You are in: Home » History of the Unit
INTERACTION WITH TECHNOLOGICAL SYSTEMS
General Notes. This material has been scanned from the original typescript. While we have done our best to remove errors, some may well remain. You can access other parts of this particular Progress Report either from the menu at the bottom of this entry, by using your browser's back function, by navigating back to the Unit History Timeline, or by accessing the relevant section of the electronic archive. Reference for this report are indexed by number and these can be found in a dedicated section also accessible from the menu at the bottom of this entry.
Over the course of the last twelve years, through a combination of externally and internally funded effort, the Unit has been building upon its programme of research on a range of issues concerned with the use of computers. Through studying the patterns emerging in people's cognitive behaviour with interactive systems, this area promised benefits both for the development of basic theory and for its application. During the period covered by this particular progress report, progress has been consolidated and extended on issues concerned with expert systems, interactive dialogues, user modelling as well as the psychology of programming and notations in general. The theoretical benefits have started to come through strongly, and the problems of bridging between laboratory data, theories and the practicalities of systems design are now receiving greater research emphasis. This section will first review our empirical contributions, followed by a summary of the main developments in modelling, a review of achievements and proposals for future work.
Project 71 - Modelling Users of Interactive Devices
71.1 Knowledge Elicitation for Expert Systems (Gammack, Young)
Research on the application of psychological techniques for the elicitation of expert knowledge for incorporation into expert systems, has been done by Gammack, APU's first CASE student, (sponsored by Unilever pic). This PhD [538] focussed on the question of the structure of and relationships between the basic concepts in a technical domain, which is known to be an issue requiring attention early in the process of knowledge elicitation.
The work is based on the argument that existing psychological techniques, developed in the laboratory and intended for use with groups of subjects in toy domains, could be adapted and would prove useful in eliciting knowledge from single experts in domains of serious industrial interest. (The domains used include the selection of appropriate statistical tests for analysing market survey data, and the choice of library sources for seeking technical information.) Early results [453] emphasise the variety of techniques available and their suitability for different kinds of knowledge, and refer to well-established structuring techniques such as multi-dimensional scaling, cluster analysis, and repertory grid, along with a newer technique, "pathfinder analysis", which this work helped to pioneer. Later results go on to stress the convergence between these different techniques and the way they lend themselves to experimental verification [499, 454]. As a contribution to the art of knowledge elicitation, these techniques offer the advantages that: (1) They are quick to use; (2) They include procedures for their own validation; (3) They can be used with single or multiple experts; and (4) They yield graphical representations. They also provide part of the foundation for future tools to aid in knowledge elicitation, which depend upon the expert and the computer system sharing a body of core domain knowledge [486].
Related work [527, 480], has developed the argument that the results of knowledge elicitation should be recorded in one or more "intermediate representations", in contrast to the prevalent practice of going straight "from chat to rules", i.e. of coding information given by the expert directly into a machine-executable form. Among the advantages of using intermediate representations are that: (1) They serve as a record of the elicited knowledge independent of the technology used; (2) They can act as a medium of communication shared by expert and elicitor; (3) They provide a more complete record of the knowledge than can be captured in a machine-executable representation; and (4) They can be used for purposes other than building expert systems. The argument is gradually becoming accepted in the field.
71.2 The Acquisition and Use of Knowledge in Interactive Dialogues (Barnard, Maclean, Wilson)
The Unit's ten year collaboration with IBM drew to a close in 1987. While the early part of this research programme focussed upon the learning of simple interactive tasks by entirely novice users, the latter part concentrated on the learning of dialogues in more complex systems and upon the transition from novice to more experienced patterns of use. Issues associated with the representation and use of task and system knowledge were the major points of concern.
Through systematically studying the learning of an integrated office system with several techniques, it was observed that the accuracy of performance on different tasks in various system environments progressed with markedly different properties and at different rates. This evidence suggested that it was unhelpful to think of learning in this context as following a unified progression from declarative to proceduralised status. Rather, a repertoire of knowledge-based skills develop at different rates and stabilising at different levels. The very variability of performance in these tasks also has profound implications for the nature, interpretation and development of valid industrial "benchmark" tests of usability [519,543].
This same study also produced striking illustrations of the complex relationships which hold between "knowledge" and "performance". For example, given a choice of methods which take differing amounts of time, there are circumstances in which even experienced users will systematically choose the slower method of achieving their goal [552, 512]. Here the precise pattern of time data can be mathematically modelled as a linear function of the task parameters. However, the method selection data are best modelled by assuming that users manage the complex decision space by relying upon a "simple compensation schema" which biases choice [440]. Likewise, what people can verbalise about systems and tasks also bears a complex relationship to their performance. Verbal descriptions of tasks that accurately characterise how a system works develop in a different way from those that do not accurately characterise how it works. Accurate and inaccurate verbal descriptions also show very different patterns of correlations with performance time and error scores. These data [536, 403], and related data collected in a subsequent experimental study [482], indicate that the modelling and prediction of good performance on the one hand and errors on the other will require concepts that differentiate the roles and functions of ideal and non-ideal knowledge as well as the circumstances under which it is accessed and used. The data also contain a cautionary tale for those relying on single sources of data. Indeed, in such complex tasks understanding relationships among the many possible measures of knowledge and performance is a major concern for the development of theory and practical assessment techniques [522, 521].
71.3 Dialogue Form and Content (Barnard, Conway, MacLean, Wilson)
As interactive systems increase in their capabilities, more flexible dialogue structures and navigational opportunities are being incorporated into system designs. Experiments covered in the last progress report suggested that, although of value, certain forms of flexibility in simple dialogue tasks led to very real costs in initial learning and transfer to more complex tasks. Recent experimental work has extended this kind of result to more complex experimental settings. With tasks involving the same number of dialogue steps, learning and transfer costs can be greater with some forms of multi-tier menus, when compared to a more constrained hierarchical menu structure [402, 534]. The requirements of actually performing this kind of experiment in a realistic way led to the practical development of a user-interface management system specifically for behavioural research, the design and implications of which have also been reported [513, 418].
Another series of studies by Conway, Norris and Bowers concentrated upon the problems of using keyword systems [497]. One of the reasons for difficulty in the effective use of such systems is that there is little normative agreement between subjects as to which keywords are most likely to prove effective in accessing targeted information: this is even the case when the keyword system is made up of semantic categories and subjects are informed and given feedback about the structure of the system. It seems that subjects will only attempt one or two normative keyword responses and if these fail to access the target information, then subjects tend to generate personal idiosyncratic keywords which, of course, do not map on to the structure of any keyword system.
71.4 Mental Models in Performance and Design (Barnard, Duff, Young)
It has often been argued that learning complex tasks should be facilitated by the prior provision of an adequate mental model of the device. The empirical evidence on the point has remained ambiguous for some time. A series of experiments by Duff has systematically compared the effects of different forms of prior knowledge about task procedures and about the underlying device model. Relative to exploratory learning and presentation of procedural instructions, prior presentation of task relevant device knowledge can and does facilitate retention and subsequent problem-solving performance. Furthermore, it does so, with slightly different properties, where the task demands are for strict sequence learning or for free order selection of command operations [498].
Earlier theoretical work on mental models of interactive devices (Young 1981, 1983) has also been applied directly to illustrate its application to a practical design problem. The basic work suggested that a "task-action mapping" could perhaps provide a useful basis for the design of interactive systems. In order to explore this hypothesis, Young and Harris [528] undertook the exercise of designing a program for editing viewdata structures, based upon a task-action mapping analysis. The exercise proved successful, with the resulting editor exhibiting an unusual structure that provides a close meshing between the user's moment-to-moment goals and the actions offered by the system.
71.5 Formalisable Models of User Knowledge (T Green)
The user's knowledge of an interactive device was modelled by Payne and Green [423] in terms of a feature-grammar, TAG (this work was started while the authors were at SAPU). TAG represents a 'competence' model of the user of an artefact, that is, it emphasizes the user's knowledge of how-to-do-it rather than the user's behaviour or the user's conceptual model of the device. The novelty of TAG is that it captures the user's ability to generalise from partial knowledge, at least in the case of 'consistent' user interfaces where similar methods are used for achieving similar goals. It is trivially obvious that novices do not learn how to do bold face printing and then learn entirely separately how to do italics printing; they can generalise. Unlike prior research on the ability to generalise, TAG provides a model which predicts exactly what forms of consistency can be exploited by the learner meeting a user interface language. It was shown that TAG correctly predicted which of a number of artificial languages would be easiest to learn, even though that language was not the one whose context-free phrase-structure grammar contained the fewest rules [424]. The status of 'competence' theories of user knowledge against 'process' theories was explored, and a number of new metrics over grammars were suggested [458].
This research on TAG was partially supported by an ESPRIT contract to develop a 'Cognitive Simulator'. In this context, formal descriptions of life-size examples were developed; it is important that the usability of HCI models be assessed against real cases. TAG is now one of the few formal methods in HCI to have been applied to a number of life-size examples; moreover, by extracting the common features from these descriptions we have obtained a model of 'cross-applicational consistency', the degree to which knowing one applications program may help with learning another [461].
Green has developed a Prolog program which interprets a TAG description, so that given a statement of the user's goal the program will state how to achieve it in a given system. The original purpose was to help check the large grammars that are being developed, but it has subsequently emerged that an executable system also helps to predict some classes of errors, notably the 'description errors' in which users apply a correct method to an inappropriate object (e.g. trying to put a circle into italics when using a drawing program).
