HSINese
`aati
`pianKoes
`
`ceeens
`
`'
`
`APPLE 1012
`
`1
`
`APPLE 1012
`
`

`

`
`
`
`
`2
`
`

`

`WITHDRAWN
`
`PROPERTY OF LIBRARY
`NAVY CENTER FOR
`ARTIFICIAL INTELLIGENCE
`Copy
`lof 2
`
`3
`
`

`

`Advances
`in
`Human—Computer
`Interaction
`
`Volume 3
`
`Edited by
`H. REX HARTSON
`DEBORAH HIX
`
`Virginia Polytechnic Institute and State University
`
`(A) ABLEX PUBLISHING CORPORATION
`
`NORWOOD, NEW JERSEY
`
`4
`
`

`

`Copyright © 1992 by Ablex Publishing Corporation
`All rights reserved. No partof this publication may be reproduced,stored in a
`retrieval system, or transmitted,in any form or by any means, electronic, me-
`chanical, photocopying, microfilming, recording, or otherwise, without per-
`mission of the publisher.
`
`Printed in the United States of America
`
`ISBN: 0-89391-751-6
`
`ISSN 0748—8602
`
`Ablex Publishing Corporation
`355 Chestnut Street
`Norwood, New Jersey 07468
`
`5
`
`

`

`Contents
`
`Preface
`
`Andrew Sears
`Catherine Plaisant
`Ben Shneiderman
`
`Dermot P. Browne
`Robert Summersgill
`
`Jakob Nielsen
`
`Expanding the Scope of Toushscreen
`Application: High Precision, Dragging
`Icons, and Refined Feedback
`
`The User Interface: The Poor Relation
`in Structured Methods
`
`Evaluating the Thinking Aloud Technique
`for Use by Computer Scientists
`
`Richard A. Wagner
`David Levinson
`David Jank
`
`Where East Meets West: Bridging the Gap
`between Anthropology and Computer
`Science
`
`Designing a Scholars’ Electronic Library:
`TheInteraction of Human Factors and
`
`Computer Science Tasks
`
`Hypermedia: Design for Browsing
`
`User Support: Considerations, Features,
`andIssues
`
`The Interactive Transaction System (ITS):
`Tools for Application Development
`
`Martha J. Lindeman
`Charles Crabb
`John R. Bonneau
`Vera Fosnot Werhli
`
`Bernice T. Glenn
`Mark H.Chignell
`
`Alison Lee
`
`Stephen J. Boies
`William E. Bennett
`John D. Gould
`Sharon L. Greene
`Charles Wiecha
`
`Author Index
`Subject Index
`
`34
`
`69
`
`83
`
`105
`
`143
`
`184
`
`229
`
`277
`283
`
`6
`
`

`

`Preface
`
`Whenwestarted working in the field of human-computer interaction
`over ten years ago, there wasn’t a field. We got strange looks from
`people whosaid “HCTisn’t research;it’s just commonsense!” Nowitis
`a fast-growing area of research and development. However, when we
`look aroundat advancesin thefield, we discover that many hard prob-
`lemsare still in front of us.
`The themeofthis series is the cooperation between behavioral scien-
`tists and computer scientists in developing human-computer inter-
`faces. Recognition of the need for multiple roles in interface develop-
`menthas led to increasing diversity in this field. We welcomethis as a
`positive indication of the contributions that a wide variety of perspec-
`tives can bringto bear on the openissuesof the field. This diversity is
`reflected in the present volume, which includes technology, meth-
`odology, techniques, case studies, applications, user support, and tools.
`This is a nice continuation from VolumeII, in which we stated in the
`preface that cooperation was needed in all the major areas of current
`work. Theory, modeling, methodologies, tools, and evaluation were
`included in VolumeII;all those sametopics plus technology and case
`studies are now addressedin this volume.
`In the half-decadesince VolumeI ofthis series appeared, in 1985, we
`have seen a dramatic increase in the conferences, journals, and other
`publications that address HCI issues. The ACM SIGCHI conferences
`have steadily increased in size every year; 1990 saw an unprecedented
`attendance of more than 2,300 participants. In fact, SIGCHI itself has
`been the fastest growing special interest group of the ACM forseveral
`years. Also since that first volume appeared, the ACM UserInterface
`Software and Technology — UIST — Symposium has been created. A
`sizable portion of the annual ACM SIGGRAPHConference is now de-
`voted to user interface work. Also since VolumeI, the Human Factors
`Society Conference has increased its attention to human-computer in-
`teraction, the Interact International Conference has grown, and the HCI
`International Conference has begun.
`
`7
`
`

