You are on page 1of 6

Applied Ergonomics 1989, 20.

4,255-260
User-computer interface

Towards a practical method of


user interface evaluation
G.I. Johnson t, C.W. Ctegg and S.J. Ravden*

MRC/ESRC Social & Applied Psychology Unit, University of Sheffield, Sheffield Sl0 2TN, UK
tnow at Nederlandse Philips Bedrijven BV, Applied Ergonomics Group, Corporate Industrial Design,
Eindhoven, The Netherlands
*now at Ciba-Geigy Plastics, Information Management Support Centre, Duxford, Cambridge, UK

This paper describes a practical method for evaluating the usability of human-computer
interfaces. The paper specifies the requirements of such a method, and then outlines our
work in developing a method to meet this specification. The method is based on the
conduct of realistic tasks with an interactive system and the subsequent systematic
elicitation of end-users' and designers' reactions to the interface using a criterion-based
evaluation checklist.
Two practical examples are used to illustrate development of the method: (a) evaluation
of a prototype production scheduling system, and (b) comparative assessment of the
usability of three prototype user interfaces to a public-access database. The paper discusses
some issues raised by the method and considers how it can be further developed.

Keywords. Computers, human-computer interaction, criterion-based evaluation

Introduction the-art). This reflects the growth of interest in this area,


caused by the increasing use of computer-based systems, the
The objectives of this paper are to:
variety of people who are interacting with them, both
• specify a set of requirements for a practical method for within and outside work, and the rapid pace of technol-
evaluating the usability of human-computer interfaces; ogical advance.
• discuss the development and application of the method Evaluation can be seen as an intrinsic element within the
in two very different evaluation settings; process of user-interface design. Ideally, evaluation should
• examine its success in meeting the stated requirements; be an iterative process throughout design and development.
However, it more commonly occurs towards the very end
• consider the method's future development. of the design process. The commercial justification for
The paper is organised in six further parts. The first consideration of the human aspects of interface design and
briefly outlines current issues in user interface design and evaluation is clear. Sutcliffe (1988), for example, points
evaluation, especially from a human factors perspective, out that the cost of poor interface design can be:
stressing the significance of 'usability'. The second section
" l . Increased mistakes in data entry and system operation.
offers a set of requirements for a method for evaluating the
Mistakes cost money to rectify and errors which go
usability of interfaces. The third section describes how the
undetected can have very damaging consequences if
method was developed and used, drawing on experience of
decisions are taken on the basis of incorrect data.
conducting two practical usability evaluations of very
different computer-based applications. In the following "2. User f r u s t r a t i o n . . , may be manifest in low productivity
section, the extent to which the method is believed to meet employee s t r e s s . . , or simple under-utilisation of the
the specified requirements is evaluated. Some views system.
concerning outstanding issues in this area are offered, and
"3. Poor system performance. The system may not handle
the paper ends with some concluding remarks.
the volume of throughput it was designed for, or the
accuracy of output may not agree with the specification.
Current issues Poor interface design makes it too cumbersome to use
There is a burgeoning literature focusing on user inter- or too obscure to learn. Extra resources and money
face design and development (e g, Rubin, 1988; Shneiderman, have to be put into the system." (p 3--4)
1987; Sutcliffe, 1988) and human-computer interaction Similar arguments have been presented by Cohill et al
(see Helander, 1988, for a thorough review of the state-of- (1988) who comment on the benefits of user interface

