HCI Bibliography Home | HCI Journals | About BIT | Journal Info | BIT Journal Volumes | Detailed Records | RefWorks | EndNote | Hide Abstracts
BIT Tables of Contents: 0102030405060708091011121314151617

Behaviour and Information Technology 7

Editors:Tom Stewart
Publisher:Taylor & Francis, Ltd.
Standard No:ISSN 0144-929X
Links:Table of Contents
  1. BIT 1988 Volume 7 Issue 1
  2. BIT 1988 Volume 7 Issue 2
  3. BIT 1988 Volume 7 Issue 3
  4. BIT 1988 Volume 7 Issue 4

BIT 1988 Volume 7 Issue 1

Editorial BIB 1-2
  Tom Stewart
Legibility of Video Display Units During Off-Angle Viewing BIBA 3-9
  Stephen J. Morrissey; Rose W. Chu
This research examined how off-angle or oblique viewing of a VDU screen and the physical location of a message on the screen influenced message legibility. Eight trained subjects viewed five-character long common words, number strings, and alphanumeric messages presented at 15 different combinations of oblique viewing angle and location of message on the VDU screen for 2.5 seconds. It was found that common words and number strings showed little overall loss in legibility except when oblique viewing angle exceeded ±32°. Alphanumeric messages were found to have a significantly lower legibility than the common words and number strings. It was recommended that for best overall legibility of the three types of messages studied here, that oblique viewing angles be less than ±32°. Predictor equations were also developed to aid in predicting loss of accuracy based on the type of message and physical features of the viewing task.
Colour Cues as Location Aids in Lengthy Texts on Screen and Paper BIBA 11-30
  P. Wright; A. Lickorish
Readers of documents on CRT displays report difficulties in remembering whereabouts in a lengthy text they previously read something. Four experiments explore whether subdividing such texts, at appropriate thematic boundaries, into five successive coloured sections can aid readers' retrieval of information. Experiment 1, using texts presented on coloured paper, showed that this use of colour helped readers relocate information. Experiment 2 presented the same texts on a CRT, but variation in the colour of the characters on the screen did not help readers relocate information. Experiment 3 replicated the findings of experiment 2, with texts differing in both content and structure from those used previously. Experiment 4, again using coloured text on a CRT display, showed that giving readers a visible guide to the ordering of the coloured sections was not sufficient to restore the advantage that coloured pages had for texts presented on paper. The implications of these findings for variation in the background and foreground colouring of multi-window displays are discussed, but the main conclusion concerns the caution needed when transferring information design solutions across media.
Selected Graph Design Variables in Four Interpretation Tasks: A Microcomputer-Based Pilot Study BIBA 31-49
  John G. Casali; Kenneth B. Gaylin
A pilot study was undertaken to assess the efficacy of selected graph types and information coding schemes in producing quick and accurate graph interpretation. Point-plots, line graphs, bar graphs, and three-dimensional bar graphs were constructed and displayed using an IBM-PC microcomputer with colour monitor. The effectiveness of each of these graph types, as well as the coding scheme used within each (colour versus monochrome), was investigated using 32 subjects in four separate experiments involving either point-reading, point-comparison, trend-reading, or trend-comparison tasks. Dependent measures including task completion time, graph interpretation error, subjective mental workload rating, and graph preference rating were obtained. In three of the tasks, colour-coded graphs were associated with lower task completion time, lower rated mental workload, and higher rated preference than monochrome graphs. However, there were no differences in subjects' error scores between colour and monochrome coding for any of the tasks. For the point-reading task, the three-dimensional bar graphs were less effective than the other graph types, though there were no differences among graph types for the other three tasks. Implications for graph design given specific data interpretation tasks are discussed, based on the results of the four experiments and related literature.
The Effect of Tree Structure on Search in a Hierarchical Menu Selection System BIBA 51-65