`

`vi
`
`PREFACE
`
`One phenomenon that arises in HCI is the introduction of “gee
`whiz” technology, such as data gloves, touchscreens, artificial realities,
`and multimedia. While interfaces using this kindof technology can be
`glamorous and engaging in themselves, they can still be neutral with
`regardto usability. Technology alone can makefancy interfaces, but not
`necessarily good interfaces. Thefirst chapter, by Sears, Plaisant, and
`Shneiderman,is an example of how behavioral analyses coupled with
`software skills can lead to creative ways to improve dramatically the
`usability of existing technology. In the case of touchscreens, technology
`has beenlimited by the imaginations of those who have applied it. This
`team effort has freed up the possibilities of touchscreen applications.
`This relationship betweentherolesis also a key aspect of Chapter2,
`by Browne, Summersgill, and Stradling, that highlights the important
`methodological differences between software development and user
`interface development,especially in the early phases of the life cycle.
`Nielsen’s Chapter 3 also addressesthe needfor bringing the behavioral
`role, and especially the empiricism it offers, into the development pro-
`cess. Because it attempts to make an empirical process usable by com-
`puter scientists during development, this is a specific example of work
`that promotes cooperation amongthe behavioral and computer science
`roles in the interface development process.
`Case studies are also important, especially in a relatively new field
`where there is a need for successstories and examplesto follow. Chap-
`ter 4 by Wagner, Levinson,and Jank and Chapter 5 by Lindeman,Crabb,
`Bonneau, and Wehrli are both examples of case studies of interface
`developmentinvolving cooperatingroles. Chapter4 states the case for
`the role of social behavioralscientists to provide user supportin specif-
`ic application areas through knowledge of information needs. This
`chapter discusses a development project from the perspective of the
`snteraction of a broad variety of team roles, and arguesthat the bound-
`aries of HCl as a discipline must be broadened even more to encompass
`the specific needs of real-world application development. In Chapter5,
`the application is a multimedia document retrieval system, and cooper-
`ation of team roles is stressed throughout the development process
`with continuous emphasis on the user’s environment and support of
`the user in the application.
`In another area of application, Chapter 6 by Glenn and Chignell
`addresses the usability of hypermedia, especially for browsing, again
`makingthe casefor usability to catch up with new technology. Here the
`authors propose that the structural properties of hypermedia, so impor-
`tant in internal design,be used to advantagein providing visual naviga-
`tion landmarksfor the user. This is followed in Chapter7 by Lee, who
`gives thought to the question of user support in general, especially for
`
`8
`
`

`

`PREFACE
`
`vii
`
`computer-based individual work. This chapter makes the case for user
`support as an important elementof the interface and proposesan in-
`teraction history facility as a user support tool.
`Chapter 8, by Boies, Bennett, Gould, Greene, and Wiecha,presents
`an outstanding example of the kind of interactive tools becomingavail-
`able to support concepts such as the ones presentedin earlier chapters.
`This chapteris especially appropriate to complete this volume because
`it ties together methodologies, techniques, and tools—all vitally impor-
`tant areas of HCI research and development.
`We are pleased with this volume andthediversity it represents in
`this exciting field of HCI. We wish to thank the authors who have
`worked hard to maketheir chapters the best possible. Wearealso grate-
`ful to the reviewers whospenta great dealof time critiquing the chap-
`ters. Jo-Anne Lee Bogner and SusanStolarski, our “wondersecretaries”
`over the course of this volume, have kept us in line, on schedule, and
`organized, helping to maketolerable the difficult job of coordinating
`and editing such a volume. We are especially pleased that Ablex Pub-
`lishing continues to encourage usin ourefforts to publish this series, in
`recognition of the importance of the advances that are being made in
`the area of human-computerinteraction.
`
`Rex Hartson
`Debby Hix
`Blacksburg, Virginia
`November 1990
`
`9
`
`