0003 6870/89/04 0255 06 $03.00 © 1989 Butterworth & Co (Publishers) Ltd AppliedErgonomics December 1989 255
User-computer interface

evaluation in terms of increased user satisfaction, increased usability evaluation studies. The next section of the paper
sales, decreased development costs, increased productivity, describes this process and the progress towards our goal of
decreased product returns, and decreased training costs a practical usability evaluation method.
(p 130-131). There is little doubt that user interface eval-
uation, as a crucial part of that process, warrants attention. Development of the method
The way in which a simple, practical evaluation can be A large part of the literature in this area concerns the use
conducted, however, remains largely unresolved. As the pace of design guidelines (see Mosier and Smith, 1986; Smith and
of technological innovation quickens, and the design of Mosier, 1986). However, there are a number of problems
user interfaces involves more complex interaction techniques, with the usability and perceived utility of such guidelines
the need for simple evaluation tools is increasing. (Johnson et al, 1986; Smith and Mosier, 1984). Whilst based
on sound empirical work within such fields as cognitive
In this paper we have not attempted to review the
psychology (see, e g, Gardner and Christie, 1987), they are
existing work on user interface and software evaluation
not easily incorporated into the design and evaluation
methods. Rather, we focus on the specification and develop-
process. The underlying principles or criteria on which
ment of a practical method. The reader is referred to Karat
most guidelines are based, however, are valid.
(1988) and Lea (1988) who provide excellent reviews of the
area. The first stage in developing the method described in
this paper was to review the literature and establish a set of
In specifying an evaluation technique we sought to pro-
major criteria for 'usable' interfaces (Ravden, 1988). These
vide the user of such a technique - for example, a software criteria, presented in Table 1, were then translated into a
engineer or user interface designer - with a method that
checklist of specific questions. The idea was that a standard
could be explained and used without reference to relatively checklist could then be employed in an evaluation setting.
time-consuming and complex laboratory-based empirical
assessments. The general requirements of such a method are The checklist forms the core of the method, and
presented in the next section. comprises specific questions about features of the interface
in relation to particular criteria. For instance, the criterion
Requirements of an evaluation method of consistency incorporates a number of questions that
The requirements outlined below are based on our
Table 1: The usability criteria
appreciation of the existing literature on user interface
evaluation, and on our work with systems and software Visual clarity: Information displayed on the screen should
engineers in collaborative research and development be clear, well prganised, unambiguous and easy to read.
ventures (such as ESPRIT projects 534 and 1217 (1199)).
We therefore propose that a method for evaluating usability Consistency: The way the system looks and works should
of user interfaces should exhibit the following nine attributes. be compatible with user conventions and expectations.
It should be:
Informative feedback: Users should be given clear,
1. systematic - i e, it should be capable of being applied in informative feedback on where they are in the system,
a relatively straightforward and routine manner; what actions they have taken, whether these actions have
2. based on existing criteria, that are derived from empirical been successful and what actions should be taken next.
and theoretical work in the area;
Explicitness: The way the system works and is structured
3. iterative - i e, it should be capable of use within realistic should be clear to the user.
iterative design processes;
Appropriate functionafity: The system shou Id meet the
4. general, in that it can be applied in a range of settings on
needs and requirements of users when carrying out tasks.
a range of interfaces;
5. participative, in the sense that it can be used and under- Flexibility and control: The interface should be sufficiently
stood by a range of different 'experts' - for example, flexible in structure, in the way information is presented
including end-users, designers and human-factors and in terms of what the user can do, to suit the needs
specialists; and requirements of all users, and to allow them to feet
in control of the system.
6. sensitive - i e, capable of relatively fine discriminations
between the usability of similar interfaces; Error prevention and control: The system should be designed
to minimise the possibility of user error, with inbuilt
7. simple to use by people who are unfamiliar with usability
facilities for detecting and handling those which do occur;
issues;
users should be able to check their inputs and to correct
8. face valid and related to realistic usage of the system; and errors or potential error situations before the input is
9. reasonably exhaustive, insofar as it considers most of the processed.
relevant issues that constitute usability - i e, it should User guidance and support: Informative, easy-to-useand
not omit major factors affecting the usability of an relevant guidance and support should be provided, both
interface. on the computer (via an on-line help facility) or in hard
Ravden and Johnson (1989) describe a method which copy document form, to help the user understand and
attempts to meet these requirements. The method itself was use the system.
developed iteratively and refined through a number of

