Wireless Devices for Mobile Commerce: User Interface Design and Usability
`
`Peter Tarasewich, College of Computer and Information Science, Northeastern University
`
`Appears in Mobile Commerce: Technology, Theory, and Applications, 2002,
`Brian E. Mennecke and Troy J. Strader, (Eds.) Hershey, PA: Idea Group Publishing, pp. 26-50.
`Copyright held by Idea Group Publishing
`
`
`INTRODUCTION
`
`
`An increasing number of technologies and applications have begun to focus on mobile
`computing and the wireless Web. Mobile commerce (m-commerce) encompasses all activities
`related to a (potential) commercial transaction conducted through communications networks that
`interface with wireless (or mobile) devices (Tarasewich, Nickerson, and Warkentin, 2001).
`Ultimately, researchers and developers must determine what tasks users really want to perform
`anytime from anywhere and decide how to ensure that information and functionality to support
`those tasks are readily available and easily accessible.
`A well-designed and usable interface to any application is critical. For example, properly
`designed websites help ensure that users can find information that they are looking for, perform
`transactions, spend time at the site, and return again. Given the uniqueness of the wireless
`environment, usability becomes even harder to ensure for m-commerce applications. The purpose
`of this chapter is to provide the reader with an overview of current wireless device interface
`technologies. It will provide guidance on designing usable m-commerce applications that take
`advantage of the benefits and respect the limitations of these devices. This chapter will also
`explore the interface design and usability challenges that the m-commerce environment still
`presents for users, researchers, and developers.
`This chapter is organized as follows. The first section describes the benefits and limitations
`of various wireless device interfaces. The next section looks at how the usability of wireless
`devices affects the feasibility and success of m-commerce applications. The third section
`discusses some of the additional challenges that developers face when designing applications for
`wireless devices. The final section reiterates the need for good wireless application design, and
`describes some of the safety and security issues related to wireless device interface design.
`
`WIRELESS DEVICES AND THEIR INTERFACES
`
`
`The devices currently most important to m-commerce can be classified according to the
`categories listed in Table 1. There is some feeling that devices will become completely generic,
`and take the place of items like televisions, pagers, radios, and telephones (Dertouzos, 1999), but
`the question remains as to what form the devices will ultimately take. This important issue will be
`investigated further in the section on mobile system developer issues later in the chapter. But first
`we look at the current interfaces of these devices, their strengths, and their limitations. The
`discussion is separated into input and output interactions. Research that has been performed with
`various types of interface devices will be discussed in the next section on usability.
`
`
`Table 1 – Wireless Device Categories
`Laptop Computer
`Handheld (e.g., Palm, Pocket PC, Blackberry)
`Telephone
`Hybrid (e.g. “smartphone” PDA/telephone combination)
`Wearable (e.g., jewelry, watches, clothing)
`Vehicle Mounted (in automobiles, boats, and airplanes)
`Specialty (e.g., the now defunct Modo)
`
`
`
`1
`
`SCEA Ex. 1044 Page 1
`
`

`
`Input Interaction with Wireless Devices
`
`
`Input interaction concerns the ways in which users enter data or commands. Common
`technologies used for input interaction with wireless devices include keyboards, keypads, styluses,
`buttons, cameras, microphones, and scanners. Each of these will be discussed in turn,
`emphasizing the benefits and limitations of each in the mobile environment.
`The keyboard still remains popular as a form of input for many types of computing devices.
`The QWERTY configuration of keys (named for the sequence of keys at the upper left of the
`keyboard), while not the most efficient layout possible, remains a standard because of its wide user
`acceptance. Laptop computers have carried the concept of QWERTY keyboards forward, although
`keys are usually made smaller to conserve room. Devices such as phones and handhelds,
`however, have generally foregone the integration of a full keyboard because of the desire to create
`a device that is as small and light as possible. The exception to this is the Blackberry device, which
`includes a miniature keyboard. The problem with this keyboard is that a user must adjust to smaller
`keys, oftentimes learning to type messages with both thumbs. Data entry and error rates can suffer
`with smaller keys as well.