`

`CHAPTER1
`
`A NewEra for High Precision
`Touchscreens
`
`Andrew Sears
`Catherine Plaisant
`Ben Shneiderman
`
`Human-ComputerInteraction Laboratory &
`Department of Computer Science
`University of Maryland
`
`INTRODUCTION
`
`Onegoal of human-computerinteraction research is to reduce the de-
`mandson users when using the computer. This can be done by reduc-
`ing the perceptual and cognitive resources required to understand the
`interface or by reducing the motoreffort to use the interface. The intro-
`duction of alternative input devices, such as the mouse and joystick,
`significantly improved some user interfaces. The touchscreen com-
`bines the advantagesof these other devices with a very direct method of
`inputting information. Users simply point at the item or action of in-
`terest, and it is selected.
`While many input devices allow interfaces to be customized, in-
`creased directness distinguishes touchscreens. Touchscreensare easy
`to learn to use, fast, and result in low error rates when interfaces are
`designed carefully. Many actions which are difficult with a mouse,
`joystick, or keyboard are simple when using a touchscreen. Making
`rapid selections at widely separated locations on the screen, signing
`your name, and dragging the handsof a clock in a circular motion are
`all simple when using a touchscreen, but may be awkwardusing other
`devices. Even when a task can be accomplished with other input de-
`vices, users may have to clear their workspace for the mouse or press
`many keys to movethe cursor.
`
`10
`
`10
`
`

`

`2
`
`SEARS, PLAISANT, AND SHNEIDERMAN
`
`Touchscreens have long been thought of as being simple to use.
`Unfortunately they have a reputation as being practical only for select-
`ing large targets and as being error prone. Recent empirical research, as
`well as advances in touchscreen hardware, have dramatically im-
`proved the performance of touchscreensandthe range of applications
`for which they can be advantageously used. Even with these advances,
`today most touchscreen applications emphasize the metaphorof “but-
`tons” being pressed onthe screen. Tasks such as dragging an object on
`the screen, moving the marker on a slider, or freehand drawing are
`rarely attempted with touchscreens, but webelieve that touchscreens
`can excel in such cases. This chapter presents recent empirical research
`whichcan providea basis for theories of touchscreen usage. We believe
`recent improvements warrant increased use of touchscreens. Human
`factors specialists, psychologists, and computerscientists have a grand
`opportunity to influence further developments andrefine theories in
`these new domains.
`
`ADVANTAGES AND PERCEIVED DISADVANTAGES
`OF TOUCHSCREENS
`
`There are many advantages to touchscreens which have made them
`popular for public access situations.
`
`Advantages
`
`Directness: Oneof the biggest benefits of a touchscreenis its directness.
`Unlike indirect devices such as a mouse, joystick, or keyboard,
`touchscreen users simply point at the desired object, andit is
`selected. There is no need to remember a complex syntax, search
`for the input device, remove visual focus from the objects of in-
`terest, or press multiple keys to movethe cursor. More important-
`ly, there is no need for users to map hand motions to cursor mo-
`tions, as required by manyother input devices. Sliding, dragging,
`andgestural input also benefit from the touchscreen’s directness.
`Speed: Thetouchscreenis the fastest selection device for manytasks.
`Users do not needto reachfor the input device whenitis time to
`makea selection as they often do with a mouseor lightpen. An
`additional advantage in manysituations is the lack of a cursor
`when users are not touching the screen. Users simply touch the
`
`11
`
`11
`
`

`

`A NEW ERA FOR HIGH PRECISION TOUCHSCREENS
`
`3
`
`desired location rather than touching a cursor and draggingit to
`the desired location.
`Ease of learning: Touchscreens are easy to learn to use. Once users
`realize that they must simply touch the screen to interact with the
`computer, they quickly master simple actions such as touching
`buttons or dragging itemsacross the screen. Unlike the mouse or
`tablet, there is no needto learn and practice spatial reorientation
`and hand-eye coordination (Nielsen & Lyngbaek, 1990).
`Flexibility: Touchscreen interfaces offer flexibility not available with a
`keyboard. Each interface can be customized for each specific task
`performed. Users can choose which keyboard layout they prefer,
`QWERTY, Alphabetic, or Dvorak, since it is displayed on the
`screen.
`
`No movingparts: The lack of moving parts contributes to the durability
`of touchscreens that has made them popular for applications such
`as information kiosks at amusement parks, office buildings, or
`museums. Unlike a mouse or keyboard, only the touchscreen
`must be accessible to users, making loss or damage of hardware
`less likely. One system, an information kiosk developed for the
`Smithsonian, traveled to museums across the country for two
`years. These touchscreens were heavily used and neverfailed.
`However, the video monitors did ultimately fail from abuse dur-
`ing shipping.
`No additional desk space: Touchscreensfree desk space for other uses.
`Many input devices, such as the keyboard and mouse, require
`desk space which maybe very limited. A related benefit is that the
`touchscreenis in a fixed location. Unlike the mouse or lightpen
`there is no need to search for the device which may be hidden
`underpapers. If the user is currently working with the computer,
`the screen must be accessible. This is particularly useful for ap-
`plications requiring only occasional pointing.
`
`Perceived Disadvantages
`
`There are also some problems that have been associated with touch-
`screens. Many of these problems have been overcomeor reduced by
`improvements in touchscreen technology or design strategies that have
`been developed for touchscreen interfaces.
`
`Lowresolution: This is one of the biggest misconceptions about touch-
`screens. Many people have reported on the low resolution of
`touchscreens. Someresearchers have claimed that the resolution
`of a touchscreenis limited bythe size of users’ fingers, and others
`
`12
`
`12
`
`