256 Applied Ergonomics December 1989


User-computer interface

could be asked concerning consistency of formatted text, • observing evaluators carrying out tasks and recording
menu structure, action sequences, use of icons and abbrevia- certain aspects of task performance is extremely valuable
tions, and so on (see Ravden and Johnson (1989) t'or details in identifying areas of difficulty (and ease) for end-users.
of the checklist). Table 2 illustrates a specific example During its development the method was employed in a
from the checklist. number of evaluation settings. In the first of these we used the
The second key feature of this method is that a checklist method to evaluate a prototype production scheduling system.
of this kind is best completed by people with experience of
undertaking realistic tasks using the system. Therefore,
Evaluating a prototype scheduling system
before completing the checklist, evaluators (system designers,
This practical study formed part of our work within an
human factors specilaists or end-users) should first carry out
ESPRIT Computer Integrated Manufacturing Project (No 534)
one or more tasks using the system. These should be rep-
and was concerned with the human factors aspects of flexible
resentative of 'real work' using the system, and should test
automated assembly. The prototype system in question was
as much of the system's functionality as possible.
developed to schedule and control the transportation of
Carrying out realistic tasks as part of the evaluation has pallets carrying components for assembly around a conveyor
a number of benefits: network to and from various assembly workstations. The
• the most effective way of exploring functionality is to software ran upon an industry-compatible microcomputer
use the system; and was intended to enable the operator of the system to
monitor, load, edit, create and save schedules for the pallet
• it enables evaluators to see the interface as part of the transportation.
overall system rather than as an artificial series of screens
and actions; The first stage in evaluating the usability of this user inter-
face was to study the tasks to be undertaken by the (potential)
• many problems, difficulties and aspects of usability are operator. This basic task analysis provided sufficient inform-
only exposed when carrying out real tasks; ation with which to create two representative tasks which the
evaluators carried out prior to completion of the checklist.
The 'evatuators' in this case fell into three categories - the
Table 2." Example of the evaluation checklist (Reprinted by system designer, human factors researchers, and design
permission of the publishers, Ellis Horwood, from engineers unfamiliar with the system.
Ravden and Johnson, 1989)
The procedure for the evaluation was as follows. Each
evatuator was presented with standard instructions detailing
the tasks to be undertaken; then attempted the tasks,
commenting as difficulties were encountered (these were
audio-taped); and completed the usability checklist con-
(bl in the same l a y o u t ? ferring with the system as necessary.
5. Does the appear in the same i n i t i a l The evaluation was able to identify certain weaknesses
position on d i s p l a y s of a s i m i l a r type?
within the user interface designed for the scheduling system.
6. Is the same item of i n f o r m a t i o n d i s p l a y e d in the Specifically, four areas emerged as requiring attention in the
same f o r m a t , w h e r e v e r it appears?
subsequent design stage. These areas were:
7. Es the format in w h i c h the user s h o u l d e n t e r
p a r t i c u l a r t y p e s of i n f o r m a t i o n on the screen (a) the use of colour, which was excessive and led to some
c o n s i s t e n t t h r o u g h t ~ u t the system!
confusion;
8. Is the method of e n t e r i n g i n f o r m a t i o n c o n s i s t e n t
throughout the system? (b) the menu positioning, which was inconsistent through-
out the interface;
9. Is the ,ict:,Jn r e q u i r e d to move the c u r s o r a r o u n d
the ( o n s l s t ~ r N t h r o u g h o u t the system?
(c) the lack of flexibility in the structure of the interface,
lO. ]s t~e rnetPod of s e e c t i n g c~ptior~s e . g . from a making transfer between elements of the system
ITl~rlt ~ corqsisten[ / h r m ~ g h o u t the s~stelrl ~
laborious and irregular; and
I]. Whet,: ,i k e v b u a r d is used, Jre the same k e y s
u~sed f?r the' sarm f u n c t i o n s t h r o u g h o u t the
(d) the clarity of data entry procedures.
~,, ster,!"
The evaluation method proved to be popular and easily
12. Are t h e r e s t a H d a r d p r o c e d u r e s for c a r r y i n g o u t used, and was shown to be successful in eliciting useful
s l m l a r , rehl',ed o p e r d t i o n s ? ( e , g , b u d d t m g a n d
delt, tlrlg infc~rmath~n ~ t a r t m g and f i n i s h i n g responses from evaluators with contrasting backgrounds.
t r d n s ~ctio,,s I
In particular, the designer of the system, initially overtly
13. Is the way the system responds to a p a r t i c u l a r cynical about the exercise, became very enthusiastic about
user a c t i o n c o n s i s t e n t at all times?
the evaluation and the relevance of the results produced. In
14. Are t h e r e , m y commer~ts (good a r bad) you w i s h to add r e g a r d i n g the abo~e
particular, he was able to use the findings to look at the next
issues '
version of the software with a clear idea of existing problems,
15. O v e r a l l , how would y o u r a t e the system in terms of c o n s i s t e n c y ? both from a 'user's' perspective and also as a result of focusing
{please t i c k a p p r o p r i a t e box below)
on specific problematic aspects of the interface. During this
evaluation it became clear that the method was performing
Ver,, Moderately Neutral Moderately Very
sati,factorv satisfactory unsatisfactory unsatisfactory well in its overall assessment of the interface, and that the
people carrying out the evaluation were impressed by the
amount and specificity of the information produced.