71.6 The Psychology of Programming (Gilmore, T Green)
Writing computer programs is a difficult real-life skill of increasing importance, bringing together many of the concerns of applied cognitive psychology. Theories of programming assume that individual fragments of written notation are conceptualised in larger units, which in the case of program notations have become known as 'programming plans'. These plans putatively express the conceptual role of each notational fragment. Strong versions of this theory assert that all programming expertise can be accounted for in plan terms; Green has criticised traditional computer science views in the light of cognitive theory and experiment, and explained why the so-called methodologies work when they work [502, 550]. Since evidence for the psychological reality of this theoretical construct was very limited, it was timely to conduct an empirical test. Gilmore and Green hypothesised that the comprehension of the plan component of programs would be increased if all parts of a single plan were highlighted together, but that comprehension of the control component would be unaffected by plan highlighting: and vice versa when the control structure was highlighted. This was indeed the case for one programming language (Pascal), showing that mental representations of program structure were notably multiple in nature, including at least plan and control structures. In the second language (Basic) the data suggested the control flow was the main form of mental representation, and that plan structures were a poor model of programmers' knowledge. [501, 410].
Adherents of 'programming plans' have also claimed that program development proceeds in units of one plan at a time. This claim was also put to the test. Expert professional programmers were closely studied solving problems in one of their speciality languages. Four languages were compared, and full details of keystrokes were recorded. Results showed that the local tactics depended on the language: Pascal programmers moved to and fro in the code making insertions here and there, significantly more so than users of other languages; Basic programmers in particular tended to run off long streams of code without jumping around. While plan structures appeared to play a role in the sequence of program development the behaviour was also apparently determined by details of the notational structure (notably its resistance to local change, termed by us 'viscosity') and the ease of parsing the notation back into plan structures (termed by us the 'role-expressiveness') [503, 504]. A theoretical analysis has been developed, the 'cognitive dimensions of notations', which will be described below.
Conway and Kahney have also explored programmers' behaviour, this time in a transfer of learning procedure where programmers attempted to write a recursive procedure in an AIM language [496]. The principle finding was that transfer of learning was facilitated only when information about the formal definition of recursion and a mapping relation were explicitly provided. They concluded that current models of learning could not account for these findings and that educational practice might benefit from the fact that formal definitions or illustrative examples alone do not facilitate learning.
71.7 Towards Approximate Models of Cognitive Activity (Barnard, A Green, Maclean, Wilson)
Theoretical advance in the domain of human-computer interaction must have the scope to encompass a broad range of system and task characteristics, as well as complex and often strategic behaviour on the part of the user. It must also furnish a means of communicating or delivering theoretical analyses into the processes of system design. One recent development has been to make empirically motivated theoretical principles explicit and deliver them through an expert system which constructs descriptive models of cognitive activity in complex tasks. The flexibility and power of AI technologies allow many of the problems associated with over-simple "design guidelines" to be circumvented [520]. Based upon Barnard's interacting cognitive subsystems framework [249], a new form of cognitive task analysis has been developed to generate an approximation over the operation of a distributed network of cognitive resources. This analysis holds that overt behaviour will be a complex function of process configurations, the procedural knowledge they embody, memory records accessed and the way in which the operation of distributed resources is co-ordinated and controlled. From this point of departure several of the earlier series of experiments were analysed to yield explicit functions inter-relating these factors [531,441].
These theoretical functions were then implemented in an expert system which incorporates the two additional kinds of rule required to (1) map input descriptions of systems, tasks and users onto the theoretical approximations; and (2) generate behavioural characterisations from those approximations. The initial knowledge bases [535, 404, 405] illustrated how a limited range of principles could work in simple tasks and how they could generalise from one form of dialogue or task (eg command names) to others (eg iconic dialogues). As well as extending basic principles, subsequent knowledge bases have explored how to build families of inter-related models covering a range of expertise levels and for different phases of extended tasks.
The use of an automated modelling system based upon expert system technology provides a direct means of delivering psychological knowledge into the design domain without requiring the applier to be an expert in the modelling technique, or even an expert interpreter of the subtleties of empirical evidence. However, the use of this modelling technique is not restricted to the delivery of an "applications" theory for human-computer interaction. As a cognitive modelling methodology, it lies in between classic verbally formulated theory and full blown AI simulations. In order for the modelling system to run, all assumptions and the logical relations asserted within a theory must be fully explicit and mapped to relevant task conditions. By being explicit, it avoids some of the problems associated with under-defining the formulation of theory in verbal terms. By representing knowledge in terms of approximations, it avoids the requirements in full blown AI simulations of specifying task relevant knowledge in its full extent. As such, the modelling technique is now at a stage where work in an applied domain can feed back to laboratory paradigms and outwards to other applied problems. Indeed, plans have been laid to develop and apply these techniques to the analysis of cognition and affect in laboratory and clinical settings (see proposals for Project 79).
71.8 Towards Programmable User Models (Young)
Another major strategy for representing and delivering psychological theory in the context of human-computer interaction has been pursued by Young and his colleagues. Since 1986 this has been done in the context of an Alvey project, in collaboration with Logica Cambridge Ltd and STC Technology Ltd, on the topic of "programmable user models" (PUMs). PUMs explore a novel approach to the problem of how to build predictive models of complex human behaviour which are accessible to professionals (in this case interface designers) who probably do not have a specialised training in psychology. A PUM is a constrained cognitive architecture that can be programmed (e.g. by an interface designer) to simulate an hypothetical user performing some range of tasks with a proposed interface. PUMs are intended as an effective way of conveying psychological considerations to the designer, by virtue of involving him/her in the actual process of building the model and making the predictive evaluations based upon it. The designer is thus in a position to see where the predictions come from and what factors they depend upon.
Following the effort spent clarifying notions of PUMs, much of the project was devoted to the investigation and assessment of different possible approaches to the integrated cognitive modelling of users. The favoured approach emerging from this work is inspired primarily by the SOAR architecture of Laird, Newell and Rosenbloom (of which more below). Later work on the project consisted of identifying the cognitive characteristics of the user most crucial for modelling human-computer interaction, and of preliminary investigations into ways of making SOAR reflect those characteristics. Although it will still be several years before there exists a running PUM, the industrial partners find that thinking about interface design problems in PUM's terms seems to help designers conceptualise the interface from the user's point of view. One interpretation of this (unexpectedly) early piece of technology transfer is that, while designers are frequently urged to "consider the interface from the user's viewpoint", such advice falls short of helping the designer know what actually to do. But to pose the question for a PUM "What 'program' must someone execute in order to use the interface?" provides the designer with a concrete and practical means for mentally stepping into the user's shoes.
Overview papers from the project serve, firstly, to explain its aims and methods [439], and secondly to propose a multi-dimensional classification of user models within which PUMs can be related to other approaches [516]. The technical papers from the project so far deal with particular aspects of human-computer interaction examined closely from the PUM viewpoint. Consider, for example, the situation facing a user sitting at a keyboard, who knows what he/she want to do, but not what steps to take to accomplish it. This kind of problem solving falls under the heading of "planning", and the project has examined the nature of planning in the HCI context [529, 517]. In contrast to the concerns of planning as studied in Artificial Intelligence, users produce just partial plans, they interleave planning with execution, and they plan opportunistically. A planning structure that would exhibit those characteristics has been proposed. Another example concerns the representation of procedural knowledge. In answer to the question "What does a user know, who knows how to do some task?", it is conventional to offer a procedure, i.e. a fragment of a program for performing that task. But such a purely procedural representation states both too little and too much. It states too little in the sense that it fails to explain how people do many things with their procedural knowledge other than just execute it. It states too much in that it severely overestimates the "knowledge content", i.e. the number of different things that have to be remembered in order to do the task. An account of procedural knowledge based upon multiple mutually-supporting representations resolves these dilemmas [530].
71.9 Towards a Theory of Notations (T Green, Simon, Young)
The instruction set for an interactive computer application (such as a spreadsheet or a word processor) can be viewed as an 'interaction language'. Research on user's knowledge of the interaction language and on the psychology of programming has led towards the beginnings of a more general theory of notations. It is known that in many creative situations, such as designing a drawing or a program, the preferred style of interaction resembles 'opportunistic planning'. This is a type of Al-derived planning, characterised by a control structure which allows the planner to develop any goal at any moment and to switch abruptly from developing a high-level goal to a low-level one and back, and allowing the actor to be driven by data and by circumstances as well as by lookahead [529]. It is possible to list, in abstract terms, many attributes of notations and to show how these relate to opportunistic planning as a form of generalised action theory. For instance, the actor needs the how-to-do-it knowledge described by TAG, but also needs to be able to parse the partly-created text into the appropriate mental representations. Certain information structures would increase the role-expressiveness, others would decrease it. Much of the work performed at SAPU has been accommodated in the same framework, as well as results from other workers. The framework has been applied successfully in industrial contexts [457].