  Kent L. Norman; John P. Chin
Search processes in a hierarchical menu selection system were investigated in a study that varied the structure of the tree. A hierarchical data base was composed of 256 gift items grouped into different clusters and presented using menus. Depth of the tree was held constant while breadth varied with level. Five structures were explored with the following number of alternatives at each of four levels: constant (4 x 4 x 4 x 4), decreasing (8 x 8 x 2 x 2), increasing (2 x 2 x 8 x 8), concave (8 x 2 x 2 x 8), and convex (2 x 8 x 8 x 2). Subjects searched for either specifically named gifts (explicit targets) or gifts appropriate for a scenario situation (scenario targets). In general, explicit targets took less time to find and fewer frames to traverse than for scenario targets. For explicit targets, the increasing menu was slightly superior to the rest. Search time was about the same across the five tree structures for explicit targets, but differed greatly for scenario targets. The concave and increasing structures were faster than the constant structure and the convex and decreasing structures were slower. Similar results were found for the number of frames traversed. The patterns of search also differed greatly among the five structures in terms of the frequency of use of the 'previous' command and the 'top' command. The 'previous' command was used most frequently with the convex menu and least frequently with the concave menu. For scenario targets, the 'top' command was used most frequently with the decreasing menu and least frequently with the increasing menu. For explicit targets, the 'top' command was used most frequently with concave menu and least frequently with increasing menu. The pattern of search indicated that if subjects moved back up the tree they tended to move to a level with eight choices rather than two. Overall, it is concluded that the concave menu is superior when searching for scenario targets and the increasing menu is slightly superior when searching for explicit targets. A theory of menu uncertainly based on information theory is proposed which helps to account for some of the results.
Monitoring Computer Users' Behaviour in Office Environments BIBA 67-78
  Alan L. Teubner; Jerry J. Vaske
Monitoring computer users' behaviour to enhance the usability of software has become increasingly popular among the developers of computer products. One objective of this paper is to compare the use of a software monitoring technique with other field observation methods. It is argued that when contrasted against traditional observation techniques, the software monitor provides a more accurate and unobtrusive approach to collecting behavioural data on larger samples of individuals for longer periods of time than either participant observation or Thinking Aloud methodologies. A second objective is to demonstrate the application of software monitors on installed multi-user systems. Technical aspects of software monitor design and discussed, with emphasis on specific practical problems in monitor design and utilization when making the transition from the laboratory to an office environment. Data from two studies are presented. The first illustrates how a software monitor can be used to determine the impact and acceptance of various software packages in an integrated office automation product. The second study examines how software monitor data were used to modify the interface to a specific electronic mail facility.
Users' Experiences of COM -- A Computer-Mediated Communication System BIBA 79-99
  Lillemor Adrianson; Erland Hjelmquist
This questionnaire study reports results on some aspects of COM -- a computer-mediated communication system. It was found that the main reasons for starting to use the COM system were the need for information, curiosity and use as a complement to the telephone. The greatest benefits were being able to obtain information and opinions and to spread information. The COM communication was judged to be dominated by users with a good ability to communicate in writing. The subjects expressed a wish for a chairman who could organize the conference discussions. They experienced spontaneity in COM, but to a certain extent also misunderstandings and aggression. COM was generally seen as easy, friendly and stimulating. The general conclusion is that COM was thought of as an efficient tool for sending and receiving information which does not involve complex communication.