Applied Ergonomics December 1989 257


User-computer interface

A number of important areas for improving the method of the time taken to carry out the task, number of times
were revealed by this study. For example, some usability referring to help facilities, incorrect menu option choices,
problems were not adequately tapped by the checklist, the and comments made during the task.
tasks proved too difficult and lengthy, and limits on the
Three tasks were developed for the evaluation: general
evaluation were posed because some of the software was not
fully operational. browsing through the database for information; a general
search (defined by two parameters) with a subsequent
As a result of this evaluation the checklist needed to browse of information; and a specific search (again requiring
include: the user to input particular search constraints), with only
• 'deeper' issues concerning the user's view of the system; one correct 'answer'. Participants were allocated to cells
within this 3 x 3 (interfaces and tasks) design, each user
• more detail in the sections of the checklist dealing with attempting one task using one of the three interfaces. No
'surface' features (e g, information format, specifically evaluator was exposed to more than one of the interfaces.
for graphical and symbolic information); and
The procedure for each evaluator was as follows. The
• a means of assessing the perceived functionality of the evaluator was introduced to the nature of the study, and
system, and its appropriateness. shown the interface in question. Evaluators were able to ask
With revisions, in line with the above, made to the coverage any questions about the evaluation and the task. The evaluator
of the checklist we then proceeded to conduct another soft- was informed about the type of system and its purpose, and
ware evaluation exercise, the assessment of alternative user presented with a brief questionnaire about his/her back-
interfaces to a public-access database. ground, experience with computers, etc. The task was carried
out with an observer noting progress. On average, the time to
carry out the task was 25 min. Following this, evaluators
Comparative evaluation of database interfaces were shown the checklist and asked to complete it for the
This evaluation study involved a comparative assessment interface they had used.
of the usability of three prototype interfaces. Each was
developed independently by separate commercial organisa- The information gathered in this evaluation was collated
tions with the requirement that they be as 'easy to learn and analysed and found to be invaluable in pinpointing the
and use' as possible. The intended user population of this strengths and weaknesses of the interfaces under examination.
large database system was the general public. Furthermore, the utility of the method was demonstrated.
The user interfaces were representative of modern public- In performing this usability evaluation, we found that
access front-ends, adopting a touch screen or touch-pad input the method was able to discriminate between what were
medium, along with coloured displays using text, icons and essentially three 'well-designed' interfaces, and that the
windows. Each system was run on an 'industry-standard' usability checklist was acceptable to a wide range of
microcomputer, without the familiar QWERTY keyboard, evaluators. Furthermore, the results of the evaluation proved
each making use of simple, non-voice audio feedback. One useful to the research sponsors who were seeking to develop
could regard these as similar to those public systems installed a new interface. The findings were specific enough to allow
by banks and building societies, the major difference being them to alter their requirement specification for potential
the complexity and degree of access to the underlying data- suppliers. In addition, they were slightly surprised by the
base. The database of information accessed by these proto- outcomes, since their favoured prototype in practice fared
type interfaces held details of recreational, training and less well in evaluation than one of the alternatives.
educational courses. All of the systems were to be used
without any supporting documentatiort (i e, there were no
user guides or manuals). Meeting the requirements - an assessment
The users recruited for the evaluation exercise were, as In the third section we outlined the requirements of an
far as was possible, representative of the general public. Thus, evaluation method for interface design. In this section our
the sample of over 60 users covered both sexes, a range of method is considered against each of these requirements.
ages (over 16 but under 60), and was drawn from a number
1. ls it systematic? The method requires that evaluators
of occupations, including hairdressers, shop assistants,
conduct representative tasks and then work systematic-
administrators, redundant mineworkers, students with
ally through a series of questions based on usability
learning difficulties, housewives, caretakers, computer
criteria. It can and should be used systematically,
science students, teachers and secretaries. All participants
although its use may be constrained by the availability
in the evaluation were unpaid volunteers.
of users, the time available and other practical consider-
The first part of the evaluation involved the analysis of ations. These, of course, are constraints which exist for
tasks carried out using similar (information retrieval) any fairly plural and rigorous method.
systems. Having gathered basic information, we devised a
2. Is it based on existing criteria? Yes, the criteria were
set of tasks to be carried out by 'users' in the evaluation. generated from current theoretical and empirical work.
These were modified following a pilot study. Specifically, As further research is conducted, new criteria may emerge
the degree of task difficulty and the number of tasks to be
and can be included in the format.
undertaken by users were reduced. The pilot study also
permitted the testing of the procedures for recording 3. Is the method iterative? Its use requires an existing proto-
observational data. The type of information sought from type. It can then be repeated with revised or new versions.
user-system interactions was consequently limited to records Thus during later stages of design the method can easily