Such a framework has been described by Green [551] as giving 'cognitive dimensions' of notations, in the sense of physicists' use of fundamental, orthogonal, dimensions. Some, including consistency and role-expressiveness, have been derived from the work described above; others emerged as soon as the framework was articulated. An example is the notion of 'viscosity', the work required to make a small change to a text structure. Still others, such as 'trail-following' and the problems of asymmetric pointers, derive from work performed while Green was at SAPU. Perhaps the most important aspect of this framework is that it enables a rational statement about the environment in which a notation is used. What is important is not that a notation has particular features, but that the notational structure and the surrounding environment act together in supporting user interaction. One task for the future is to enumerate some of the techniques by which environments can alleviate or complement the particularities of the notations.
71.10 Confronting the Problems of Scope and Applicability (Barnard, Simon, Whittington, Young)
The application of cognitive models to the phenomena of human-computer interaction has frequently been criticised because most models tend to be of strictly limited scope and also tend to lag well behind commercial developments in interface design. Although cumulative scientific experimentation is undeniably important, it tends to emphasise the accuracy of model predictions at the expense of their scope and applicability. One proposal to redress the balance is to encapsulate a systematically sampled range of behavioural phenomena in idealised scenarios. These are not unlike test sentences in linguistics. By "test driving" a range of models against such scenarios in a form of theoretical matrix, it is possible rapidly to extract common properties of different models, and to capitalise on just those properties most likely to lead to enhancements of scope and applicability [438, 547]. The three approaches outlined above (TAG, SOAR and Interacting cognitive subsystems) were considered along with other approaches in an early theoretical matrix.
Further Consolidation, Reviews and General Dissemination
In addition to developing new lines of empirical enquiry, and firming up theoretical approaches, older work has been consolidated, with significant effort directed at wider dissemination of research results.
Some further work [526] has been done extending earlier research into the human interface aspects of expert systems, the argument being that interacting with systems which are "quasi-intelligent" rather than merely complex poses novel issues. Some of these issues are akin to those that arise during interaction between people, such as the respective social roles of the two participants.
There have been further publications on the learning of command-argument structures [417], and the learning of extended task sequences [511]. In addition, all the work on command names has been collated and synthesised in a book chapter for the research community [445]; and a complete review of the whole command name field presented in a definitive handbook of human-computer interaction [533, 444]. Alternative approaches to task analysis, including those originating at APU, have also been reviewed and structured [462, 544], and more general overview papers have appeared (Shallice, 1982, [411]).
The potential application of a range of cognitive techniques provided input to an ESPRIT project concerning human factors tooling [495]. The work outlined earlier on TAG also contributed a number of internal reports to this project. The nature of evidence and theories has also been presented and discussed in book chapters directed at human factors [443, 532] and software engineers [442]. A chapter reviewing information-processing models in HCI was also prepared as a contribution to the definition of European research objectives arising out of the FAST programme [457].
In the case of AI, a tutorial introduction has been written [560], based on one of a series of videotaped lectures, explaining how "production systems", originally introduced into Cognitive Psychology and Artificial Intelligence as a technique for modelling human problem solving, have properties that enable them to be used as the core of expert system technology. Other previous work by Young consists of models of cognitive skill cast in the form of a "production system". Such models raise problems as to how they are to be evaluated appropriately, since data from a subject are used in deciding what model to assemble, and conventional techniques such as split-halves are not always applicable. A paper [515] based on work in a PhD thesis by Tony Priest (then a graduate student at the Open University, supervised by Young) proposes and compares a number of methods for performing this kind of evaluation.
The Unit has also been intimately involved with the development of European associations in the area of HCI. Green has been associated with a biennial series of European meetings on cognitive ergonomics since their inception in 1981, and has co-edited books based on the meetings [399, 401]; at the 1986 meeting it was agreed to form a formal constituted organisation, the European Association for Cognitive Ergonomics, and Green joined the founding committee. The first conference held by the new association took place in Cambridge in 1988. Green is also associated with the Psychology of Programming Interest Group, which is co-operating with its French counterpart in editing a book of invited contributions on the psychology of programming [400]. Similarly, Young is coordinating the European interest in SOAR. Members of the Unit have also delivered invited keynote addresses at European meetings on these topics [456].
This is also an appropriate place to note an Alvey contract held jointly between SAPU (Clegg) and APU (Green), aimed at presenting the techniques of both HCI and organisational psychology in a more accessible form. This project delivered a small book on "How to Evaluate your company's new technology" [398] intended for non-specialist readers.
71.11 Rank Xerox EuroPARC
In 1987 Rank Xerox opened its Cambridge EuroPARC, a laboratory for research in human-computer interaction with strong links to its famous parent laboratory Xerox PARC (Palo Alto Research Center). APU staff collaborated closely with Xerox over the setting-up of EuroPARC, and indeed the presence of APU is known to be one of the considerations influencing the choice of Cambridge as a site for the laboratory.
This close collaboration has continued, with a previous member of APU staff (A MacLean) moving to employment at EuroPARC. T R G Green was seconded there for two years, and Young is acting as a consultant, whilst others (P J Barnard, R Patterson) are involved in various kinds of informal research collaboration. EuroPARC are also the industrial sponsor for an MRC Partnership Award research studentship (E Churchill).
An early piece of collaboration between APU and EuroPARC covered in earlier section concerns the re-analysis and modelling of data from the APU/IBM project on users' preferences between two possible methods for accomplishing a task [440]. More recent collaborative work marks the beginning of what is intended to be a fairly long-term investigation and development of "Design Rationale" [419]. The observation behind this work is that when something is designed (say, the user interface to an interactive software system), typically the only thing preserved from the process of design is the designed object itself, in this case the interface. What is lost is the exploration of alternative options, and the web of argumentation and analysis that justifies the final design. Design rationale aims to capture those missing aspects. It is of considerable psychological relevance, as many of the arguments supporting design decisions that affect usability make appeal to considerations of the user, such as partial models of the user or the user's mental model of the system.
Green's framework on cognitive dimensions of notations, described earlier, requires considering a wide range of exemplars and environments. His secondment to EuroPARC enabled him to become conversant with the state of the art in visual programming, object-oriented programming, and program support environments. A short review of recent developments and some of the unanswered psychological questions they raise can be found in [400].
Project 72 - Cognitive Demands of Technical Communications
The value of studying technical communication, apart from any practical applications, is that people working with these materials recruit a wide range of cognitive processes in addition to those commonly thought of as "reading" and "writing". The project objective is a better understanding of the determinants of people's strategic communication choices. Readers choose what, when and how to read. Authors decide what information to provide, in which representational forms, with what sequences and links among items. One way of enhancing our understanding of these processes is to examine how information design influences people's communication choices. This is evident in the converging approaches adopted for the project which seek to determine on the one hand, the way cognitive processes are recruited when people design written materials, and on the other hand the way readers' performance is influenced by design decisions within, three disparate domains of technical information (electronic documents, instructions and questions, numerical information).
72.1 Creating Technical Communications (Hull, Lickorish, Wright)
Studies of writing processes enable comparisons between reading and writing strategies. They also enable sources of communication difficulty to be distinguished. Possible sources include lack of knowledge (e.g. about design options or readers' information needs) and inappropriate strategic choices (e.g. in the selection of representational forms). Our research suggests that both factors are implicated in the problems writers experience. We have found that adults' written instructions are frequently under-informative. The lack of information is apparently 'deliberate' because these writers can modulate their omissions as a function of the contextual support available to those following the instructions [432]. Such production problems might not be serious if people applied adequate revision skills to their texts. However, revision processes were found to be influenced by linguistic factors (e.g. verb form) which could differ from the factors (e.g. clause location) which influenced readers' comprehension [463]. Subsequent data suggested that revision deficiencies were not adequately remedied by advice given as guidelines.
We have examined design issues relating to information for helping people find their way inside modern complex buildings. Our experimental studies involved the use of both maps and wall signs in a hospital outpatient's department, and our data highlighted the importance of empirical procedures as part of the design process for information of this kind [436]. We discovered a discrepancy between the category of navigation information which enhanced users' satisfaction (maps) and that which yielded faster performance (wall signs) [433]. This discrepancy illustrates the risks inherent in too simple a notion of the reader's "task" when people interact with technical materials.
A range of techniques for text evaluation have been compared and their role in the design process has been discussed [464]. We have outlined the contribution of cognitive skills analysis to the improvement of a wide range of technical communications, and also the nature of the research effort needed in this area [466]. This framework has been contrasted with other approaches to technical writing, for example that represented by the Plain Language movements in Europe and North America. We have responded to numerous requests to relate the research literature on how readers read technical materials, to the needs of specific groups of writers - e.g. educators [426], technical writers in industry [471] and computer documentation [469, 472]. In illustrating that written information is important in many jobs, we emphasized that contemporary functional literacy must include the skills for designing the wide range of technical communications needed in the work place [427].
72.2 Electronic Documents (Lickorish, Stark, Wright)
To take advantage of the large storage capacities offered by media such as optical disks and CD ROMs, new information structures are being devised that exploit functionality not available with paper-based displays. Readers are offered alternative organisations of the same content (e.g. chronological/topical) and can move rapidly to different parts of the information space. Such documents are called hypertexts when the content is visual and predominantly static, and hypermedia when they involve several modalities and dynamic visual displays. These forms of communication raise many issues about readers' abilities to find and integrate information. Although these cognitive issues are not new, electronic documents are an important research tool because they offer an environment in which readers' strategies, particularly those relating to information seeking, can be monitored unobtrusively.
