My work adopts a temporal perspective on the usability of software applications used in distributed and multi-tasking computing environments, such as Internet browsers used for searching the Web. It is motivated by the need to understand how users go about what Woods calls "disturbance management", [Woods, 88; Sarter and Woods, 91]: how we behave in response to delays in computer processes and, to a lesser extent, what we do in response to interruptions by such processes. Both delay and interruption constitute inappropriate timing and cause usability problems as a result of the failure of users' immediate expectation or anticipation.
The paper suggests that findings in the psychology of time should
be applied to an analysis of users' temporal requirements for usable
computer applications. A number of features of human temporal
experience and time perception which affect interactive behaviour are
outlined and their importance to usability is discussed. Finally, a
number of practical suggestions for Web browser design are made,
derived indirectly from psychological theories.
We can propose general definitions of what it is for the timing of interface behaviour to seem "just right", but we cannot meaningfully establish what it is for the timing to be too fast or too slow. This is the conclusion to be drawn from Shneiderman's treatment of response time studies: there are too many user, task and environment variables at play, to determine any generally acceptable interface response rate, [Shneiderman 87].
Here are two other versions of the "just right" definition:
i) the timing of behaviour at the Web interface always conforms to users' expectations, formed on the basis of prior experience of use of a browser, or on the basis of current status information.
This definition embraces the view that delays may be made acceptable if the user is accustomed to them. It supports the practice of introducing response-time choke. It is consistent with the view that delays can be made tolerable to users, if the interface affords the user an understanding of progress in computer activity, whilst he or she is waiting, [Conn, 95]. With the use of progress indicators and the notion of temporal affordances, we can design and build interfaces which partially satisfy this second definition of temporal usability.
A second usability criterion, closely related to that above:
ii) the experienced user never has to devote conscious attention-directed awareness to the timing of interface behaviour.
Delays (and interruptions) exist, but the user is never directly confronted by them and left to puzzle about them. In the case of delay, an "appropriate" alternative task is proffered by the interface together with support for the user to resume the suspended task at an "appropriate" time. This idea is not new. It was advocated more than a decade ago, appearing in work by Cypher [86] and by Miyata and Norman [86]. However, such task management techniques have barely been developed in present day Web browsers and commercial window managers and experienced users are left to develop such management strategies for themselves. For example, experienced Web users may develop a strategy in which they run multiple browser sessions, to counteract the boredom and frustration of delay. Should the Web interface actively encourage a less expert user to pursue such a strategy and what is the cognitive load associated with managing multiple sessions in parallel?
The central point here, and the focus of the paper, is that we
should be able to design and build interfaces which better meet these
two definitions of temporal usability, if we understand how and when
users perceive temporal information in the (interactive) environment,
if we understand how users formulate their expectations and
explanations for the timing of (interactive) events and if we
understand the costs involved in task switching behaviours, [see
Allport et al 94]. However, with a few exceptions, the literature in
HCI is marked by a general absence of consideration for this
subjective human experience of time, [for exceptions see Dix 90,
Decortis et al 91; O'Donnell and Draper 95].
The literature in the psychology of time and the study of decision making should be a fruitful source of findings and models which we might examine and apply to user timing in HCI. Such an attempt is not straightforward, as the literature in time psychology is marked by an absence of widely accepted theory on the fundamental mechanisms responsible for time perception, temporal judgments and the dynamics of voluntary task selection. There are no quantitative laws, analogous to Fitts' law, that are known to govern the timing of human cognitive responses, but there are some general frameworks for analysing human temporal interpretations of the environment.
The paper reports on Michon's model of human timing as a number of flexible, adaptive tuning processes, from which the properties of temporal perception and judgment emerge, [Michon 85 and 93]. Jones and Boltz have added to this analysis, suggesting that, at a cognitive level, human "tuning" is aided by coherent event structures in the environment, [Jones and Boltz 89; Boltz 92]. We ask what it means for Web browser behaviour to be temporally coherent? Is a temporally coherent event structure just that which corresponds to a well-rehearsed script or a frequently activated schema? Is there a role for predictably rhythmic delivery of information at the Web interface, in the management of delay. Can temporally regulated or paced delivery of information counteract the frustration of delay?
We examine temporal concepts and metaphors. Do users think
spatially about time? What explanations are sought for delays? Pilot
studies indicate that users' "explanations" for delay are shallow.
Amongst experienced Web users, there appears to be a resignation to
temporally indeterminate behaviour. What is the significance to Web
usage of the reported interconnection of "delay" and "risk" and what
effect does this have on inhibiting/ promoting users task shifting in
the face of delay? Users may be unwilling to abandon a delayed search
once a certain amount of time has already been devoted to
waiting.