258 Applied Ergonomics December 1989


User-computer interface

encompass, andindeed it can encourage, iterative working. Relative importance of usability issues
In principle there is no reason why the method cannot be One potential criticism of the method is that it fails to
presented to designers working at earlier conceptual distinguish between major and minor problems. Is poor use
stages as a means of influencing their design ideas, but of colour potentially a serious problem? In a comparison
it has no formal application at this stage; although the between two competing systems, is it more or less important
criteria would remain relevant during conceptual design, than inconsistent menu formatting?
the mode of presentation may need to be significantly
One approach to this difficulty would be to try to weight
different.
the significance of certain criteria, and the particular issues
4. Is it general? This issue is unproven. Whilst we have found within them. Our own view, however, is that such an enter-
the method applicable to several different kinds of inter- prise, at least for the forseeable future, is misguided. It
face, it seems likely that different systems may require would need to take account of a wide range of variables
the definition and elaboration of new criteria and new such as the nature of the task, the system and the user
questions. We hope the method is sufficiently flexible to population. There would be a real danger that the method
cope with such system-specific additions and deletions, became cumbersome and that more energy was invested in
and indeed we anticipate that it will be most effectively the abstract weighting of priorities than in debate about user
used in this way. preferences.
5. Is it participative? Thus far we have found the method Our view of the method at this stage of development is
does allow for, and indeed encourages, participation. For that it is a tool for helping users, designers and others
example, we have tbund that designers and users have identify the strengths and weaknesses of an interface. They
had little difficulty in using the method. We should add, themselves are in the best position to decide the relative
however, that the outcomes of the evaluations have importance of particular issues, and what action needs to be
required some input from people with some expertise in taken. The best safeguard that the outcomes of such eval-
this area. We are' not claiming the method is 'free-standing'. uations are useful is not to resort to quasi-mathematical
6. Is it sensitive? Again this is unproven, but in our second weighting formulae, but to ensure that end-users are
study described above the method seemed well able to involved in discussing and interpreting the results.
identify specific problems with the three different inter-
faces. It was also able to separate out their relative merits
and demerits, even though each had been specifically Quantifying the method
designed to be 'user-friendly'. To some people a method of this kind should ideally
result in quantifiable outcomes which allow some com-
7. Is it simple to use? This is more problematic and raises a parison between interfaces, and perhaps even some notional
number of issues of general relevance to any type of scoring of the 'usability' of a particular system (e g, in
relatively specialist evaluation. The method is certainly percentage terms). For the reasons outlined in the previous
easy to administer and easy for users to conceptualise. section, we believe this an unrealistic goal. For example, a
However, like any other form of evaluation, the inter- single problem concerned with functionality may render a
pretation of the results require some expertise. system virtually unusable to a sub-group of users, even
8. Is it face valid? The users of the method have appreciated though it 'scores' highly on all other dimensions. Epistem-
that it relies heavily on the conduct of realistic tasks ologically, our position is that different stakeholders
which should be chosen both to represent its typical use (Mitroff, 1984) interpret and perceive their work and their
and also to explore the functionality of the system. Our IT systems quite differently, and thereby the notion of a
experience is that the method appears face valid. unitary system for 'scoring' usability is misguided.