Contemporary reading theories do not explain why people choose not to read material which is available to them - e.g. clearly legible on the screen. We have outlined a model summarising the determinants of readers' choices about not reading and have pointed out that novel forms of discourse, such as hypertexts, challenge our understanding of many reading strategies [472]. In particular we have emphasized the relation between the content structure (e.g. hierarchical or modular) and the way readers will wish to move about within the text [478]. The authors of electronic documents have many more design options than do the authors of printed texts. We have articulated the constituents of this enlarged design space in order to be able to integrate potentially conflicting data from the many case-studies of hypertexts being reported in the literature [428]. Our experimental studies of readers moving through electronic texts have shown that alternative navigation systems differ in their cognitive demands, as indicated by readers' use of memory aids such as notetaking [523].
An involvement with "electronic journals" led to our examining the problems facing readers of lengthy electronic texts [467]. People had particular difficulty remembering where they had read something previously. Further studies showed that spatial cues were more effective retrieval aids than colour cues because of their sequencing properties [425, 434]. Our data from a small international survey showed that behavioural researchers make considerable use of electronic documents both at work and at home [474].
72.3 Instructions and Questions (Hull, Wilkins, Wright)
Written instructions allow direct exploration of comprehension processes, especially the mental representations that readers create when they plan and carry out actions on the basis of what they read. The important issues concern on the one hand the extent to which readers exploit the optionality they have in forming these representations so that remembering and manipulating the information can be done easily, and on the other hand the extent to which writers giving instructions understand their readers' information needs. Communicating the knowledge gained about how technical materials are read to those who are professional writers and information designers is one way of trying to meet writers' communication needs.
We have demonstrated differences in readers' representation of the information in negative conditionals (unless, if not) for contexts where the conditionals have a logical equivalence [430]. Subsequently we showed that the patterns of performance observed across a range of syntactic forms resulted from the way readers linked the negative element to other information in the representations that they created [431]. It appears that readers do not spontaneously exploit the optionality they have in choosing which representations to use, but are heavily influenced by the surface characteristics of the written information. This enhanced understanding of readers' representations of verbal instructions has been related to the design of written procedures [476]. We have also responded to requests to apply our earlier research on questions to the design of business forms and the design of forms for clinical trials in medical research [477].
Instructions for the operation of domestic electrical appliances can place subtle cognitive demands upon users; and users' failure to meet these demands can be dangerous. Our preliminary investigation of the task of wiring an electric plug have brought to light some of the ways in which design can guide effective usage [510].
72.4 Numerical Information (Hull, Lickerish, Wright)
Numerical information in the form of tables and graphs is a common component of technical materials and one that allows detailed exploration of readers' strategies for information-seeking. When numeric data can be accessed interactively, as in an electronic document, this permits examination of readers' metacognitive processes relating to information integration - i.e. their choices of representational forms when engaged in making comparisons and decisions on the basis of the information displayed. The research objectives are to ascertain how information design influences cognitive performance, and whether people are sufficiently aware of these design consequences to select appropriate display formats. Such awareness may be a precondition for successfully communicating quantitative numerical information.
We have identified ways of redesigning the numerical information presented in complex fraud trials, so as to make it more comprehensible to jurors [546, 435]. Current studies of people using simple graphs show that discrepancies can arise between the display formats yielding faster performance and those which users prefer. This suggests that in the multistage process of making comparisons within graphical data, different stages are influenced by different design factors.
Project 73 - Study and Modelling of Transport System Users
The behaviour of transport system users has been actively researched at the Unit since its inception. Over that time the Unit has made a unique contribution to this field and its output has strongly influenced the way in which this research is organised elsewhere.
An attempt has been made to produce a programme of research which spans a range of perceptual and cognitive issues, both theoretical and applied, involved in the driving task. These issues concentrate on the acquisition and organisation of skilled behaviour and decision making. Five areas of current research are described below, all of which are to some extent supported by external funds. In addition to this work, and his recent appointment as Professor of Traffic Science at Groningen, Brown continues to act in an advisory capacity to international, national and local government/public bodies concerned with transportation issues.
73.1 Individual Differences and Components of Driving Skill (Duncan, Brown, P Williams)
Using an instrumented car driven in normal traffic, Duncan, Williams and Brown have developed procedures for on-the-road assessment of multiple driving skills. The project, which was externally funded by a research grant to Duncan and Brown from the Transport and Road Research Laboratory (March 1986 - March 1988), aimed to supplement the limited information that can be obtained from global measures of driving proficiency, especially accident rate [422, 491]. Skills assessed include aspects of car control, scanning, anticipation and setting safety margins. Studies have dealt with the effects of experience, the nature of individual differences, and dual task interference.
Studies of experience [408] have compared normal, experienced motorists with both novices and acknowledged experts. In car control experienced motorists resemble experts, while novices are poorer. In scanning, anticipation and safety margins, however, it is the experienced drivers who perform poorly. The skills which apparently deteriorate with experience are those whose failure is rarely punished by immediate adverse consequences. Thus experience by no means guarantees driving expertise, suggesting there might be merit in periodic retraining or retesting. More theoretically, behavioural sequences learned under tuition may later be unstable if they are not obviously linked to general goals [12, 452], including accident avoidance.
Studies of individual differences [409] have dealt with the balance between specific skills and general characteristics that might make one person a "better" or "worse" driver than another. In experienced motorists the finding is that specific skills dominate. Reliabilities within an individual are high, but intercorrelations between skills are uniformly low. It seems that after practice the quality of skilled performance depends almost entirely on very specific learning. Thus individual differences do not reflect any finite set of "abilities". In practical terms, there seems little promise in selection tests given before learning this complex skill.
A dual task study produced results of particular theoretical significance [409]. The research assessed changes in performance across component skills induced by a secondary task chosen to be demanding but to share little obvious content with driving. Performance decrements were modest and, across component skills, the profile of decrements agreed well with a profile of (similarly modest) correlations with "general intelligence" or g. The findings are linked elsewhere in the report to Duncan's wider work on executive functions. It is well known that practice reduces both dual task interference and correlations with g. These results suggest that both changes reflect the same diminishing role of general executive functions as specific skills develop. In support of this view, we have found that head injury patients with clear signs of executive dysfunction show little impairment in driving skills.
73.2 Estimating Time-to-Coincidence/Collision (Groeger, Brown, P Williams/Grande)
Deciding how long it will take to reach a distant object towards which one is travelling is an important component of several driving tasks. Early applied work by Brown [494] showed, in the context of indirect vision devices, that restricting the visual field impairs estimates of when a target object will be reached, usually resulting in 'undershooting'. This research was followed up by Groeger and Brown [506, 414] who showed that subjects' judgements of where they were in relation to an object could be "shifted" by presenting veridical, or non-veridical, auditory speed information and visual information simultaneously.
73.3 Risk Assessment and Decision Making (Groeger, Brown, Chapman)
A relatively high proportion (ca 25%) of road accidents are attributable to mismatches between road system demands and drivers' responses [537]. In a subsequent series of papers [487, 449, 548, 396, 413, 397, 415, 412, 460] Brown and
Groeger have attributed such mismatches to inadequacies in drivers' perception of hazard in the traffic scene and their assessment of their own ability to deal safely with that hazard. However, most research on "subjective risk" has confounded these two variables, thus preventing its findings being translated into effective accident countermeasures [449].
73.4 The Role of Feedback in Developing Driving Skill (Groeger, Brown, Grande)
The process of learning to drive and the utility of driver training programmes have been discussed in recent papers by Brown and Groeger [446, 451, 452], and inadequacies in the training and long term support of novice drivers have been adduced as contributory factors in the accidents of inexperienced drivers. The central role of knowledge of results in honing performance of skills other than driving has long been accepted, and its role in the driving task should be investigated.
Generally, skills differ in the source (i.e. task-intrinsic/task extrinsic) and predictability of feedback. Not all goals and events when driving alone, for instance, are accompanied by feedback on the success of performance, whereas levels of feedback are obviously high when a person is learning to drive, with the driving instructor being the primary source. When the learner driver later comes to drive without supervision, the extent of extrinsic feedback is greatly diminished, and inability to rely on task-intrinsic feedback may lead to deterioration of skill and contribute to an increased likelihood of accident [452].
73.5 Errors in Performance (Groeger, Brown)
Groeger and Brown have highlighted the interpretation of errors as a prime contributory factor in the understanding of skill organisation and development [507, 397, 415, 412]. The lack of reliable information on driver error in the behavioural antecedents of road accidents is a major failing of routine police reports (STATS 19). It is ethically inadmissible to provoke most of the more interesting and informative errors, for experimental purposes, in road and traffic studies. Yet 'safe' laboratory studies often fail to capture, or simulate validly, the essential characteristics of complex driving tasks.
REFERENCES
AI - Authored Books
394. BROWN, I.D. Driver Behaviour: The. Human Factor in Road Transport. Amsterdam: Elsevier. Manuscript in preparation for 1990 publication.
A2 - Edited Books
395. BROWN, I.D., Goldsmith, R, Coombes, K.E. and Sinclair, M. (Eds.) (1985) Ergonomics International 85 (Proceedings of the 9th Congress of the International Ergonomics Association, Bournemouth, 2-6 September 1985). London: Taylor and Francis Ltd.
396. BROWN, I.D. and GROEGER, J. (Eds.) (1988) Risk perception and decision taking during the transition between novice and experienced driver status. Ergonomics, (Special Issue), 31, 585-597.
