overhead of representing design deliberation in a structured form. This paper examines some of the cognitive attributes of various DR notations, focussing on how generic cognitive dimensions of notations can be operationalised to clarify the relationship between DR notations and the DR authoring process. It is shown how cognitive dimensions analysis makes explicit some of the requirements for DR authoring environments, highlighting the importance of supporting intermediate DR representations; these 'rough DRs' evolve as design reasoning unfolds, to reflect new perspectives on the design space.
Keywords: Design rationale, Cognitive dimensions, Notations, Argumentation

Applications

Helping the Police with Their Enquiries BIBAK 347-358
  Andy Woods; Dermot P. Browne; John Friend
A software development project was undertaken to show how modern User Interfacing Techniques could significantly improve the usability of an existing Police System. The following reports the User-Centred approach taken to this work, including difficulties encountered. A formal evaluation was undertaken of the work. This is reported at some length as an example of the difficulties of performing usability evaluations within viable timescales.
Keywords: Empirical study, Graphical user interfaces, Style document, User-centred design, Smalltalk-80, Consistency
A Study of Conversational Turn-Taking in a Communication Aid for the Disabled BIBAK 359-371
  R. Woodburn; R. Procter; J. L. Arnott; A. F. Newell
This paper examines the potential benefits of conversational computer-mediated communication aids for the speech and hearing impaired. The importance of turn-taking protocols for effective conversation management is discussed. Results are presented of a preliminary study of turn-taking behaviour in a text-based conversational aid, and the effectiveness of an explicit turn-taking signalling mechanism is assessed.
Keywords: Computer-mediated communications, Aids for the disabled, Conversational protocols
Navigating the Interface by Sound for Blind Users BIBAK 373-383
  Ian J. Pitt; Alistair D. N. Edwards
The increasing reliance on visual forms of communication in modern computer interfaces poses severe problems for blind users. A possible solution is to make greater use of auditory communication. Speech has obvious applications, but is slow and hence not useful in situations where immediate feedback to the user is essential, such as when locating items using a mouse. Experiments have been carried out in order to ascertain the best way in which to use non-speech sounds to guide the user in locating such targets. The guiding principle has been to modulate the sounds in a manner which is as natural as possible, so that people can exploit their every-day listening skills. Some success has been achieved, particularly through the use of stereo sounds to give two-dimensional spatial sound guidance.
Keywords: Non-speech audio, Blind users, Graphical user interfaces, Navigation

Task Analysis

User Modelling: The Task Oriented Modelling (TOM) Approach to the Designer's Model BIBAK 387-402
  Dan Diaper; Mark Addison
User models as psychological models useful to HCI practitioners and system designers are discussed and a form of designer's user model based on logical, as opposed to psychological, behaviourism is proposed. A small survey of recently published work on user models suggests that many models have a weak empirical basis.
Keywords: User models, Designer's user models, Task oriented models, Task analysis, Air traffic control
Assessing the Programming Language PML as a Task Analysis Method and Product BIBAK 403-417
  Andy Whitefield; Julie Wight; Andrew Life; Martin Colbert
This paper reports a feasibility study to investigate the possible use of the programming language PML (Process Modelling Language) for task analysis. In so doing, it proposes a number of criteria for assessing task analyses. The study used PML to analyse and describe the task of producing a multi-author research proposal. To assess the suitability of PML for this purpose, a number of criteria were identified, concerning the effectiveness of: the task analysis product; the task analysis method notational support; and the task analysis method procedural support. The assessment of PML suggests that it may well be suitable for task analysis, generating a task analysis product particularly appropriate for software engineers developing multi-role systems, and offering a clear notation for a task analysis method. Its principal weakness is the lack of procedural support it provides as a task analysis method.
Keywords: Task analysis, Process modelling, Multi-user systems

Space and Location

Cognitive Representations of Space in the Design and Use of Geographical Information Systems BIBAK 421-433
  D. J. Medyckyj-Scott; M. Blades
Geographical Information Systems (GIS) are multi-user, multi-functional systems concerned with handling and analysing spatially-referenced data. While technically sophisticated they are still poor with regard to their degree of usability. Although ergonomic principles can be applied to improve the usability of such systems, a more effective approach is to reduce the amount of cognitive and behavioural modification required by the user by incorporating ideas of how they represent and think about space into the design of GIS.
   The purpose of this paper is to summarize some of the research currently being conducted into GIS design using this cognitive-behaviourist approach (i.e. user interfaces, spatial query languages, visual representations, conceptual design). The paper begins by describing our current knowledge about how people mentally represent and transform space and then describes how this knowledge is being and might be applied to the design of more usable GIS.
Keywords: Spatial representations, Graphical information systems (GIS), Cognition, Models, User interfaces
The Principle of Locality Used for Hypertext Presentation: Navigation and Browsing in CONCORDE BIBAK 435-452
  M. Hofmann; H. Langendorfer; K. Laue; E. Lubben
This paper presents a comparison of various graph-layout algorithms used for an hypertext overview. The solution finally implemented in a prototype system called CONCORDE is discussed.
Keywords: Graph-layout algorithms, Hypermedia, Locality, Navigation, Smalltalk-80, User interface design