Issue
Article
Vol.30 No.4, October 1998
Article
Issue
Issue |
Article |
Vol.30 No.4, October 1998 |
Article |
Issue |
The workshop was held at Queen Mary and Westfield College, University of London from 2-4 July 1997 and attracted participants from academia and industry in Europe and USA.
The motivation for the workshop arose from research being conducted under an EPSRC funded project, "Representations for supporting communication between users and designers" at QMW, University of London. The background to this research is the assumption that good communication underpins collaborative design. Since neither users nor designers have unlimited time to devote to collaborative design, effective representations of, for example, users' work or design solutions, are necessary. The aim of the project is to investigate different forms of representation for communication between users and designers with the goal of improving the efficiency and quality of interactive work systems. The workshop was organised by the authors in order to investigate how existing research might contribute to research on this project, and also to enable discussion of research issues arising from an international audience active in the field of investigating representations employed in interactive software development (ISD).
The workshop objectives were to gain a better understanding of the nature of representation and to acquaint delegates with current research conducted on representations and their use in ISD. We aimed to further the state of the art in representations in ISD research by identifying new research issues to be addressed, and by relating, comparing and contrasting current research on similar representations in different contexts, and dissimilar representations within the same context. Additionally, it was hoped that we could identify new uses for existing representations.
A total of ten papers were presented, following five themes: general representational issues; frameworks for representations; scenarios in interactive software development; semi-formal or formal representations, and representations for assessing usability.
In the opening address, Hilary Johnson was concerned with what a representation might be and the nature of representation. A representation it was argued was something that stands for something else, a description, a portrayal or copy of something else. Although, there are correspondences or likenesses between the representation and what it represents, a representation is usually on a smaller scale than, or is a simplified version of, the thing it represents. A representation is a systematic reduction of the thing to be represented for the purposes of manipulation. However, it is also possible for the representation to include an exaggerated element of the represented world.
Objects, properties and relationships are the constituents of the represented worlds; however, not all representations equally or adequately represent the constituents of interest. The utility of a representation can only be judged by how well it supports us reaching a conclusion about the thing being represented by looking only at the representation. In the context of interactive software development, we need to know the goal(s) of constructing the representation(s) in order to assess what objects, their properties and relationships need to be represented to enable us to make valid conclusions about aspects of those systems, their usage, and their development.
The first theme of the workshop on general representational issues was concerned with the need for effective modelling notations, developing criteria for choosing effective notations for ease of understanding, and considering representations not only as tools in a design process but also as the outputs. Jones, in her talk, argued that communication in interactive system design often occurs around a paper-based model rather than an implemented prototype and that it is necessary to select a notation which is easy for users to understand. However, many notations are immature or have an inadequate or unsound theoretical basis, thus highlighting the need to develop criteria for choosing effective notations as representations that users and designers can understand. A selection of criteria were emphasized in the talk ranging from the number of symbols in the notation, their discriminability and the structure inherent in the notation, etc. The conclusion to the talk was that an informed choice of modelling notation meant identifying criteria for effective notations relevant for the particular situation at hand.
Under the same general representational issues theme, Dourish made the point that although there are many different types of representation in interactive system design, all computer systems share one common form of representation -- a computational representation. The emphasis in this talk was with the form of representation used in interactive system development and the uses to which the representations are put. An interesting comment was that representations come to be used as though they were the things for which they stand, i.e. the representing world comes to be considered as the same as the represented world. There were two key elements to the research described in Dourish's talk; the first element relates to how the correspondence between representations and their referents are managed and maintained. The second element relates to the fact that the representations exist not only for the process of design but also for its products, i.e. the artefacts and other outcomes of design.