The paper concludes by questioning where we can go, on the basis of known psychological findings, in interface design for temporal disturbance management and temporal usability.
Can the behaviour of Web browser interfaces be made to promote models of the system which support appropriate user explanations for delay? What role is there for the promotion of physical/ spatial reasoning in the system image presented at the interface? Can we adapt the idea of "spatial" navigational aids in Web browsers, to support better temporal navigation during information retrieval tasks? [Bieber and Wan 94] Is it sufficient to give Web users a deeper technical understanding of delay, or does usability arise out of giving Web users something better to do whilst they wait?
Should temporal coherence be built in to interface behaviour at the Web browser level or at a higher level of user task management? How do we interpret and make practical use of conflicting psychological factors in the promotion of task and application switching, when users are faced with delays?
To date, an understanding of temporal "disturbance management" has
only been pursued in the fields of interface design for process
control and safety critical systems, [Decortis et al 91; Sarter and
Woods 91]. With the increased use of Web based resources for
personal, academic and commercial purposes, there is a good case for
investigating human temporal disturbance management in the retrieval
of distributed information. Psychology can inform the design of
interfaces which promote effective disturbance management, but only
indirectly. Much more experimental evidence is required, obtained
specifically from observing the controlled use of Internet tools, to
construct models of human computer usage and usability under
conditions of both interruption and delay.
Allport et al 94: Allport A, Styles E A, Hsieh S L, "Shifting intentional set - exploring the dynamic control of tasks", Attention and Performance 15, pp 421-452, 1994
Bieber and Wan 94: Bieber M and Wan J, "Backtracking in a multiple window hypertext environment", Proceedings of ECHT 94 (5th ACM Conference on Hypermedia Technology), 1994
Boltz 92: Boltz M, "The incidental learning and remembering of event durations" in "Time, Action and Cognition: Towards Bridging the Gap", eds F Macar, V Pouthas and W J Friedman, Kluwer Academic Publishers (NATO ASI series), pp 153-163, 1992
Conn 95: Conn A P, "Time affordances: the time factor in diagnostic usability heuristics", Proceedings of CHI 95 (ACM), 186-193, 1995
Cypher 86: Cypher A, "The Structure of Users' Activities", in Norman D A and Draper S W eds, "User Centred System Design", Lawrence Erlbaum Associates, pp 243-263, 1986
Decortis et al 91: Decortis F, de Keyser V, Cacciabue P C, Volta G, "The Temporal Dimension of Man Machine Interaction," in "Human Computer Interaction and Complex Systems", Weir G R S and Alty J L eds, Academic Press, pp 51-72, 1991
Dix 90: Dix A J, "Non-determinism as a Paradigm for Understanding the User Interface", in "Formal Methods in Human Computer Interaction", Harrison M D and Thimbleby H W (eds), CUP, 1990
Jones and Boltz 89: Jones M R and Boltz M, "Dynamic attending and responses to time", Psychological Review 96 (3), pp 459-491, 1989
Michon 85: Michon J A, "The compleat time experiencer", in Michon J A and Jackson J L (eds), "Time, Mind and Behaviour", Springer Verlag, pp 20-52, 1985
Michon 93: "Concerning the Time Sense: The Seven Pillars of Time Psychology", Psychologica Belgica 33(2), 329-345, 1993
Miyata and Norman 86: Miyata Y and Norman D A, "Psychological Issues in Support of Multiple Activities", in "User Centred System Design - New Perspectives on Human-Computer Interaction", D A Norman and S W Draper eds, Lawrence Erlbaum Associates, pp 265-284, 1986
O'Donnell and Draper 95: O'Donnell P and Draper S W, "How machine delays change user strategies", in Papers from a Workshop on Temporal Aspects of Usability, C Johnson and P Gray (eds), University of Glasgow GIST Technical Report G95-1, 1995
Sarter and Woods 91: Sarter N B and Woods D D, "Situation awareness: a critical but ill-defined phenomenon", International Journal of Aviation Psychology 1 (1), pp 45-57, 1991
Shneiderman 87: Shneiderman B, "Response time and display rate", in "Designing the user interface: strategies for effective human-computer interaction", Addison-Wesley, 1987
Woods 88: Woods D D, "Coping with complexity: the psychology of human behaviour in complex systems", in "Tasks, errors and mental models", L P Goodstein, H B Andersen and S E Olsen eds, Taylor and Francis, pp 128-148, 1988