397. BROWN, I.D. and Janssen, W. (Eds.) (1988) Risky decision making in transport operation. Ergonomics, (Special Issue), 31, 403-668.
398. Clegg, G., Warr, P., GREEN, T., Monk, A., Kemp, N., Allison, G. and Lansdale, M. (Eds.) (1988) People and Computers: How to Evaluate your Company's New Technology. Chichester: Ellis Horwood Ltd.
399. Falzon, P., GREEN, T.R.G., Hoc, J.M., Streitz, N. and Van der Veer, G.C. (Eds.) (1986) Proceedings of the Third European Conference on Cognitive Ergonomics, Paris, France: INRIA.
400. Hoc, J-M., GREEN, T.R.G., GILMORE, D.J. and Samurcay, R. (Eds.) The Psychology of Programming. London: Academic Press, in press.
401. Van der Veer, G., GREEN, T.R.G., Hoc, J.M. and Murray, D (Eds.) (1988) Working with Computers: Theory Versus Outcome. London: Academic Press. London: Academic Press Ltd.
B - Refereed Journal Articles
402. BARNARD, P., MACLEAN, A. and WILSON, M. (1988) Navigating integrated facilities: Initiating and terminating interaction sequences. In CHI '88. Human Factors in Computing Systems. New York: ACM, pp.121-126.
403. BARNARD, P.J., WILSON, M. and MACLEAN, A. (1986) The elicitation of system knowledge by picture probes. In CHI'8 6 Human Factors in Computer Systems. New York: ACM, pp.235-240. (See also APU/1914)
404. BARNARD, P.J., WILSON, M. and MACLEAN, A. (1987) Approximate modelling of cognitive activity: Towards an expert system design aid. In CHI & GI'87, Human Factors in Computing Systems and Graphics Interface. New York: ACM, pp.21-26.
405. BARNARD, P., WILSON, M. and MACLEAN, A. (1988) Approximate modelling of cognitive activity with an expert system: A theory-based strategy for developing an interactive design tool. The Computer Journal, 31, 445-456.
406. BROWN, I.D. (1987) Burgeoning issues in ergonomics, 1975-80. Ergonomics, 30, 13-18.
407. BROWN, I.D. Drivers' margins of safety considered as a focus for research on error. Ergonomics (Special Issue), in press.
408. DUNCAN, J., WILLIAMS, P. and BROWN, I.D. Components of driving skill: Experience does not mean expertise. Ergonomics, in press.
409. DUNCAN, J., WILLIAMS, P., NIMMO-SMITH, M.I. and BROWN, I.D. The control of skilled behaviour: Learning, intelligence, and distraction. In D. Meyer and S. Kornblum (Eds.), Attention and Performance XIV. Hillsdale, N.J.: Lawrence Erlbaum Associates, in press.
410. GILMORE, D.J. and GREEN, T.R.G. (1988) Programming plans and programming expertise. Quarterly Journal of Experimental Psychology, 40A, 423-442.
411. GREEN, T.R.G. (1986) Cognitive aspects of HCI. Computer Bulletin, 2, 7-9.
412. GROEGER, J.A. (1989) Conceptual bases of drivers' errors. Irish Journal of Psychology, 10, 276-290.
413. GROEGER, J.A. and BROWN, I.D. (1989) Assessing one's own and others' driving ability: Influences of sex, age and experience. Accident Analysis and Prevention, 21, 155-168.
414. GROEGER, J.A. and BROWN, I.D. Cognitive components of veridical time-to-coincidence estimation. (Manuscript submitted to Perception).
415. GROEGER, J.A. and CHAPMAN, P. Errors and bias in assessments of danger and frequency of traffic situations. Ergonomics (Special Issue), in press.
416. GRUDIN, J. and BARNARD, P.J. (1985) When does an abbreviation become a word? And related questions. In CHI'85, Human Factors in Computer Systems, New York: ACM, pp.121-125.
417. Hammond, N.V., BARNARD, P.J., Morton, J., Long, J.B. and Clark, LA. (1987) Characterizing user performance in command-driven dialogue. Behaviour and Information Technology, 6, 159-205. (See also APU/1739-84)
418. MACLEAN, A. (1987) Human factors and the design of user interface management systems: EASIE as a case study. Information and Software Technology, 29, 192-201.
419. MACLEAN, A., YOUNG, R.M. and Moran, T.P. (1989) Design rationale: The argument behind the artifact. In CHI'89, Human Factors in Computing Systems, pp.247-252. New York: ACM Press, pp.247-252.
420. MCKENNA, F.P. (1985) Do safety measures really work? An examination of risk homeostasis theory. Ergonomics, 28, 489-498.
421. MCKENNA, F.P. (1985) Evidence and assumptions relevant to risk homeostasis. Ergonomics, 28, 1539-1541.
422. MCKENNA, F.P., DUNCAN, J. and BROWN, I.D. (1986) Cognitive abilities and safety on the road: A re-examination of individual differences in dichotic listening and search for embedded figures. Ergonomics, 29, 649-663.
423. Payne, S.J. and GREEN, T.R.G. (1986) Task-action grammars: A model of the mental representation of task languages. Hum an-Computer Interaction, 2, 93-133.
424. Payne, S.J. and GREEN, T.R.G. (1989) The structure of command languages: An experiment on task-action grammar. International Journal of Man-Machine. Studies, 30, 213-234.
425. Tombaugh, J., LICKORISH, A. and WRIGHT, P. (1987) Multi-window displays for readers of lengthy texts. International Journal of Man-Machine Studies, 26, 597-615.
426. WRIGHT, P. (1987) Writing technical information. In E.Z. Rothkopf (Ed.), Review of Research in Education, XIV. Washington, DC: American Educational Research Association, pp.327-385.
427. WRIGHT, P. (1988) Functional literacy: Reading and writing at work. Ergonomics, 31, 265-290.
428. WRIGHT, P. (1989) Alternative interfaces for hypertexts. Hypermedia, 1, 146-166.
429. WRIGHT, P. Varieties of strategic reading: Some interface requirements. Machine Mediated Learning (Special Issue), Proceedings of AERA Symposium, March 1989, in press.
430. WRIGHT, P. and HULL, A.J. (1986) Answering questions about negative conditionals. Journal of Memory and Language, 25, 691-709.
431. WRIGHT, P. and HULL, A. (1988) Reading to do: Creating contingent action plans. British Journal of Psychology, 79, 187-211.
432. WRIGHT, P. and HULL, A.J. How people give verbal instructions. Applied Cognitive Psychology, in press.
433. WRIGHT, P., HULL, A.J. and LICKORISH, A. Navigating in a hospital outpatients' department: The relative merits of maps and wall signs. (Manuscript submitted).
434. WRIGHT, P. and LICKORISH, A. (1988) Colour cues as location aids in lengthy texts on screen and paper. Behaviour and Information Technology, 7, 11-30.
435. WRIGHT, P., LICKORISH, A. and HULL, A. (1989) Numerical evidence in commercial fraud trials. Information Design Journal, 5, 171-181.
436. WRIGHT, P., LICKORISH, A. and HULL, A.J. The importance of iterative procedures in the design of location maps for the built environment. Information Design Journal, in press.
437. Wright, R. and LOGIE, R.H. (1988) How young house burglars choose targets. The Howard Journal of Criminal Justice, 27, 92-104.
438. YOUNG, R.M. and BARNARD, P. (1987) The use of scenarios in human- computer interaction research: Turbocharging the tortoise of cumulative science. In J.M. Carroll and P. Tanner (Eds.), CHI & GI'87, Human Factors in Computing Systems and Graphics Interface. New York: ACM, pp.291-296.
439. YOUNG, R.M., GREEN, T.R.G. and SIMON, T. (1989) Programmable user models for predictive evaluation of interface designs. In CHI'89, Human Factors in Computing Systems. New York: ACM Press, pp.15-19.
440. YOUNG, R.M. and MACLEAN, A. (1988) Choosing between methods: Analysing the user's decision space in terms of schemes and linear models. In E. Soloway, D. Frye and S.B. Sheppard (Eds.), CHI'88, Human Factors in Computing Systems. New York: ACM, pp.139-143.
C - Invited Chapters and Commentaries
441. BARNARD, P.J. (1987) Cognitive resources and the learning of human-computer dialogs. In J.M. Carroll (Ed.), Interfacing Thought: Cognitive Aspects of Human-Computer Interaction. Cambridge, MA: MIT Press, Ch.6, pp.112-158. (See also APU/1912)
442. BARNARD, P.J. Applied Cognitive Psychology: A research methodology for Human-Computer Interaction. In Downton, A.C. (Ed.), Engineering the Human-Computer Interface. London: McGraw Hill, in press.
443. BARNARD, P.J. The contributions of applied cognitive psychology to the study of human-computer interaction. In B. Shackel and S. Richardson (Eds.), Human Factors for Informatics Usability. Cambridge: Cambridge University Press, in press.
444. BARNARD, P. and GRUDIN, J. (1988) Command names. In M. Helander (Ed.), Handbook of Human-Computer Interaction. Elsevier Science Publishers, B.V. (North-Holland), pp.237-255. (See also N.I. 2201)
445. BARNARD, P., GRUDIN, J. and MACLEAN, A. (1989) Developing a science base for the naming of computer commands. In J.B. Long and A. Whitefield (Eds.), Cognitive Ergonomics and Human-Computer Interaction. Cambridge: Cambridge University Press, pp.95-133.