Book Reviews

"Foundations of Programming," by Jacques Arsac BIB 101-102
  Tom Axford
"Information Technology and the Role of the Librarian," by William Masterson BIB 101-102
  Hillary Dyer

BIT 1988 Volume 7 Issue 2

Editorial BIB 107
  Tom Stewart
The Knowledge Elicitation Problem: A Psychological Perspective BIBA 111-130
  J. St B. T. Evans
The knowledge elicitation problem arises from the need to acquire the knowledge of human experts in an explicit form suitable for encoding in a computer program such as an expert system. This is very difficult to perform successfully because of the size and complexity of knowledge structures in the human brain, and because much procedural knowledge is tacit and unavailable to conscious verbal report via interview methods. The present paper draws upon an extensive review of research in the field of cognitive psychology in an attempt to offer a practical approach to this problem. First, a wide range of cognitive theories concerning the nature of knowledge representation in humans is considered, and a synthesis of the current state of theory is provided. Second, attention is drawn to a number of performance factors which may constrain the exhibition of a person's underlying cognitive competence. There then follows a review and discussion of a number of alternative psychological methodologies that might be applied to the elicitation of different types of human knowledge. Finally, some suggestions are made for the application of the psychological work discussed to the practical problem of knowledge elicitation.
On the Memorability of Icons in an Information Retrieval Task BIBA 131-151
  M. W. Lansdale
Information retrieval involves the balance of two mnemonic processes: recognition of items presented to the user, and recall of where wanted documents might be. Iconic methods of human-computer interaction are seen to assist the recognition processes by virtue of the enrichment of cues provided. However, the principle of cue enrichment could apply equally to the process of recall, which is arguably a process more needing of support. This paper reports two exploratory experiments using icons to support the recall process in information retrieval. The results indicate no exceptional levels of recall. However, some aspects of users' performance suggest icons used in this way have some interesting and exploitable mnemonic properties. In particular, they may be useful in enhancing and supporting the search process by rapidly limiting the number of documents through which a user might be asked to search.
Expert-Novice Knowledge Organization: An Empirical Investigation Using Computer Program Recall BIBA 153-171
  Iris Vessey
Expertise in a given domain is regarded as being manifested in the knowledge structures or chunks that experts possess. This research sought to use the chunking hypothesis of expertise, operationalized via computer program recall, to distinguish the more expert from the less expert computer programmers in a group of programming professionals. Two expertise levels were required to investigate differences in debugging processes of expert and novice programmers. The programmer classification produced by the recall pretest, however, explained little variation in debug time and the number of errors subjects made, when compared with a programmer classification based on the effectiveness of subjects' chunking processes.
   Subjects' recalled programs were examined to determine whether the information structures in the program used for recall matched the declarative knowledge structures programmers could be expected to possess. Examination of the program suggests that programmers may not have been expected to possess two of the knowledge structures represented in the program. Examination of the programmers' recalled programs suggests that those programmers classified as experts by the process classification may have had greater difficulty in recalling this program than those classified as novices, leading to the hypothesis that experts may be affected more than novices by non-matching knowledge structures.
An Empirical Comparison of Menu Selection, Command Entry and Combined Modes of Computer Control BIBA 173-182
  Jonathan F. Antin
Command entry and menu selection control modes, along with a combination of the two, were compared using objective performance and subjective preference measures. Experienced as well as novice users of an interactive computer aided design system participated in the study. Although command entry produced performance that was superior to the other two modes, there was a strong user preference for the combined mode. Therefore, it is recommended that combined modes of control be considered for future systems, as long as there is a means to deactivate the menus. These results dispel the notion that menus are necessarily viewed as a hindrance by experienced users of a computer system.
The Effects of Feedback During Delays in Simulated Teletext Reception BIBA 183-191
  Miguel A. Planas; William C. Treurniet
In a teletext service, where pages of information are retrieved from a random sequence of pages broadcast cyclically on a video channel, the mean and variance of system response times are directly related to the number of pages in the cycle. Previous experiments have shown that viewers express annoyance with delays in system response, and that the frequency of such expressions increases non-linearly with delay. This paper examines possible explanations for the reported annoyance. A first experiment investigated the effect on annoyance of different kinds of feedback to the viewer following a page request. Results showed that annoyance grew at a slower rate when continuous, rhythmic feedback was present. Further, knowledge of the actual duration of the delay had no effect on viewers' annoyance. A second experiment showed that the same continuous feedback shortened viewers' estimates of the durations of the delays. These results suggest that annoyance is due to the experienced delay in system response time, and not to mere awareness of the duration.
The Social Shaping of Technology: The Case of a CNC Lathe BIBA 193-204
  P. T. Kidd
The design of most automated equipment is guided primarily by technical and economic considerations. The social science aspects of the system, if considered at all, are usually confined to the important questions of work organization, or man-machine interface design, or both. Social scientists are rarely given the opportunity to participate in the design of the often complex details of the actual technology that lies behind the man-machine interface. This paper gives consideration to the social shaping of computer numerically controlled lathe technology, which implies that social, technical and economic considerations are used in the design of the technology. The characteristics of a decision support system, which was designed jointly by engineers and social scientists with the aim of allowing the existing skills of machinists to evolve into new skills in relation to the changed technology, are described. The implications for the role and values of system designers that this system implies are also discussed.
The Interaction between the Use of Information Technology and Organizational Culture BIBA 205-213
  Yves V. H. Morieux; Ewan Sutherland
The interaction between the use of information technology, (IT) in organizations and that organization's culture is examined. The interaction is considered from the early stages of specification through to the regular use of the systems. The changes in the technological artifacts which result from the use of IT are discussed. Some suggestions about control of the interaction are made.

