Stephen Cummaford
Ergonomics & HCI Unit, University College London, 26 Bedford Way, London, WC1H OAP
ABSTRACT
There is a need for more formal HCI design knowledge, such that effective design knowledge may be specified in a format which facilitates re-use. A conception of Engineering Design Principles (EDPs) is presented, as a framework within which to systematically relate design knowledge to performance. It is argued that the specification of these relations supports validation, leading to a higher likelihood that application of an EDP to an appropriate design problem will result in a satisfactory design solution. A hierarchy of classes of design problem is presented, and discussed in context of the ongoing research project.
Keywords Cognitive engineering, design, knowledge re-use, performance, principles, validation.
INTRODUCTION
Human-computer interaction practitioners have designed many effective technologies. However, the knowledge applied during development of a successful design solution is not often stated explicitly. Such ‘craft’ skills are difficult to represent explicitly, and as such, cannot easily be tested, and so validated [4]. Such under-specification limits the possibility of successful communication, and so re-use, of effective design knowledge. Engineering Design Principles (EDPs) benefit designers by facilitating more complete communication of designers’ knowledge. Design knowledge applied to produce a successful solution to a design problem can be carried forward to solve similar design problems. Validated design knowledge allows designers to generate new solutions with less prototyping and testing, thus reducing system development costs.
ENGINEERING DESIGN PRINCIPLES
The concept of specifying a complete, class-level, design solution was presented in the Engineering Conception of HCI (HCIe), which was developed within the Ergonomics & HCI Unit, UCL. [3]. The HCIe conception characterises the worksystem, tasks, and interaction structure in terms which support expression of performance, in terms of task quality achieved, and costs incurred whilst performing the work. The focus of my thesis is to develop coherent representations of the relationships between worksystem, tasks, interaction structure, and performance. Such representations support validation of the relationships between a design solution and performance. This conception of principles, as relating to specific classes of design problem, allows design knowledge to be represented at varying levels of generality, as classes may contain instances which are themselves classes.
EDP components
The EDP conception relates the elements of the HCIe conception, by specifying design representations which allow the relationship between the design solution and performance to be measured. An EDP consists of related representations which support coherent specification of a class-level effective design solution. Examples for each concept are provided, from a recent e-store transaction systems case-study, in { } brackets.
Scope: identifies the class of design problems for which this EDP has been validated {physical goods transaction systems supporting order collation and payment, for use by general public without specific training}.
Product goal: the work which is to be done {online order collation and payment }.
Domain model: contains objects which are transformed by the worksystem. Objects have attributes {book, has an owner} and the product goal is expressed in terms of object-attribute value transformations {change book owner from ‘vendor’ to ‘customer’ }.
Task-goal structure: specifies the interactive behaviours to be performed by the worksystem as a hierarchical box diagram, which decomposes the work to be done into behaviour primitives. These behaviours achieve the product goal.
User and computer models: specify the user and computer structures and behaviours which are sufficient to perform the task-goal structure {knowledge of credit card use, shipping address; realtime credit card transaction processing}. These models are used to determine costs incurred by the worksystem, an aspect of performance.
Achievable task quality: statement of how well the work was achieved by previous artefacts designed with this EDP, an aspect of performance { in empirical test, 100% of users completed transaction, no set-up or training required for users with 6+ months internet experience }. Cost matrix an expression of the costs incurred by the worksystem whilst performing the task-goal structure. It is constructed by listing the elements of the user model on the y axis {read, calculate, working memory} and the subtasks of the product goal on the x axis { order item, enter payment details}. The cost matrix is used to record the number of times each user model element is activated during each task. This is assessed empirically, by observation of user trials, and so measures actual user costs, rather than ideal performance. The cost matrix has proved useful in systematic comparison of competing design solutions [2].
Validation: The EDP conception supports expression of design knowledge by specifying an effective design solution at some level of generality. Performance is measured empirically, by testing with implemented design solutions specified by an EDP. Establishing performance supports assessment of whether the design solution is actually effective, rather the extent of its functionality [5]. The concepts contained in the EDP conception, it is argued [1], support validation by evaluation of the relations between performance and worksystem components performing a task-goal structure.
CLASSES OF DESIGN PROBLEM
A class of transaction system design problems has been identified to inform the development of EDPs. A transaction system is the order collation and payment component of an internet store. The parent class has three instances (subclasses), each of which is also a class. The subclasses of transaction system design problem are: physical goods (e.g. books); online services (e.g. video on demand); and financial (e.g. loans). A case study, addressing the design of physical goods transaction systems, has been conducted. This case study involved specification of a class-level design problem, by abstraction over the requirements for several e-stores, and specification of a design solution, which was then implemented and tested. There were particularly high costs associated with finding the total price for goods, including shipping, in the existing system. This was because the user must enter address and credit card details before the price was calculated. The re-designed system featured two popup menus to select a country and a shipping option, after which the total price could be displayed throughout the interaction, thus reducing user costs. Further details from this case-study are shown on the accompanying poster.
FUTURE RESEARCH
The next stage of the research project is to develop class-based knowledge for the subclass of online services transaction systems. The class hierarchy hypothesised earlier will then be evaluated. The hypothesised class hierarchy will be partially validated, if commonalities are found between the first and EDPs, such that design knowledge may be abstracted and represented at the general class level. The abstracted, class-level knowledge will then be validated by application to further design problems which are instances of the third subclass. The final stage of the project will be to address the usability of the design representations, such that the benefits of using the EDP framework are realised, whilst incurring an acceptable cost of use to designers.
ACKNOWLEDGEMENTS
I would like to thank my supervisor, Professor John Long, for his enthusiastic and insightful input throughout this project. This work was funded by the UK Engineering and Physical Research Council.
REFERENCES
[1] Cumrnaford and Long (1998) Towards a conception of HCI engineering design principles, in T.R.G. Green, L. Bannon, C.P. Warren & J. Buckley (eds.) Proceedings of ECCE-9, the Ninth European Conference on Cognitive Ergonomics. EACE, pp. 79-84.
[2] Cummaford, S. and Long, J. B. (1999) Costs Matrix: systematic comparison of competing design solutions, in S. Brewster, A. Cawsey & G. Cockton (eds.), Proceedings of Human-Computer Interaction INTERACT ’99 Volume 2, IOS Press, pp.25-26. [3] Dowell, J. and Long, J. B. (1989) Towards a conception for an engineering discipline of human factors. Ergonomics 32, 1513-1536 [4] Long, J. B. (1996) Specifying relations between research and the design of human-computer interactions. International Journal of Human Computer Studies, 44, 6, 875-920. [5] Newman, W. M. (1997) Better or Just Different? On the benefits of designing interactive systems in terms of critical parameters, in G. C. van der Veer, A. Henderson & S. Coles (eds.), Proceedings of the Symposium on Designing Interactive Systems (DIS ’97), ACM Press, pp.239-245.