446. Biehl, B. and BROWN, I.D. (1987) A comparison of driver training in the Federal Republic of Germany and Great Britain. In Vergleich der Verkehrssicherheit in der Bundesrepublik Deutschland und Grossbritannien, Forschungsprojekt 8507 der Bundesanstalt fur Strassenwesen, 1987, Chapter 5.5, pp.259-279.
447. BROWN, I.D. (1985) Fatigue. In P.A.B. Raffle (Ed.), Medical Aspects of Fitness to Drive. London: Medical Commission on Accident Prevention, 4th Edition, Chapter 10, pp.79-88 (refs: 109-112).
448. BROWN, I.D. (1986) Prospects for improving road safety. (Ergonomics Society Lecture 1985) presented at the University of Nottingham, 29 March 1985). Ergonomics, 29, 1495-1505.
449. BROWN, I.D. (1987) Predisposing factors in the alcohol- and drug- impairment of young drivers' performance. In T. Benjamin (Ed.), Young Drivers Impaired by Alcohol and Other Drugs. London: Royal Society of Medicine, pp.165-183.
450. BROWN, I.D. Accident reporting and analysis. In J.R. Wilson and E.N. Corlett (Eds.), Evaluation of Human Work: A Practical Ergonomics Methodology. London: Taylor and Francis Ltd., in press.
451. BROWN, I.D. and Biehl, B. (1987) A comparison of traffic education in the Federal Republic of Germany and Great Britain. In Vergleich der Verkehrssicherheit in der Bundesrepublik Deutschland und Grossbritannien, Forschungsprojekt 8507 der Bundesanstalt fur Strassenwesen, 1987, Chapter 5.6, pp.291-314.
452. BROWN, I.D., GROEGER, J.A. and Biehl, B. (1987) Is driver training contributing enough towards road safety? In J.A. Rothengatter and R.A. de Bruin (Eds.), Road Users and Traffic Safety. Assen/Maastricht, The Netherlands: Van Gorcum, pp.135-156.
453. GAMMACK, J.G. (1987) Different techniques and different aspects of declarative knowledge. In A. Kidd (Ed.), Knowledge Acquisition for Expert Systems: A Practical Handbook. New York: Plenum Press, pp.137-163.
454. GAMMACK, J.G. (1987) Formalising implicit domain structure. In C.J. Pavel in and M.D. Wilson (Eds.), Knowledge Acquisition for Engineering Applications. Rutherford Appleton Laboratory Report RAL-87-055, pp.29-38. (This paper is a revised/adapted version of APU/2072.)
455. GREEN, T.R.G. (1989) Cognitive dimensions of notations. In A. Sutcliffe and L. Macaulay (Eds.), People and Computers V. Cambridge: Cambridge University Press, pp.443-460.
456. GREEN, T.R.G. Limited theories as a framework for human-computer interaction. In D. Ackermann and M. Tauber (Eds.), Mental Models and Human-Computer Interaction 1. Elsevier Science Publishers B.V. (North-Holland), in press.
457. GREEN, T.R.G. User modelling: The information-processing perspective. In J. Rasmussen and H.B. Andersen (Eds.), Research Directions in Cognitive Science: A European Perspective, Vol. 3: Human Computer Interaction. London: Lawrence Erlbaum Associates, in press.
458. GREEN, T.R.G., SCHIELE, F. and Payne, S.J. (1988) Formalisable models of user knowledge in human-computer interaction. In G. Van der Veer, T.R.G. Green, J.M. Hoc and D. Murray (Eds.), Working with Computers: Theory Versus Outcome. London: Academic Press, pp.3-46.
459. GROEGER, J.A. (1986) Developing driver behaviour: A computational approach. In S. Piemont and P. Stenner (Eds.), Proceedings of CEC Workshop on Education and Training to Prevent Breakdowns in Adaptation. Trieste, Instituto Tecnico Nautico, pp.39-48.
460. GROEGER, J.A. Concepts of danger: The unknown risks we run. In H. Bohm (Ed.), Psychological Statistics and Models of Accidents in Traffic Systems. Bremen: Commission of the European Communities, in press.
461. SCHIELE, F. and GREEN, T.R.G. HCI Formalisms and cognitive psychology: The case of Task-Action Grammar. In M. Harrison and H. Thimbleby (Eds.), Formal Methods in Human-Computer Interaction. Cambridge: Cambridge University Press, in press.
462. WILSON, M., BARNARD, P., GREEN, T.R.G. and MACLEAN, A. (1988) Knowledge-based task analysis for human-computer systems. In G. Van der Veer, T.R.G. Green, J.M. Hoc and D. Murray (Eds.), Working with Computers: Theory Versus Outcome. London: Academic Press, pp.47-87.
463. WRIGHT, P. (1985) Editorial policies and processes. In T.M. Duffy and R.H. Waller (Eds.), Designing Usable Texts. New York: Academic Press Ltd., Ch.4, pp.63-96.
464. WRIGHT, P. (1985) Is evaluation a myth? Assessing text assessment procedures. In D. Jonassen (Ed.), The Technology of Text, Vol. 2. Englewood Cliffs, N.J.: Educational Technology Publication, pp.418-435.
465. WRIGHT, P. (1985) Prerequisites of writing for computer users. Keynote address at the Second Conference on Writing for the Computer Industry. Plymouth, New Hampshire: Plymouth State College, pp.l-21.
466. WRIGHT, P. (1986) Phenomena, function and design: Does information make a difference: In D.J. Oborne (Ed.), Contemporary Ergonomics 1986 (Proceedings of the Ergonomics Society 1986 Annual Conference, Durham, England 8-11 April 1986). London: Taylor and Francis, pp.l-18.
467. WRIGHT, P. (1987) Reading and writing for electronic journals. In B.K. Britten and S.M. Glynn (Eds.), Executive Control Processes in Reading. Hillsdale, N.J.: Lawrence Erlbaum Associates, pp.23-55.
468. WRIGHT, P. (1987) Shifting sands and shipwrecks in forms design: What sort of maps do we need? In C. Jansen and M. Steehouder (Eds.), Formulieren als Communicatiemiddel. Amsterdam: LINEA, pp.65-75.
469. WRIGHT, P. (1988) Issues of content and presentation in document design. In M. Helander (Ed.), Handbook of Human-Computer Interaction. Elsevier Science Publishers B.V. (North-Holland), pp.629-652.
470. WRIGHT, P. (1988) The need for theories of NOT reading: Some psychological aspects of the human-computer interface. In H. Bouma (Ed.), Working Models of Human Perception. London: Academic Press Ltd., pp.319-340.
471. WRIGHT, P. (1989) Can research assist technical communication? Keynote Address in Proceedings of the 36th International Technical Communication Conference. Chicago: Society for Technical Communication, pp.RT3-RT6.
472. WRIGHT, P. (1988) Communicating with the user. In N. Heaton and M. Sinclair (Eds.), Designing End-user Interfaces State of the Art Report 15.8. Maidenhead: Pergamon Infotech Ltd., pp.123-129.
473. WRIGHT, P. Designing and evaluating documentation for I.T. users. In B. Shackel and S. Richardson (Eds.), Human Factors for Informatics Usability. Cambridge: Cambridge University Press, in press.
474. WRIGHT, P. Homework: An international comparison of behavioural researchers' use of computers for work at home. In M. Feeney and K. Merry (Eds.), Information Technology and the Research Process. London: Bowker-Saur, in press.
475. WRIGHT, P. Hypertexts as an interface for learners: Some human factors issues. In D. Jonassen and H. Mandl (Eds.), Designing Hypermedia for Learning. Berlin, Heidelberg: Springer-Verlag, in press.
476. WRIGHT, P. Reading technical texts: Strategies for procedural instructions. In J. Ulijn and L. Olson (Eds.), The Proceedings of the 2nd Eindhoven Symposium on Language for Special Purposes. University of Michigan Press, in press.
477. WRIGHT, P. and Haybittle, J. (1986) Designing clinical trials forms to collect the right data. In H. Glenny and P. Nelmes (Eds.), Handbook of Clinical Drug Research. Oxford: Blackwell Scientific Publications Ltd., pp.247-270.
478. WRIGHT, P. and LICKORISH, A. (1989) The influence of discourse structure on display and navigation in hypertexts. In N. Williams and P. Holt (Eds.), Computers and Writing. Ablex, Ch.7, pp.88-120.
479. YOUNG, R.M. (1987) Production systems for modelling human cognition. In E. Scanlon and T. O'Shea (Eds.), Educational Computing. Chichester: John Wiley and Sons in association with The Open University, pp.209-220.
480. YOUNG, R.M. (1988) Role of intermediate representations in knowledge elicitation. In D.S. Moralee (Ed.), Research and Development in Expert Systems IV. Cambridge: Cambridge University Press, pp.287-288.
D - Conference Proceedings
481. BARNARD, P.J. and Harrison, M. (1989) Integrating cognitive and system models in Human-Computer Interaction. In A. Sutcliffe and L. Macaulay (Eds.), People and Computers V. Cambridge: Cambridge University Press, pp.87-103.
482. BARNARD, P.J., ELLIS, J. and MACLEAN, A. (1989) Relating ideal and non-ideal knowledge to performance. In A. Sutcliffe and L. Macaulay (Eds.), People and Computers V. Cambridge: Cambridge University Press, pp.461-473.