`

`4
`
`SEARS, PLAISANT, AND SHNEIDERMAN
`
`have claimedthatselection of single characters would be slowifit
`was even possible. Recent research has shownthat targets 0.4 x
`0.6mm could be selected with touchscreens (Sears & Shneider-
`man, 1991). The same research concluded that targets 1.7 x
`2.2mm could beselected as fast with a touchscreen as they could
`with a mouse, with similar errorrates.
`Arm fatigue: This could be one of the mostsignificant problems with
`touchscreens. Using a touchscreen at the angle most monitors are
`currently mounted can lead to arm fatigue, making them difficult
`to use for extended periodsof time. Renewedinterest in reducing
`fatigue appears to have resulted in simple changes to the touch-
`screen positionthatwill significantly reduce this problem (see the
`section on Workstation Design on page 24 for more details).
`Parallax: When touchscreens werefirst introduced, the infrared tech-
`nology was prevalent. Early infrared touchscreens had the touch
`sensing devices mounted abovethe surface of the monitor. When
`users’ fingers were close enoughto the screen, the infrared beams
`would be broken, resulting in a touch. This could occur long
`before the user meant to touch the screen. Newerinfrared touch-
`screens, and all other technologies, sense touches much closerto
`the monitor surface, if not directly on the surface, reducing the
`problem with parallax. Software strategies have also been ex-
`plored that reduce problems created by residual parallax by cor-
`recting for offsets created by the parallax and providing feedback
`to users about their exact position.
`Glare and smudges: Glare and smudgeson the monitor are of concern
`to many designers. Mounting the monitorat a better angle, using
`lightly ground glass surfaces, and paying careful attention to the
`lighting near the workstation can significantly reduce the glare
`problem. Smudgesare unattractive and can obscurethe display.
`Reducing smudges simply requires users to clean the monitor
`occasionally. On the other hand wefind that some touchscreens
`have less problems with accumulating dust
`than standard
`monitors. In our laboratory environment, we find ourselves clean-
`ing the mouse pad and mechanicalparts more often than we clean
`the touchscreens.
`Obscuring of the screen: Thefact that users use their fingers to make a
`selection by touching the screen implies that the users’ hand will
`obscure a part of the screen. Careful design of the interface, plac-
`ing selectable items in locations that will keep the user’s hand
`from obscuring the screen, can significantly reduce this problem.
`Whenpossible, the handedness of users should be considered
`
`13
`
`13
`
`