9. Is it reasonably exhaustive? We would expect that the


method will need editing as it is applied to a wider range Relationship to performance metrics
of interfaces. In time we hope that the criteria and check- The relationship between self-reported usability as
list are extended to form a menu from which prospective assessed using this method and objectively measured indices
users can select relevant evaluation topics. of system use and performance remains unclear. Ideally,
Our conclusion is that the method lares reasonably well both forms of evaluation should be included if resources
against the requirements specified earlier in this paper. permit. It remains an important research issue to examine
Clearly, however, we make no claims that it is a 'finished' how and when usability, as assessed in this method, becomes
piece of work ready for universal application. In the next manifest in performance problems.
section of this paper we discuss the prospects for the method
and some of the substantive and methodological issues
which arise from our experience with its application. Applicability of the method
The two key issues here were discussed briefly earlier.
The first concerns the stages in the design process at which
the method can be applied. Our experience is that it is best
Discussion used when operational prototype interfaces are available for
test. We are not especially confident that giving copies of
This part of the paper considers some important issues the checklist to designers who are still conceptualising their
arising from the development and use of practical methods work (i e, they are in the early stages of design) is especially
of the kind reported in this paper. productive.

Applied Ergonomics December 1989 259


Uwr-computer interface

The second issue concerns the potential range of appli- A final more speculative comment should be made. It
cations or systems. In deliberately attempting to devise a seems likely that the role of methods such as this is essenti-
generalisable method, it is perhaps incumbent upon us to ally to help people to avoid developing interfaces that are
specify those circumstances where it is, and is not, appropriate. difficult to learn and use. Its function may thus be seen as
For example, is it beyond the scope of the method to apply it rather negative. Whilst it is conceivable that, with regular
to a video-conferencing system which relies heavily on the use use, designers may internalise these criteria and the more
of voice input/output and televisual media? Clearly the user-centred perspective that lies behind them, and that this
method should now be employed in as many evaluation may promote new ideas and approaches, we would not
settings as possible so as to establish its scope of utility. As we claim this method encourages innovative and creative
said earlier, our ambition would be to develop a menu of designs. Perhaps other techniques and tools are required
criteria and checklists which would cover a wide range of inter. for such purposes?
faces, from which evaluators should select the most relevant.