483. BELLAMY, R.K.E. (1988) Agents and actions: A framework for describing the interaction between users and intelligent systems. In Proceedings of Alvey Workshop on Multiple Agent Systems.
484. BELLAMY, R.K.E. and GILMORE, D.J. Programming plans: Internal or external structures. In K.J. Gilhooly, M.T.G. Keane, R.H. Logie and G. Erdos (Eds.), Lines of Thinking: Reflections on the Psychology of Thought, Vol. 2. Chichester: John Wiley & Sons Limited, in press.
485. BELLAMY, R.K.E. and GREEN, T.R.G. (1986) "Damn it, I've done it again": An investigation of action slips. In P. Falzon, T.R.G. Green, J.M. Hoc, N. Streitz and G.C. Van der Veer (Eds.), Proceedings of the 3rd European Conference on Cognitive Ergonomics. Paris: INRIA, pp.3-9.
486. Bradshaw, J. and YOUNG, R.M. Shared causal knowledge as a basis for communication between expert and knowledge acquisition system. In J. Boose, B. Gaines and M. Linster (Eds.), Proceedings of the 2nd European Workshop on Knowledge Acquisition. Bonn, in press.
487. BROWN, I.D. (1985) Concepts and definitions in road safety. In Proceedings of "Evaluation 85", International meeting on the evaluation of local traffic safety measures, Paris, 20-23 May 1985, Vol.11, pp.413-422.
488. BROWN, I.D. (1985) How have ergonomists responded to the changing demands of new technology? In Proceedings of 147th Annual Meeting of the British Association for the Advancement of Science, University of Strathclyde, 26-30 August.
489. BROWN, I.D. (1985) Realising ergonomics matters in industrial health and safety. In Proceedings of Healthy Industry - Its Realisation", 8-9 July 1985, University of Warwick.
490. BROWN, I.D. (1988) Drivers' perception of motion. In A.G. Gale, M.H. Freeman, CM. Haslegrave, P. Smith and S.P. Taylor (Eds.), Vision in Vehicles II. Amsterdam: Elsevier Science Publishers, B.V. (North-Holland), pp.25-26.
491. BROWN, I.D. Functional requirements of driving. In Cars and Casualties, Proceedings of the 8th Berzelius Symposium, Stockholm, 17-18 March 1986. Stockholm: The Swedish Society of Medicine, in press.
492. BROWN, I.D. On the social dilemma of motorway safety. In Proceedings of 'International Symposium on Driver Behaviour in a Social Context'. Paris, France, 16-18 May 1989, in press.
493. BROWN, I.D. and GROEGER, J.A. (Eds.) Errors in the operation of transport systems. Ergonomics (Special Issue), 33, London: Taylor & Francis, in press.
494. BROWN, I.D. and McFaddon, S.M. (1986) Display parameters for driver control of vehicles using indirect viewing. In A.G. Gale et al. (Eds.), Vision in Vehicles. Elsevier Science Publishers, B.V. (North-Holland), pp.265-274.
495. Byerley, P., BARNARD, P., Carr, D., Foster, A., Fowler, T., Saffin, R. and Ward, G. (1987) Cognitive simulator for user-interface design. In ESPRIT 7: Achievements and Impact, Part 2, Amsterdam: North-Holland, pp.1101-1109.
496. CONWAY, M.A. and Kahney, H. (1987) Transfer of learning in inference problems: Learning to program recursive functions. In J. Hallam and C. Mellish (Eds.), Advances in Artificial Intelligence. Chichester: John Wiley and Sons, pp.239-250.
497. CONWAY, M.A., NORRIS, D. and BOWERS, J. (1985) Cognitive factors in information systems. In Proceedings of the Eleventh International Symposium on Human Factors in Telecommunications, (meeting at CCETT, France, 9-13 Sept.).
498. DUFF, S.C. (1989) Reduction of action uncertainty in process control systems: The role of device knowledge. In E.D. Megaw (Ed.), Contemporary Ergonomics 1989 - Proceedings of the Ergonomics Society's 1989 Annual Conference, Reading, England. London: Taylor and Francis, pp.213-219.
499. GAMMACK, J.G. (1987) Modelling expert knowledge using cognitively compatible structures. In The Proceedings of the Third International Expert Systems Conference, London 2-4th June 1987, Oxford: Learned Information (Europe) Limited, pp.191-200.
500. GAMMACK, J.G. and YOUNG, R.M. (1985) Psychological techniques for eliciting expert knowledge. In M.A. Bramer (Ed.), Research and Development in Expert Systems. Cambridge: Cambridge University Press, pp.105-112.
501. GILMORE, D.J. and GREEN, T.R.G. (1987) Are 'Programming plans': psychologically real - outside Pascal? In H.-J. Bullinger and B. Shackel
(Eds.), Human-Computer Interaction - INTERACT '87. Elsevier Science Publishers, B.V. (North-Holland), pp.497-503.
502. GREEN, T.R.G. (1986) Computer languages: Everything you always wanted to know but no-one can tell you. In F. Klix and H. Wandke (Eds.), Proceedings of the 1st Macinter Conference on Man-Computer Interaction (East Berlin, 1984). Elsevier Science Pub., B.V. (North-Holland), pp.249-259.
503. GREEN, T.R.G., BELLAMY, R.K.E. and Parker, J.M. (1987) Parsing and gnisrap: A model of device use. In H.-J. Bullinger and B. Shackel (Eds.), Human-Computer Interaction - INTERACT 7. Elsevier Science Publishers, B.V. (North-Holland), pp.65-70.
504. GREEN, T.R.G., BELLAMY, R.K.E. and Parker, J.M. (1988) Parsing and gnisrap: A model of device use. In G.M. Olson and E. Soloway (Eds.), Empirical Studies of Programmers, Vol. 2. Ablex.
505. GROEGER, J.A. (1988) Underlying structures: Driver models and model drivers. In J.A. Rottengatter and R.A. de Bruin (Eds.), Road User Behaviour: Theory and Research. Assen/Maastricht, The Netherlands: Van Gorcum, pp.518-526.
506. GROEGER, J.A. and BROWN, I.D. (1988) Motion perception is not direct with indirect viewing systems. In A.G. Gale M.H. Freeman, CM. Haslegrave, P. Smith and S.P. Taylor (Eds.), Vision in Vehicles II. Amsterdam: Elsevier Science Publishers B.V. (North-Holland), pp.27-34.
507. GROEGER, J.A. and BROWN, I.D. (1988) Mistakes and misunderstandings: Interpreting drivers' errors. In S. Oppe (Ed.), Traffic Safety Theory and Research Methods, Session. 3, Theoretical Analysis and Models. Amsterdam: SWOV, pp.1-23.
508. GROEGER, J.A. and Cavallo, V. Judgements of time-to-collision and time- to-coincidence. In A.G. Gale et al (Eds.), Vision in Vehicles III, in press.
509. GROEGER, J.A., GRANDE, G. and BROWN, I.D. Accuracy and safety: Effects of different training procedures on a time-to-coincidence task. In A.G. Gale et al (Eds.), Vision in Vehicles III, in press.
510. HULL, A., WILKINS, A.J. and BADDELEY, A.D. (1988) Cognitive psychology and the wiring of plugs. In M.M. Gruneberg, P. Morris and R.N. Sykes (Eds.), Practical Aspects of Memory: Current Research and Issues, Vol. 1: Memory in Everyday Life. Chichester: John Wiley, pp.514-518.
511. Jorgensen, A. and BARNARD, P.J. (1986) An experiment on the effect of task structure in interactive computer systems. In Proceedings of Work with Display Units, Stockholm, May 12-15, pp.777-780.
512. MACLEAN, A., BARNARD, P. and WILSON, M. (1985) Evaluating the human interface of a data entry system: User choice and performance measures yield different trade-off functions. In P. Johnson and S. Cook (Eds.), Proceedings of HCI 5, People and Computers: Designing the Interface, (Sept. 1985). Cambridge: Cambridge University Press, pp.172-185.
513. MACLEAN, A., BARNARD, P. and WILSON, M. (1986) Rapid prototyping of dialogue for human factors research: The EASIE approach. In M.D. Harrison and A.F. Monk (Eds.), People and Computers: Designing for Usability. Cambridge: Cambridge University Press, pp.180-195.
514. MCKENNA, F.P. (1985) Prepared discussion: Risk homeostasis in an experimental context. In L. Evans and R.C. Schwing (Eds.), Human Behavior and Traffic Safety. New York: Plenum Press, pp.143-144.
515. Priest, T. and YOUNG, R.M. (1988) Methods for evaluating micro-theory systems. In J. Self (Ed.), Artificial Intelligence and Human Learning. Intelligent Computer-Aided Instruction. London: Chapman and Hall Computing, pp. 124-137.
516. SIMON, T. (1988) Analysing the scope of cognitive models in Human- Computer Interaction: A trade-off approach. In D.M. Jones and R. Winder (Eds.), People and Computers IV. Cambridge: Cambridge University Press, pp.79-93.
517. SIMON, T. and YOUNG, R.M. (1988) GOMS meets STRIPS: The integration of planning with skilled procedure execution in Human-Computer Interaction. In D.M. Jones and R. Winder (Eds.), People and Computers IV. Cambridge: Cambridge University Press, pp.581-594.