`

`A NEW ERAFOR HIGH PRECISION TOUCHSCREENS
`
`5
`
`whendesigninginterfaces, or users could be allowed to customize
`the software for the left or right hand.
`Limited tactile feedback: Visual and audible feedback should be used
`to compensate for limited tactile feedback in button applications.
`Tactile feedback is particularly important when performing rapid
`button presses without watching the screen. An exampleis typing
`on a touchscreen. Whenusers type on a traditional keyboard, the
`edges of the keys help orient their hands and the motion of the
`keys indicates when they are pressed. Thesecuesare not available
`with touchscreen keyboards. Visual and audible feedback can sup-
`plement the physical contact with the screen to help compensate
`for the absence of key motion,but identifying whenthe edgeof the
`touchscreen key is touched is more difficult. When performing
`tasks that involve sliding and dragging, the friction between the
`users’ finger and the screen provides sometactile feedback. Al-
`though this problem is not uniqueto touchscreens, it is an impor-
`tant consideration when designing touchscreen interfaces. Cur-
`rently research is being conductedto improve user performancefor
`“typing” with touchscreens(Sears, 1990; Sears, Revis, Crittenden,
`& Shneiderman, 1991; Plaisant & Sears, 1991).
`their
`Undesired touches: When using touchscreens users may rest
`hands on the screen for extra support or to reduce arm strain, or
`they may inadvertently touch the screen with anotherfinger. This
`causes touchscreen hardwareto lose track of the location users
`wish to touch. Research with touchscreens that recognize multi-
`ple touch locations may proveuseful in eliminating this problem.
`Price: Touchscreen prices are getting lower, butarestill relatively ex-
`pensive. Touchscreens range from approximately 350 to over
`1,000 dollars. This is considerably more than most mice,
`joysticks, or lightpens.
`
`Manyof these problems have either been overcome or reduced, and
`usage is steadily increasing. Many of the problems associated with
`parallax and glare have been overcome by advances in touchscreen
`hardware. Design guidelines can significantly reduce the problemsas-
`sociated with obscuring the screen, the lack of tactile feedback, and
`undesired touches. There is renewed humanfactors research into re-
`ducing fatigue that appears promising. The price of touchscreens is
`decreasing as technology improves and touchscreen useincreases. It is
`anticipated that when manufacturers begin producing monitors with
`touchscreensinstalled at the factory, the price of touchscreens should
`drop significantly.
`
`14
`
`14
`
`

`

`6
`
`SEARS, PLAISANT, AND SHNEIDERMAN
`
`the historical prejudices
`We cannot resist mentioning some of
`against the touchscreen. Many of the pioneer touchscreens did have
`severe limitations. As a result, many peoplestill picture touchscreens
`as low-precision, high-error rate input devices. Touchscreens can be
`reliably used to select relatively small targets (approximately 2 mm
`square). Touchscreens do not require a large and intrusive frame glued
`or taped on a monitor as manyearly versions did. They can be mounted
`directly onto the surface of the monitor and all supplemental hardware
`can be installed inside the monitor. In summary, touchscreens have
`improved dramatically in recent years, and, as a result, high-precision,
`low-error rate tasks can now be performed using a touchscreen. Now
`researchers can explore newstrategies and applications to guide practi-
`tioners,
`
`COMPARISON WITH OTHER INPUT DEVICES
`
`Touchscreens have been compared empirically to mice, lightpens, key-
`boards, joysticks, and other devices. The majority of human factors
`studies tested touchscreens against various devices for selecting pre-
`defined, stationary targets. Time and error rates were measured, and
`some studies measured usersatisfaction. In general, these studies have
`shownthat touchscreens are the fastest device for selecting stationary
`targets (Muratore, 1987; Ostroff & Shneiderman, 1988; Ahlstr6m & Len-
`man, 1987; Karat, McDonald, & Anderson, 1986). Unfortunately, touch-
`screens have also been shownto be the most error-prone input device
`(Muratore, 1987; Ahlstrom & Lenman, 1987). However, muchofthis
`research emphasizedrelatively large targets, and few usedalternative
`selection strategies that may improve user performance, making this
`research of limited use for higher resolution tasks, such as character
`selection or graphics input.
`A recent study (Sears & Shneiderman, 1991) compared the touch-
`screen to the mouseforthe selection of various size targets when using
`thelift-off selection strategy (the lift-off strategy will be described in the
`following section). This study showedthat usingthis selection strategy
`can result in very low error rates for the touchscreen, It also showed
`that selection of very small targets (0.4 < 0.6mm) is possible with the
`touchscreen, refuting claims that the size of the user’s finger determines
`the minimumtarget size. This study showedthat selecting targets that
`are approximately the size of a character is as fast with a touchscreen as
`with a mouse.
`Other studies have compared various input devices for selection
`tasks when users must also type on a keyboard. One study compared a
`
`15
`
`15
`
`