Human-Computer Interaction and Electronic Communication

Editorial Introduction BIB 215
  David J. Pullinger
HICOM -- Enter a New Communications Network BIB 216-218
  Mark Shuttleworth
Requirements for Mailbox Group Working BIB 219-222
  Paul Wilson
The State of the COSMOS -- Current Research and Future Possibilities BIB 223-226
  Sylvia Wilbur
Using an Electronic Communication System BIB 227-230
  D. J. Pullinger

BIT 1988 Volume 7 Issue 3

Editorial BIB 231-233
  Tom Stewart
Evaluating Expert Created, Idiographic Command Sets for Novice, Nomothetic Purposes BIBA 235-261
  Perry R. Morrison
Several studies have demonstrated performance benefits associated with self-defined computer commands (aliases). This study further investigated the possibility of empirically establishing pools of self-generated (idiographic) command names for novice (nomothetic) use by others. Experiment 1 showed that when given the commands and functions from which they were derived, independent expert (E) and novice (N) groups were able to discriminate between bona fide and bogus aliases at above chance levels, despite surface heterogeneity, although Es were more able to do this than Ns. Experiment 2 compared the understandability of E and N created aliases for independent groups of Es and Ns. Results showed that E aliases were more understandable than N aliases and that Es understood all aliases better. That is, Es exhibited a decoding advantage (due to experience) and an advantage in encoding semantic content in their aliases. In Experiment 3 Rosenberg's 'command suggestiveness' index showed that the mean suggestiveness of E aliases was significantly higher than that of N aliases. Moreover, for experiments 1 and 2, subjects' confidence in matching aliases to their parent functions was significantly correlated with suggestiveness. To test the utility of the suggestiveness metric, lists of high, medium and low suggestiveness aliases were constructed and subjects learned all lists in counterbalanced order. Recall using command functions as cues showed that more of the higher suggestiveness aliases were remembered. It was concluded that despite aliases surface heterogeneity, they possess sufficient semantic content to allow identification of their original functions. Moreover, Es produce more meaningful aliases and experiment 4 revealed that this may be due to greater suggestiveness. Generalizing from these findings, it appears appropriate that for complex systems where novice understanding is limited, performance may be facilitated by establishing E alias pools from which the most efficacious are empirically selected using the methods proposed by Rosenberg.
Callers' Perceptions of Post-Dialling Delays: The Effects of a New Signalling Technology BIBA 263-274
  Stephen J. Lupker; Gregory J. Fleet; Brian R. Shelton
The present paper reports an investigation of the potential impact of introducing common-channel signalling (CCS) into the current telephone network. This technology would have the effect of greatly diminishing post-dialling delay (PDD). As such, its main benefits would be obtained by introducing it into the toll network, in which PDDs are much longer than in the local network. The issues examined concerned potential 'contrast' effects, in that reducing PDDs in the toll network may cause callers to be less patient with normal PDDs in the local network. Three laboratory studies were undertaken to evaluate caller impatience and abandonment under (1) the current system, (2) the current local system with a simulated new toll system and (3) the current toll system with a simulated new local system. Ratings of impatience and abandonment increased on local calls when the new technology was implemented on the toll network, but not vice versa. The explanation offered is based on a 'cognitive' contrast effect resulting from callers' expectations that toll PDDs should always be longer than local PDDs. The implications of this effect for caller behaviour with the introduction of CCS are discussed. Any negative effects on local call behaviour are outweighed by the much shorter PDDs on the toll network and should be counteracted by the gradual introduction of CCS.
Hierarchical Planning as Method for Task Analysis: The Example of Office Task Analysis BIBA 275-293
  Suzanne Sebillotte
The paradigm of hierarchical planning from artificial intelligence literature is used to describe 133 office tasks. Various levels of abstraction are described. Their number varies with the tasks. From several examples we demonstrate that some of these levels can be grouped and that any task can be broken down using a four-level model: the most abstract level of the task formulation; the expert level, which represents specific context procedures (or subtasks); the highest common level, these being common procedures which are domain independent; the lowest verbalizable level, or elementary actions. Then the common procedures are considered as possible functions in a computer aided system. From a detailed analysis of some of these common procedures, we emphasize the obligatory or optional features of certain actions. Using an example we propose a function description which takes these features into account.
The Effects of an Active Development of the Mental Model in the Training Process: Experimental Results in a Word Processing System BIBA 295-304
  Michael Frese; Karen Albrecht; Alexandra Altmann; Jutta Lang; Patricia Von Papstein; Reinhard Peyerl; Jochen Prumper; Heike Schulte-Gocking; Isabell Wankmuller; Rigas Wendel
Three different training programmes for a word processing system were experimentally compared: (1) a sequential programme, which taught low-level skills and which did not help the user actively to develop a mental model, (2) a hierarchical programme, which provided an explicit and integrated conceptual model of the system to the user and (3) a programme in which the users were asked to develop hypotheses on the functioning of the software and in which they were encouraged to use an active and exploratory approach. From an action theory point of view it was hypothesized that the third group would be superior to the first group. In an experimental study with two training sessions each of two hours and a two-hour testing session (n=15), this was shown to be the case for several performance criteria (error time, transfer and experimenter rating). Additionally, an interindividual difference variable to measure the individual learning style was used, giving results that could be interpreted in a similar way to the experimental results.
Attitudes to Computers of Managers in the Hospitality Industry BIBA 305-321
  P. R. Gamble
Service industries are an increasingly important part of most developed economies as employers of displaced manufacturing labour. Computers are currently used in fewer than 30% of British service organizations and it has been suggested that this low level of utilization is attributable to cost. A study of managers in the largest service industry in the United Kingdom, the hospitality industry, shows that it is not cost attitudes but management attitudes which are the major inhibitory factor. Furthermore, these findings are consistent with other studies of attitudes to computers of senior managers in British industry. A comparison with a study of the attitudes of some American professional persons suggests that this is not necessarily attributable simply to educational level. In such circumstances it seems probable that penetration levels will increase slowly and that creative applications of computers in service industries, which may reduce their propensity to absorb labour, are not imminent.
The Contributions of Cognitive Engineering to the Design and Use of Expert Systems BIBA 323-342
  Chaya Garg-Janardan; Gavriel Salvendy
The cognitive basis of several issues critical to building expert systems is reviewed. Required and desirable attributes of a knowledge elicitation methodology are identified. It is argued that knowledge should be elicited from several levels of expertise. Plausible contributions of research in human problem solving to knowledge representation and design of control structure are examined. Implications of the literature on human learning and on expert-novice differences for machine learning are reviewed. The applicability of systems, which rely on user modelling, to the design of intelligent interfaces is discussed.