Concluding remarks References


This paper has described the development of a practical
Clegg, C.W., Warr, P.B., Green, T.R.G., Monk, A., Kemp, N,J.,
Allison, G., and Lansdale, M. 1988, 'People and computers
method for evaluating the usability of human-computer - H o w to evaluate your company's new technology'. Ellis
interfaces. The paper has proposed the requirements of such Horwood, Chichester.
a method, examined its development with respect to two
Cohill, A.M., Gilfoil, D.M., Pilitsis, J.V., and Carey, T. 1988,
case studies, and discussed the method's general suitability.
Measuring the utility of application software. In: Hartson,
The method, which employs a typical framework for H.R. and Hix, D. (Eds), 'Advances in human-computer
user-centred interface evaluation, is based upon a detailed interaction, Volume II'. Ablex, Norwood, New Jersey.
checklist of criterion-driven questions. The use of the Gardner, M.M., and Christie, B. (Eds) 1987, 'Applying
method depends to a large extent upon those factors that cognitive psychology to user-interface design'. John Wiley
influence all such evaluation exercises including: & Sons, Chichester.
• availability of time within the design and development Helander, M. (Ed) 1988, 'Handbook of human-computer
cycle; interaction'. North-Holland, Amsterdam.
• availability and willingness of user (and designers) to Johnson, G.I., Clegg, C.W., and Ravden, S.J. 1986,
participate in the evaluation; 'Human factors in the design of a flexible assembly system:
Design guidelines for design guidelines': In: Oborne, D.J.
• definition of tasks that are comprehensive and repre- (Ed), 'Contemporary Ergonomics 1986'. Taylor and Francis,
sentative of the intended use of the application; and London.
• availability of operational prototypes. Karat, J. 1988, 'Software evaluation methodologies'. In:
Helander, M. (Ed), 'Handbook of human-computer inter-
However, the method adopts a different perspective to action'. North-Holland, Amsterdam.
many usability evaluation approaches in that it:
Lea, M. 1988, 'Evaluating user interface designs'. In:
• is aimed at being especially usable by evaluators without Rubin, T. 'User interface design'. Ellis Horwood, Chichester.
experience in the collection and interpretation of Mitroff, l.I. 1984, 'Stakeholders of the mind'. Jossey-Bass,
behavioural data; San Francisco.
• requires 'user' involvement and the systematic collection Mosier, J.N., and Smith, S.L. 1986, Behaviour and ln/br
of their judgements; Technol, 5, 39--46. Application of guidelines for designing
user interface software.
• explicitly encompasses both representative tasks and
usability criteria. Ravden, S.J. 1988, Int J Computer Applications in Technol,
1, 35-42. Ergonomics criteria for design of the software
One of the most positive attributes of the method is the interface between human and computer in CIM.
degree to which it appears valid to users. Whilst the language Ravden, S.J., and Johnson, G.I. 1989, 'Evaluating usability
of usability evaluation is often a frustrating hurdle for of human-computer interfaces: A practical method'. Ellis
system designers and users (e g, cognitive compatibility, Horwood, Chichester.
users' conceptual mental models, and the formulae derived
Rubin, T. 1988, 'User interface design'. Ellis Horwood,
from keystroke data), the checklist offers a straightforward Chichester.
operational definition of usability and is relatively trans-
parent in how it works. The void that is often apparent Shneiderman, B. 1987, 'Designing the user interface:
Strategies for effective human-computer interaction'.
between the production of learning curves, usability plots or
Addiscn Wesley, Reading, Mass, USA.
annotated verbal protocols, and their subsequent translation
to recommendations for modification, is markedly reduced. Smith, S.L., and Mosier, J.N. 1984, Behaviour and lnfi~r
For example, if the consistency of command/option names Technol, 3(3), 195-203. The user interface to computer-
based systems: A survey of current software design practice.
or the informational content of error messages are problems,
then they are quickly identified. Of course, the idea that we Smith, S.L., and Mosier, J.N. 1986, 'Guidelines for user-
examine the utility of such evaluation techniques from the system interface software'. Technical report ESD-TR-86-
perspective of the end-user shoulc( be encouraged with 278, MTR 10090. Mitre Corporation, Bedford, Mass.
respect to all 'products' and 'systems' produced by the Sutcliffe, A. 1988, 'Human-computer interface design'.
human factors community. Macmillan Education, London.

260 AppliedErgonomics December 1989

You might also like