`

`A NEW ERA FOR HIGH PRECISION TOUCHSCREENS
`
`7
`
`touchscreen, mouse, and keyboardfor a selection task. Results indicate
`that the touchscreen waspreferred for tasks that do not involve a typing
`subtask, and the keyboard was preferred whenthe typing subtask was
`included. The touchscreen was preferred to the mousein both situa-
`tions. The touchscreen wasalso the fastest device for cursor position-
`ing in both situations (Karat et al., 1986).
`Unfortunately, there have been few comparisons of the touchscreen
`to other input devicesfor tasks other than target selection. Tasks such
`as dragging objects and outlining an object offer new opportunities for
`researchers. Informal observations indicate that tasks involving uncon-
`strained movements maybesignificantly easier with the touchscreen
`than many other devices. Of course, a stylus interface or tablet may
`prove superior for sometasks dueto their similarity to writing with a
`pen. Additional studies are needed to understandthe range of tasks for
`which touchscreens can and cannotbe used.
`Overall, touchscreens appearto be the fastest device for selecting
`relatively large targets. They can also be used for selecting smaller
`targets if the correct selection strategies are used. Error rates can also be
`reducedto a point wherethey are negligible if the correct strategies are
`used. Tasks such as dragging objects on the screen or marking the
`border of an irregular region, also appear to be very promising with
`touchscreens. The last section of this chapter contains references to
`many papers dealing with touchscreens.
`
`DESIGNING TOUCHSCREEN APPLICATIONS
`
`A Modelof User Interaction
`
`We might consider a modelof operation that divides touchscreen usage
`into seven stages. These are based on the syntactic/semantic (Shneider-
`man, 1987) and seven stages (Norman, 1988) models. Theuser:
`
`1. Formulates a plan of what needs to be done in the task domain,
`2. Examines the current computerscreen to identify all touchable and
`nontouchable areas that represent actions and objects relevant to
`the task,
`Identifies the desired touchable area by the action or object,
`3.
`4. Reachesoutto touch (the syntax is simply a touch) the desired area
`and receives feedback from hand position and from on-screen
`changes(a cursor, selectable areas inverting,etc.),
`5. Confirms that the finger is on the desired touchable area andlifts-
`off to activate,
`
`16
`
`16
`
`

`

`8
`
`SEARS, PLAISANT, AND SHNEIDERMAN
`
`6. Confirms that the desired touchable area has been activated,
`7.
`Interprets and evaluates the result of the touch in terms of whether
`the task domain goalis furthered.
`
`This model describes interaction whenthelift-off selection strategy is
`used with a touchscreen. Simple modifications can adapt this model
`for other selection strategies, or for other input devices.
`The central benefit of using a pointing device rather than a keyboard
`is reducing the syntactic load by replacing typing with pointing. The
`directness of touchscreensfurther simplifies the task by allowing users
`to simply point directly at the object or action. Instead of detailed
`instructions about whatto type (the syntax) to select an action or object,
`users simply touch a visual representation of the object or action (the
`semantic description). Touchscreens avoid the distraction of looking
`from the screen to the input device and back while remembering the
`desired syntax. If designers choose a properset of touchable objects and
`actions, then progress can be rapid with low errorrates.
`Thetraditional touchscreen strategy has been Jand-on, in which the
`first location that users’ touch initiated an action. This is acceptable,
`even preferable, when there are a few large targets. However, as the task
`domain complexity increases, the numberof choices can increase dra-
`matically and underminethe efficacy of the land-on strategy.
`With high resolution touchscreens that support continuous feed-
`back, the finger touch may producea cursorthat can be dragged across
`the screen, and activation occurs when thefingeris lifted off the sur-
`face. This can not only improvetheselection of menu items or buttons
`but also open a new world ofinteraction techniques suchasthe direct
`manipulation of metaphors (moving a cursor ona slider, selecting a
`color on a color wheel, etc.), freehand drawing, and symbolic gestural
`input. There are manyother possible strategies and combinations of
`strategies that can be used.
`The following sections will discuss various interaction techniques,
`and how each technique can be used with a touchscreen.
`
`Traditional Button or Menu Selection
`
`Menuorbutton selection tasks typically require the selection of pre-
`defined targets represented on the screen. The majority of current ap-
`plications for touchscreens involve tasks such as these. This section
`discusses factors that play important roles in button selection tasks
`including the visual representation, size, and location of targets and
`several alternative selection strategies.
`Visual representation of touchable areas. Users are constantly con-
`
`17
`
`17
`
`