The second theme of the workshop was concerned with frameworks for representations in design. Maiden, presenting the Sutcliffe et al. research, outlined a framework for investigating design representations that account for users' information needs, the display media and artefacts. The framework was considered to have two functions: highlighting issues to be investigated to understand how representations function in a particular task context; and, as an evaluative approach for assessing the suitability of representations for a particular task. A framework with four levels of abstraction was described, taking requirements engineering as an illustrative example. A second framework developed by O'Neill, Johnson and Johnson addressed similar themes; relating forms of representation, types of communication and software development activities. This research was described as an initial attempt at a framework and was illustrated with two cooperative development projects as examples. Three fundamental questions were posed in the talk given by O'Neill: what is the nature of the representation used?; what are the purposes of using the representation, and who is using the representation? These issues were common across both development projects.
Scenarios in interactive software development, comprised the third workshop theme. The talks in this session were concerned with the concept of scenarios and breakdowns in the context of participatory design, and providing a framework to assist designers in describing scenarios.
Beynon-Davies gave the first talk, describing and relating two concepts in design; i.e. scenarios and breakdowns. The talk mainly focused on breakdowns but attempted through a study to show how scenarios and design breakdowns have a complementary status in design. The argument was that most participatory design examples in the literature occurred as academic exercises whereas this study occurred within a commercial setting. The talk concluded with reflections on the use of design breakdowns and scenarios.
In the second talk under the scenarios theme, Harrison, outlined a conceptual framework for scenario-based enquiry as a means of questioning the design of an interactive system. The aim of developing the framework is to provide a more systematic method of questioning and exploring interactive system designs based on focusing questions about a scenario. One thrust of the talk was to guide designers enquiry and to choose scenarios based on interactional problems and operational experience. The initial framework to assist designers in describing scenarios consisted of agents, states, goals, events and communication. The talk concluded with an example scenario from air traffic control to illustrate the use of the framework.
The fourth workshop theme was concerned with the use of formal and semi-formal representations. The first talk in this section by Markopoulos et al. described the development of a conceptual framework and a semi-formal representation scheme for the modular specification of user interfaces. The specification of user interface designs in the form of an Abstraction-Display-Controller (ADC) interactor is, according to Markopoulos, different to other formal or semi-formal representations as it is architectural, describing how the interface is structured as a composition of elementary activities. The talk outlined at some length the dimensions for the description of an ADC interactor, its architecture and dialogue. The talk concluded by arguing that no single representational modelling approach is adequate on its own but rather a range of representation techniques should be combined, each focusing on a different aspect of the design problem. This argument mirrors the comment by Sutcliffe et al., that it is necessary to look at how different representations interact.
The second paper in this section was by Paternò and Meniconi and in this talk Paternò outlined their use of "patterns" for representing dialogues at different abstraction levels. The talk highlighted the most important elements which should be included in a pattern definition. Design task patterns were then discussed and an instance of a task pattern was provided in order to exemplify how useful task patterns can be in specifying design solutions for communication and discussion between users and developers. The talk concluded with Paternò arguing that the patterns approach supports the possibility of creating user interface design environments which are more powerful than current toolkits. A full language of patterns for designing interactive applications is now being developed, along with a supporting toolkit.
The final section of the workshop was concerned with representations for usability evaluation. Blandford described an approach to usability evaluation which involved the use of an Ontological Sketch Model (OSM), where the analyst generates a structured but informal representation of the ontology of the system. The aim of this research is to develop a technique useful by non-specialists to evaluate system designs. The technique involves the use of an OSM notation which in the talk was applied to a prototype sketching program, J-Sketch, to assess whether the technique was both useful and usable. Important points arising from the study, in relation to representations and their content, relate to the fact that entities and actions were easily comprehended and described but that relationships between entities presented the evaluators with considerable problems. Further research is being conducted in this area, both to refine the method, and to consider the training needs for developing the representations.
Lavery and Cockton, provided the last talk in the workshop, making the important point that most representations have been focused on design specification, whilst the role of representation in usability evaluation has received much less attention. Furthermore, the representations generally in usability evaluation have tended to be models or representations of the evaluated system, rather than representations of the results of evaluating the interactive system, that is usability problem reports are themselves representations. Their aim is to develop a common representation for both predictions about usability and user test results. The talk concluded with a comment that questions about the reliability and quality of usability inspection methods and reports demonstrates that the nature of the representations are critical to producing effective and efficient usability evaluations. In the past usability evaluation has concentrated rather on the methods and procedures for evaluation rather than the representation of the evaluation results and predictions.