BIT 1988 Volume 7 Issue 4

Editorial BIB 343-344
  Tom Stewart
Technology Adaptation: A Typology for Strategic Human Resource Management BIBA 345-359
  Urs E. Gattiker
Organizational adjustment to technological change and its impact upon human resources should be based on an understanding of employee perceptions and behavioural responses to such developments. An analytic framework which relates strategic choice to internal labour market and employee determinism is derived from the methods employed by organizational theorists and psychologists to generate a set of typologies. The framework divides variables in four classes: (1) natural selection, with minimum choice and adaptation or selection out; (2) differentiation, with high choice and high employee determinism and adaptation within constraints; (3) strategic choice, with maximum choice and adaptation by design; and (4) undifferentiated choice, with incremental choice and adaptation by chance. Useful in reconciling previously inconsistent or incompatible theories of technological change, the framework is offered as a co-ordination device to integrate the results of divergent approaches to technological adaptation studies.
An Empirical Investigation of Two Electronic Mail Systems BIBA 361-372
  Frank Safayeni; Eric Lee; James MacGregor
Two electronic mail systems were studied within an organization: an executive system for upper management and a general system for all employees. The results from 130 interviews indicated that both systems were perceived to contribute to productivity, a reduction in use of memos and a reduction in telephone calls. However, the executive system was perceived to be more satisfactory and more effective than the general system. The ratio of positive comments to negative comments was used as an independent measure of satisfaction with each electronic mail system. The user comments were also classified and discussed in terms of types of perceived advantages and problems with these systems.
A Cognitively Based Methodology for Evaluating Human Performance in the Computer-Aided Design Task Domain BIBA 373-397
  Joseph Sharit; Donna L. Cuomo
This article describes a methodology for evaluating human performance in the computer aided design (CAD) task environment. The methodology is based primarily on cognitive theoretic frameworks that are consistent with processes presumed to underlie human design activities. The motivation for its development stems from rapid software and hardware advances in CAD systems and our relative lack of understanding of how these enhancements affect human design performance for (1) fundamentally different types of tasks and (2) different levels of complexity for a particular task. This methodology is currently being applied to computer aided architectural design, an area where artificial intelligence (AI), enhanced geometric modelling and other system features are being debated in terms of their usefulness in aiding the human's design activities.
Problems Associated with the Off-Line Programming of Robots BIBA 399-416
  Al Humrich; Iain Wilson
The introduction of robots into any organization forms part of a considerable investment in new technology over a range of applications in the search for efficiency and increased productivity. The greater consistency and quality associated with robot operations compared with that of the human operator, is regarded as an advantage both in relation to the manufacturing process and to the product. However, in order to make the most efficient use of robots, the ability to generate good robot programs must be developed. Traditional robot programming techniques are extremely slow and laborious. Off-line programming by textual input alone is an equally tedious process. In this paper, existing and potential problems associated with off-line programming are examined. Various commercial and experimental robot languages and their relative important features are described. Opinions of manufacturers of CNC (Computer Numerically Controlled) machines and robots concerning both user interfaces and potential candidate users are discussed. Both the training and the abilities of the candidate robot programmers are important aspects of any robot programming system, although obviously much will depend on the sophistication of the particular application. Recommendations are made concerning issues that should be taken into account when developing future off-line programming systems.
Human Intelligence Models and Their Implications for Expert System Structure and Research BIBA 417-430
  John Cook; A. Dale Whittaker; Ronald H. Thieme; Owen R. Smith; Gavriel Salvendy