`

`A NEW ERA FOR HIGH PRECISION TOUCHSCREENS
`
`9
`
`fronted with the question of what is a touchable area. A consistent
`principle or small set of principles greatly reduces the burden on users
`(Apple, 1987). Possibilities include: realistic button shapes, rounded
`rectangles, shadowed boxes, distinctive color text, distinctive color
`background,
`tabs on a book, or standard icons. Designers must re-
`memberthatit can be frustrating for users whoidentify what they think
`is a selectable object, try to select it, and discoverthatit is not selecta-
`ble after all. If instructions or icons are used, they should be chosen to
`be consistent with the tasks and users’ expectations.
`There is no simple solution for indicating whatis selectable. In some
`systems,all selectable objects appear in the same shape (Figure 2) or
`color (Figure 1). Once users learn this rule, all targets can easily be
`identified. Other systems place simple instructions on the screen, pos-
`sibly near each different target. The instructions may be as simple as
`“Touch the desired amount,” indicating whatis selectable and how to
`select it (Figure 3). The goal is to choose an option that works best for
`the tasks being performed, and to be consistent once the choice has
`been made.
`Making objects significantly different from the remainder of the
`screen can result in users ignoring the remainderof the screen. In some
`
`GOVA - Introduction
`
`PAGE 1 OF 2
`
`Welcome to GOVA,
`
`the Guide to Opportunities in Volunteer Archaeology.
`
`GOVA will tell you about archaeological projects all over the world
`
`(see Map
`
`9)
`
`that welcomes volunteers. You can use GOVA to answer
`
`practical questions like how to join a dig, or how much it costs.
`
`To Use GOVA, simply touch a blue word with your fingertip-- not your
`
`fingernail-- and release.
`
`For example,
`
`if you touch the blue word
`
`"volunteers" at the third line of this page, you will get an article
`about volunteers.
`
`To turn to the next page touch the blue word NEXT PAGE.
`
`INDEX
`
`RETURN TO MAP 9, THE WORLD
`
`NEXT PAGE
`
`RESTART
`
`Touch a highlighted word with your fingertip
`Figure 1. Hyperties, a hypertext system, allows users to traverse a database
`of articles including text, graphic and videodisc images. All selectable items
`are in bold (these appear in light blue on the actual system).
`
`18
`
`18
`
`

`

`10
`SEARS, PLAISANT, AND SHNEIDERMAN
`situations, designers may not wantto make it obvious whatis selecta-
`ble: when the designer wantsto force users to explore the entire screen,
`whenalmost everything is selectable, or when an existing image cannot
`be overlaid oraltered. In these situations, the designer can either force
`users to select what they think is selectable, or a mechanism can be
`provided to reveal all selectable regions. A special key or a “Reveal”
`button may makeall targets temporarily visible. Targets may also be
`shown whenuserstry to touch a nonselectableregion. This method can
`be both frustrating and slow if there are many selectable regions on the
`screen, making a “Reveal” button preferable whenpossible.
`As a general guideline, if users have specific goals, making targets
`obvious may speed performance. However, if the purposeis to explore
`and gain general knowledge about the system, it may be desirable to
`make the targets blend in with the remainderof the screen.
`Feedback. Feedback plays an importantrole in every user interface.
`Feedback indicates wherethe useris currently touchingthe screen, that
`an action could be taken, or that an action has been taken. Feedback
`provides confirmation to usersthat the correct actions are aboutto be or
`have been performed,as described in the modelpresented earlier.
`If users are allowed to drag their fingers before makinga selection it
`is often advantageous to provide a cursor near the users’ fingers show-
`ing exactly wherea selection will be madeif theylift their fingers. This
`is particularly important if the targets are very small, less important if
`targets are large. It
`is also important to indicate when a selection is
`possible.If the lift-off strategy (or any other strategy that uses the re-
`movalof the users’ fingers as input) is used, then as users drag their
`fingers onto a target, users should receive feedback that a selection
`could be made. This could be visual, by inverting or flashing the target,
`or audible, by making a short tone. Once a selection has been made,
`feedback should indicatethat an action is about to be taken. This confir-
`mation could be visual or audible. Visual confirmation has the advan-
`tage that the specific action to be taken could be indicated(the selected
`target could be inverted temporarily indicating exactly whataction will
`be taken). Audible feedback has the advantage that users eyes can beoff
`the targets, however, audible tones are more difficult to distinguish
`(making them less useful for indicating the exact action to be taken). If
`audible feedback is to be used, the volume must be set carefully. If the
`system is used in a public place or in an open workarea, loud tones
`may be very annoying;if the tones are too soft, they may be missed.
`User control is useful in this situation.
`Feedback about possible selections can often replace the cursor. In
`the examplein Figure 4, as users drag their fingers across the calendar,
`
`19
`
`19
`
`

`

`A NEW ERA FOR HIGH PRECISION TOUCHSCREENS
`
`11
`
`they select a day by lifting their fingers when the des