`Smaller mobile devices usually rely on a more limited keypad for input. Most mobile phones
`use a standard 12-button numeric keypad, sometimes augmented by several special purpose keys
`(such as “clear” and “ok”). Each of the keys 2 through 9 also corresponds to a set of three or four
`letters. There are several approaches to entering text using a keypad. In the first, known as the
`multi-press input method, the user must hit a numeric key that also corresponds to the desired
`letter. For example, the letter “s” would require that the “7” key (labeled with “pqrs”) be depressed
`four times. A capital “S” would then require eight or more keystrokes. A user must also pause or
`press an additional key to move onto the next letter. A different method that uses two-key input
`requires selecting a letter’s group with the first key press and the location of the desired key with
`the second. For example, the letter ‘E’ (the second character on the “3” key which is labeled “def”)
`requires the key press sequence 3-2. Another approach uses dictionaries of words and linguistic
`models to “guess” the word intended by a series of keystrokes. For example, the sequence 8-4-3
`(corresponding to “tuv”-“ghi”-“def”) might produce the word “the” out of all possible letter
`combinations.
`One way to eliminate the use of a keypad for text entry is to attach a temporary keyboard to
`the device being used. Several vendors have developed miniature and/or full-size folding
`keyboards for this purpose. A more radically designed alternative is the Matias Half Keyboard
`(Figure 1), which contains only those keys from the left-hand side of a traditional keyboard. When
`the space bar is pressed, the same keys function as the right-hand side. Another alternative is a
`fabric keyboard, being developed by ElectroTextiles, that can be rolled up for storage (Figure 2).
`Researchers are also developing “non-keyboards” in the form of gloves (Goldstein et al., 1999) or
`“finger rings” (Fukumoto and Tonomura, 1997) that sense finger movements of users typing on a
`virtual keyboard and use software
`to
`interpret
`the movements. Essential Reality
`(www.essentialreality.com) is producing a glove called P5 that can be programmed to respond to
`users’ hand gestures with combinations of keystrokes and mouse clicks. A potential problem with
`these types of devices is the additional training time that might be needed to use the device
`effectively.
`Another way to eliminate the use of a keypad (and keyboard as well) is to use a stylus to
`write input directly on the screen of the device, a process known as gesture recognition. With this
`method, the device must recognize each character or symbol that is written, which can take a good
`deal of processing time and oftentimes suffers from inaccuracy. Palm has developed a proprietary
`system for character recognition (called Graffiti) that seems more accurate than other recognition
`systems, but forces the user to conform to a writing style for letters that is somewhat different than
`normal. Another gesture recognition technique is Jot (often used with Pocket PC devices). In both
`cases, the user must learn which pen strokes represent a particular character to the device, rather
`than the device interpreting the handwriting of the user. As an alternative to keypads, Smart
`
`
`
`2
`
`SCEA Ex. 1044 Page 2
`
`

`
`Design (www.smartdesign.com) is developing a system called Thumbscript that replaces a keypad
`on phones with a nine-point grid. Users tap a keystroke sequence on the grid for each character
`(Roman letter or Asian character) that they wish to input.
`As an alternative to gesture recognition, keyboards (or other key configurations) can be
`created virtually on a screen, with each key being “pushed” by touching it with a stylus. These so-
`called “soft-keyboards” are sometimes implemented in sections (e.g., the alphabetic characters
`separated from numbers and other characters) to save screen space and create larger keys.
`Styluses can also be used to activate icons, menu choices, or hyperlinks displayed on a screen.
`Virtual keyboards currently suffer from a lack of tactile feedback often found on keyboards and
`some keypads, although feedback can be provided through sounds generated as keys are
`“pressed.”
`Mobile device input can also be achieved through “mouse buttons,” thumbwheels, and
`other special-purpose buttons. The user interface of the telematics system OnStar consists of just
`three buttons, labeled “call”, “help”, and “off”. Mobile phones often have dedicated buttons with
`labels such as “call”, “ok”, and “clear” in addition to a numeric keypad. Mouse buttons are toggle
`switches that allow one-dimensional cursor movement. An alternative to a mouse button is the
`“navi-roller,” which allows scrolling by rolling and selection by clicking. Small joysticks, which allow
`two-dimensional cursor movement, are sometimes found integrated into the keyboards of laptop
`computers, and more recently on mobile phones. Handheld devices usually have a mouse button
`and a few other special-purpose buttons, but no keyboard or keypad. CyMouse by Maui Innovative
`Peripherals (maui-innovative.com) is an eight-ounce headset that acts as a wireless mouse. A
`version called Miracle Mouse is aimed at providing more control options to people with physical
`disabilities. The now defunct Modo device (Figure 3), which featured one-handed operation, had a
`thumbwheel to move between selections and to scroll text up and down. Pressing the wheel
`activated the current selection. Some other handheld devices also feature a similar built-in
`thumbwheel. However, the location of the thumbwheel limits which hand can hold the device for
`one-handed operation.