In order to determine the features which contribute to intelligent behaviour, several models of human intelligence were studied. This evaluation identified ten features present in human intelligent behaviour which may be relevant to expert systems design. These features were rated, by experts in the field of expert systems, on the extent of their presence in expert systems as compared with intelligent human behaviour. The four features receiving the lowest ratings were social competence, comprehension, automatization and memory management. The paper concludes with a discussion of these four features and outlines what needs to be done and why in order to incorporate them into the structure of expert systems.
What We Know and What We Need to Know: The User Model versus the User's Model in Human-Computer Interaction BIBA 431-442
  Pamela Briggs
An argument is made for the importance of studying the real, as opposed to idealized, behaviour of the computer user. Formal methods which model user behaviour in terms of production rules are criticized because they fail to account for the unique behaviour which results either from problems arising in the normal work routine, or from novices who create their own patterns of interaction with the machine. This latter point is illustrated with reference to a study of novice users. How well are such users able to identify the knowledge they need when learning how to use a new system, and what kinds of knowledge of the system do they seek? It seems that in the absence of a suitable, generalizable model of a word processing system, these users structure their own learning experience badly, making poor use of the little experience they have. The behaviour of such users deviates markedly from that of the 'ideal user' captured within formal task descriptions.
The Concept of an Information Management System and its Use within Design Studies BIBA 443-455
  M. A. Tainsh
This paper describes the concept of an information management system (IMS) and its association with a user centered approach to the design of equipment which acts as a medium for the user to attain a goal. The case is taken where such mediating devices (MDs) involve computer based processing, storage and communications arrangements, and transactions with either a physical environment or engineered facilities. Particular attention is drawn to the designer's role in defining the data transformations within an MD, and hence to information that is available to the user and the means of carrying out transactions with it, transforming it into knowledge and attaining a goal. There is also special emphasis on the usability (= operability + trainability) aspects of the user's interface with the MD. The means of exploring design options are described along with the possible assessment techniques. It is concluded that the concept of an IMS offers considerable benefits for studying a range of design issues and matters of scientific importance.
A Feature Matching Approach to the Retrieval of Graphical Information BIBA 457-465
  James N. MacGregor; Eric S. Lee
This paper is concerned with how to provide easy and effective access to computer databases containing pictures rather than text. Experiences with several types of graphic material suggested that hierarchical menu indexes were unlikely to be effective. What seemed to be required was a probabilistic method which could tolerate a certain amount of error and uncertainty. As a solution, we propose a feature matching approach in which users describe what they want to retrieve. Descriptions are elicited by a set of queries. The system then matches user descriptions with descriptions of items in the database, and responds with the best fitting items.
   Initial tests suggest that this type of system may have the qualities we were looking for. First, it appears to tolerate errors, in that the user's description may disagree with the database's in a number of ways and still lead to the correct item. Second, it permits uncertainty, in that the user can choose the non-committal response of 'maybe' without the retrieval process coming to a halt. Third, it appears to require no training or manuals, only a few brief instructions on screen. It therefore seems well suited to the non-expert user.
Effects of Display Format on Proof-Reading with VDUs BIBA 467-478
  Anthony Creed; Ian Dennis; Stephen Newstead
Two experiments are reported which compared proof-reading performance in various VDU display formats. Experiment 1 found that displaying text one paragraph at a time improved the accuracy of performance, relative to a full screen condition, but at the expense of speed. Subject also preferred using the paragraph format. Display contrast (positive vs negative) had no effect on performance. Experiment 2 supported the findings of Experiment 1, and found increased accuracy when text was further subdivided into sentences, but speed was again reduced. Possible explanations for the format effect are presented, and its practical implications are considered.