518. STARK, H.A. (1989) What do readers do to pop-ups, and pop-ups do to readers? In: Proceedings of Hypertext 2 Conference, York, July 1989.
519. WILSON, M.D, BARNARD, P.J., and MACLEAN, A. (1985) Analysing the learning of command sequences in a menu system. In P. Johnson and S. Cook (Eds.), People and Computers: Designing the Interface. Cambridge: Cambridge University Press, pp.63-75.
520. WILSON, M., BARNARD, P. and MACLEAN, A. (1986) Using an expert
system to convey HCI information. In M.D. Harrison and A.F. Monk (Eds.), People and Computers: Designing for Usability. Cambridge: Cambridge University Press, pp.482-497.
521. WILSON, M., BARNARD, P. and MACLEAN, A. An investigation of the learning of a computer system. In P. Falzon (Eds), Cognitive Ergonomics, Understanding Learning and Designing Human-Computer Interaction. London: Academic Press Ltd., in press.
522. WILSON, M. and MACLEAN, A. (1986) Assessing cognitive aspects of learning and using computer systems. In P. Falzon, T.R.G. Green, J.M. Hoc, N. Streitz and G.C. Van der Veer (Eds.), Proceedings of the Third European Conference on Cognitive Ergonomics. Paris, France: INRIA, pp.197-207.
523. WRIGHT, P. and LICKORISH, A. (1989) An empirical comparison of two navigation systems for hypertexts. In Proceedings of Hypertext 2 Conference, York, July 1989, Session P.3.
524. WRIGHT, P., LICKORISH, A. and HULL, A.J. (1988) The importance of iterative procedures in the design of location maps for the built environment. In Proceedings of Design Research Society Conference on Information for Designers.
525. WRIGHT, P. and Norrish, P. (1988) Colour Cues in Technical Texts. Proceedings of BCS Displays Group Seminar on "Colour and Computer Graphics", December.
526. YOUNG, R.M. (1987) Interacting with quasi-intelligent machines. In Proceedings of the BPS Conference on Information Technology and People: Designing for the Future (January 1987). Leicester: British Psychological Society, pp.56-58.
527. YOUNG, R.M. and GAMMACK, J.G. (1987) Role of converging techniques and intermediate representations in knowledge elicitation. In T.R. Addis (Ed.), Proceedings of First European Workshop on Knowledge Acquisition for Knowledge-Based Systems. Reading, University, Computer Science Department, Section D - Techniques D.7.
528. YOUNG, R.M. and HARRIS, J.E, (1986) A viewdata-structure editor designed around a task/action mapping. In M.D. Harrison and A.F. Monk (Eds.), People and Computers: Designing for Usability. Cambridge: Cambridge University Press, pp.435-446.
529. YOUNG, R.M. and SIMON, T. (1987) Planning in the context of human-computer interaction. In D. Diaper and R. Winder (Eds.), People and Computers III: Proceedings of the Conference on Human-Computer Interaction. Cambridge: Cambridge University Press, pp.363-370.
530. YOUNG, R.M., SIMON, T., Lewis, A.E. and Tang, H. (1989) Multiple mutually-supporting representations for procedural knowledge. In A.G. Cohn (Ed.), Proceedings of the Seventh Conference of the Society for the Study of Artificial Intelligence and Simulation of Behaviour. London: Pitman Publishing, pp.21-30.
E - Technical Reports, Theses and Tests
531. BARNARD, P.J. (1986) Cognitive resources and the learning of human-computer dialogues. Hursley Human Factors Laboratory Report No. HF118, March 1986.
532. BARNARD, P. (1987) The contributions of applied cognitive psychology to the study of human-computer interaction. Hursley Human Factors Laboratory Report No. HF 133, September 1987. (See also N.I. 2205)
533. BARNARD, P. and GRUDIN, J. (1988) Command names. MCC Technical Report Number ACA-HI-039-88, February 1988. (See also N.I. 2202)
534. BARNARD, P., MACLEAN, A. and WILSON, M. (1987) Navigating integrated facilities: Initiating and terminating interaction sequences. Hursley Human Factors Laboratory Report No. HF 134, September 1987. (See also N.I. 2203)
535. BARNARD, P., WILSON, M. and MACLEAN, A. (1986) Approximate modelling of cognitive activity with an expert system: A concept demonstrator for an interactive design tool. IBM Hursley Human Factors Report HF123, Sept. 1986, pp. 1-37.
536. BARNARD, P., WILSON, M. and MACLEAN, A. (1986) The elicitation of system knowledge by picture probes. Hursley Human Factors Laboratory Report No. HF117, February 1986.
537. BROWN, I.D. (1985) Biggs, R., Rainbird, R. and Quimby, A. Perception of danger. Report of Joint Working Party on Safety Research. London: Institution of Highways and Transportation/Medical Commission on Accident Prevention.
538. GAMMACK, J.G. (1987) Eliciting expert conceptual structure using converging techniques. Unpublished PhD Thesis, University of Cambridge.
539. GREEN, T.R.G. (1987) Process modelling with PML: An assessment using Cognitive Ergonomics. Unpublished report commissioned by STC Technology Ltd.
540. GREEN, T.R.G. (1987) Teleshopping: Perspectives from HCI and cognitive psychology. Unpublished report commissioned by General Information Systems Ltd., Croxton, Cambs.
541. Probert, D. and YOUNG, R.M. (1985) Report on the joint MMI/IKBS workshop on applications driven research. SERC/Alvey Directorate.
542. Shackel, B., Florentin, J. and WRIGHT, P. (Eds.) (1986) BLENDS: the Computer Human Factors Journal. LIR Report 47. ISBN 0 7123 3073 9.
543. WILSON, M.D., BARNARD, P.J. and MACLEAN, A. (1985) A user learning of core command sequences in a menu system. IBM Hursley Human Factors Laboratory Report No. HF114, September, 117 pages.
544. WILSON, M.D., BARNARD, P.J. and MACLEAN, A. (1986) Task analyses in Human-Computer Interaction. IBM Hursley Human Factors Report HF122, August 1986, pp.1-63. (See also APU/1955)
545. WILSON, M., BARNARD, P. and MACLEAN, A. (1987) Learning to use a computer system: Its assessment via multiple behavioural methods. Hursley Human Factors Laboratory Report No. HFR 135, September 1987. (See also N.I. 2097)
546. WRIGHT, P., LICKORISH, A. and HULL, A. (1986) Presenting numerical information to fraud trial juries. In Improving the Presentation of Information to Juries in Fraud Trials, Report to Lord Roskill Committee on Fraud Trials. London: H.M.S.O., pp.17-39.
547. YOUNG, R.M., BARNARD, P., SIMON, T. and WITITTINGTON, J. (1989) How would your favourite user model cope with these scenarios? S1GCHI Bulletin, 20, 51-55.
F - Dissemination
548. BROWN, I.D. (1986) The motorist's attitude towards and perception of road traffic offences. In Proceedings of Seminar on Traffic Law and the Motorist, 25 March, 1986, London: London Centre for Transport Planning, pp.8-18.
549. BROWN, I.D. (1988) Motorway design and usage as causes of behavioural problems among drivers. Invited paper presented to Motorway Safety Seminar, Preston, Lancashire, 30-31 January 1988.
550. GREEN, T.R.G. (1986) Design and use of programming languages. In J.K. Skwirzynski (Ed.), Software System Design Methods, NATO ASI Series Vol.F22. Berlin, Heidelberg: Springer-Verlag, pp.213-241.
551. GREEN, T.R.G. The nature of programming. In J-M. Hoc, T.R.G. Green, R. Samurqay and D.J. Gilmore (Eds.), The Psychology of Programming. London: Academic Press, pp.21-44.
552. MACLEAN, A., BARNARD, P. and WILSON, M. (1985) A comparison
between performance and preference trade-offs in a data entry task. In Proceedings of the Eleventh International Symposium on Human Factors in Telecommunications, (meeting at CCETT, France, 9-13 Sept. 1985).
553. POULTON, E.C. (1987) Ergonomics. In R.L. Gregory (Ed.), Oxford Companion to the Mind. Oxford: Oxford University Press, pp.226-228.
554. WRIGHT, P. (1985) Form and function. Review of J.F. Burgess 'Human Factors in Forms Design'. Contemporary Psychology, 30, 75-76.
555. WRIGHT, P. (1985) Review of: M. Nystrand (Ed.), "What writers know: The language, process and structure of written discourse", New York: Academic Press, 1982. Applied Psycholinguistics, 6, 354-356.
556. WRIGHT, P. (1988) Beyond plain English. RaPAL (Research and Practice in Adult Literacy) Bulletin, 5, 15-17.
557. WRIGHT, P. (1988) Cricket graph 1.1. The Psychologist, 1, p.113.
558. WRIGHT, P. (1988) Filemaker plus. The Psychologist, 1, 34-35.
559. WRIGHT, P., LICKORISH, A. and Whalley, P. (1985) Experimental comparison of reading lengthy texts on either CRT screen or paper. In Forum 85 sponsored by the International Council for Technical Communication, Helsingor, Denmark, pp.105-112.
560. YOUNG, R.M. (1987) An introduction to production systems. In T. O'Shea, J.S. Self and G. Thomas (Eds.), Intelligent Knowledge-Based Systems: An Introduction. London: Harper and Row, pp.68-82.
Other sections in the 1985-1989 report
3. LANGUAGE, SPEECH, READING AND WRITING