`Using human speech as input to mobile devices is also becoming increasingly practical as
`voice recognition technology continues to improve. Whether or not voice interfaces will ultimately
`succeed as a primary form of input depends on how well certain limitations of the technology can
`be overcome. These limitations include the need to train devices to recognize a user’s voice, the
`relative slowness of voice versus other input means, and the difficulty in using visual information
`(e.g., graphics) with voice input. Benefits of voice input include the ability of users to interact with
`the device in their natural language. Voice input allows those users who cannot type or use a
`stylus to interact with a device. It may also be a viable interface alternative for devices too small for
`buttons or for those without a screen. However, voice input suffers from possible privacy and social
`issues. For example, users may feel uncomfortable speaking input aloud instead of typing or
`writing it, and certain places (e.g., libraries) might restrict the use of voice input to maintain a quiet
`environment. One option that allows a voice interface with mobile devices, but does not require
`direct Internet access from the mobile device, is Voice Extensible Markup Language (VXML). This
`standard allows consistent access to Web applications from both the wired and wireless
`environments.
`With the shrinking size of camera lenses and the increasing sophistication of digital
`photography, video is becoming more common as a form of input with mobile devices. Some
`laptops, phones, and handheld devices have built-in or attachable cameras. DoCoMo has been
`developing specialized mobile Internet appliances, some of which are cameras that can take
`pictures, adorn them with overlays, and send them to users with similar devices or i-mode phones.
`Video might also be used as input through the recognition of hand gestures or facial expressions.
`Similarly, scanners may also become part of the wireless environment. They can be used
`for reading text, bar codes, or other symbols. Wireless devices that scan UPC symbols as input
`could be part of in-store mobile commerce applications used for comparison-shopping or for
`purchasing merchandise without the need of a cash register and sales attendant.
`
`
`
`3
`
`SCEA Ex. 1044 Page 3
`
`

`
`Finally, input can come from technologies that sense location, or from those that can
`receive information from their environment based on their location. The Global Positioning System
`(GPS), a set of satellites owned and operated by the U.S. Department of Defense, allows any
`device equipped with a GPS receiver to determine its geographic location within about 10 meters.
`All mobile phones sold in the U.S. will be required to have the ability to determine their location.
`Bluetooth technology, which allows short-range communications, will allow mobile devices to
`receive information automatically when they are in close proximity of another Bluetooth-equipped
`device. As we will discuss later, location is a key factor in designing useable mobile applications.
`However, privacy issues dealing with the use of location data must also be addressed.
`
`Output Interaction with Wireless Devices
`
`
`Output interaction concerns the ways in which users receive data, prompts, or the results of
`a command. Common technologies used for output interaction with wireless devices include video
`screens and speakers. Both of these will be discussed in turn, emphasizing the benefits and
`limitations of each in the mobile environment.
`The liquid crystal display (LCD) screen is the primary technology used to produce output in
`the form of images and text on current wireless devices. Screen size varies greatly from one type
`of device to another. Most mobile phones have small (1” to 2” square) screens that can display 4 to
`8 lines of 10 to 20 alphanumeric characters each. Handheld devices have relatively larger screens
`(about 3” by 4”) that are more suitable for graphics as well as text, but are still limited by low screen
`resolutions (usually 240 by 320 pixels). Most phones and handhelds have monochrome screens,
`although more are being sold with color screens, which can increase device usability. Laptops
`have fairly large color screens (up to 15” diagonal) with resolutions that compare favorably to
`desktop monitors. Vehicle-mounted devices have screens ranging from smaller than the size found
`on phones to the size found on small laptops, depending on the intended purpose of the device
`(e.g., displaying song titles versus a map of a city).
`The current limitations of screens on wireless devices are their size, resolution, and color
`capabilities, all of which are usually less than those found on desktop computers. These limitations
`make it difficult to display large amounts of text and graphic-based output (e.g., maps, charts, or
`Web pages). There are also tradeoffs in improving the screen characteristics of mobile devices.