The workshop was organised in order to provide plenty of discussion time, both of individual papers, papers within sessions and research themes arising from those papers. In all about half the allotted time of the workshop was devoted to discussion. There were a number of common topics which were returned to throughout the workshop. The nature of representation was one of these topics, with both expressive power and functional affordances of representations, providing the basis for lively debate. In addition, the distinction between models and representations was also alluded to. Other discussions surrounded the concept of scenarios, as general structures to be extensively used in HCI as a vehicle for user-designer collaboration, or as specific instances. There was a general tendency to concentrate on representations that were or could be useful in the early stages of development. Issues for further research concerned the need to assess the utility of different representational formats within both academic and commercial settings.
There are two tangible outputs of the workshop for those who are interested in what occurred but were unable to come, the first is a proceedings which is available from workshop organisers at QMW. The second output is a special issue of the journal Human Computer Interaction on Representations for interactive software development which is due to be published in the near future.
Our thanks go to the people who participated in the workshop. We would also like to thank the authors for the high quality of their papers: Paul Beynon-Davies and Steve Holmes, University of Hertfordshire; Ann Blandford, Middlesex University and Thomas Green, University of Leeds; Carol Britton and Sara Jones, University of Hertfordshire; Paul Dourish, Apple Computer Inc; Michael Harrison, R.Fields and P. Wright, University of York; Darryn Lavery, University of Glasgow and Gilbert Cockton, University of Sunderland; Panos Markopoulos and Jon Rowson, QMW, University of London; Fabio Paternò and Silvia Meniconi CNUCE-C:N:R Pisa; Alistair Sutcliffe, N.Maiden and B.Bright, City University.
Thanks also to the following members of the workshop review committee:
John Carroll, Virginia Tech, USA
Gilbert Cockton, University of Sunderland, UK
Joëlle Coutaz, CLIPS-IMAG, Grenoble, France
David Gilmore, IDEO, USA
Michael Harrison, University of York, UK
Steve Payne, University of Cardiff, UK
Mike Smith, Harlequin Ltd., UK
Keith Stenning, University of Edinburgh, UK
We are grateful to EPSRC for funding the RESCUED project (EPSRC grant number GR/K79154).
Hilary Johnson is a Lecturer in Human Computer Interaction in the Computer Science Department at QMW, University of London. Her research interests include task analysis and modelling, principled approaches to interface design and evaluation.
Peter Johnson is a Professor of Human Computer Interaction and Head of the Computer Science Department at QMW, University of London. His research interests include: theory of HCI; tasks, context and situation in design; distributed working and mobile computing.
Eamonn O'Neill has just completed his PhD at QMW, University of London on "User-developer cooperation in software development: building common ground and usable systems". He was a Postdoctoral Researcher on the EPSRC funded project, "Representations to support communications between users and designers" before taking up a Lectureship in the Department of Computer Science at Queen Mary and Westfield College, University of London. His research interests include participatory design, cooperative task analysis and common ground in the design and use of interactive systems.
Hilary Johnson, Department of Computer Science, Queen Mary and Westfield College, University of London, Mile End Road, London, E1 4NS. UK.
Email: hilaryj@dcs.qmw.ac.uk
Peter Johnson, Department of Computer Science, Queen Mary and Westfield College, University of London, Mile End Road, London, E1 4NS. UK.
Email: pete@dcs.qmw.ac.uk
Eamonn O'Neill, Department of Computer Science, Queen Mary and Westfield College, University of London, Mile End Road, London, E1 4NS. UK.
Email: eamonn@dcs.qmw.ac.uk
Issue |
Article |
Vol.30 No.4, October 1998 |
Article |
Issue |