`Increasing screen size will increase the size and weight of a device. Color screens with high
`resolutions use more power than their monochrome counterparts, resulting in increased battery
`weight and/or less time before the battery needs to be recharged (although research into better
`batteries continues).
`There are, however, some recent technological developments that may address some of
`the disadvantages of current wireless device screens. Flexible screens are on the horizon, which
`may eventually allow screens that can be rolled or folded up. E Ink (www.eink.com) and Gyricon
`Media (www.gyriconmedia.com) are developing displays with electronic ink technology (e-paper),
`first in black and white, but possibly in color in the future. The screens hold an image until voltage
`is applied to produce a new image, using less overall power than LCD screens.
`Monocular units or goggles can be used with magnifying glasses to enlarge small displays
`(less than an inch diagonal) so that they look like an 800 x 600 resolution monitor. Goggle-type
`products
`include
`InViso’s
`eShade
`(www.inviso.com/products),
`Sony’s Glasstron
`(www.ita.sel.sony.com/products/av/glasstron),
`and Olympus’
`Eye-Trek
`(www.eye-trek-
`olympus.com, see Figure 4). Microvision (www.mvis.com) is developing a device that projects an
`image, pixel by pixel, directly onto the viewer’s retina. Heads-up displays, which have seen limited
`use in automobiles in the past, might also be used for vehicle-mounted devices. These types of
`devices allow viewing of color images with similar sizes and resolutions as those found on desktop
`computers. Potential concerns with these technologies include interference with users’ other visual
`inputs, and the social acceptance of wearing and using such technologies.
`
`
`
`4
`
`SCEA Ex. 1044 Page 4
`
`

`
`Sound is the other primary form of output from a wireless device. Forms of this output range
`from words to music to various beeps, buzzes, and other noises. These can be created through
`speakers or through headphones. Newer laptops usually have a set of speakers built in for stereo
`sound production. Most smaller mobile devices have a single speaker at best. Stereo speakers can
`be used to generate sounds coming from a particular direction, which as we shall see later can be
`used to enhance usability. This same effect can be achieved through headphones, but at the cost
`of possible interference with a user’s other audio input (i.e., sounds from the environment).
`Sound output may be a viable interface alternative for devices without a screen, although there
`may be difficulties in presenting certain visual information (e.g., graphics). Voice output is also
`generally produced and comprehended slower than visual output. On the positive side, sound
`allows those users who cannot see a screen to receive output. Ultimately, it may be that multi-
`modal browsing, where voice and visual output are combined, may be best suited for wireless
`devices (Nah and Davis, 2001).
`
`WIRELESS DEVICE USABILITY
`
`
`This section looks at the usability of wireless devices and how usability affects the feasibility
`and success of m-commerce applications. Some of the recent research on interface design and
`usability for mobile and wireless devices will be discussed, along with usability issues present with
`wireless devices. The section will also consider whether or not current HCI standards can be
`applied to wireless devices, and what further research issues regarding the usability of wireless
`devices need to be addressed.
`Usability can be defined as the quality of a system with respect to ease of learning, ease of
`use, and user satisfaction (Rosson and Carroll, 2002). It also deals with the potential of a system to
`accomplish the goals of the user. Usability testing asks users to perform certain tasks with a device
`and application while recording measures such as task time, error rate, and the user's perception
`of the experience. Methods for evaluating usability include empirical testing, heuristic evaluations,
`cognitive walkthroughs, and analytic methods such as GOMS (goals, operators, methods, and
`selection rules).
`Many of these same usability methods can be applied successfully in order to test the
`usability of a particular application on a device, or compare usability across different devices or
`configurations. Chan and Fang (2001) reported on research in progress that is conducting a
`heuristic evaluation and cognitive walkthrough of fifteen m-commerce sites across three different
`device platforms (Palm, Pocket PC, and WAP phone). Their preliminary results indicate that many
`Web sites are trying to duplicate their wired Web architecture and design for the wireless Web,
`resulting in poor navigation and information overload.
`Likewise, many of the current principles of interface design can be transferred to newer
`devices, although soundly applying these principles may be more difficult due to the unique nature
`of mobile systems and devices. Fundamental rules such as consistency, shortcuts for advanced
`users, the use of feedback, error prevention, easy reversal of actions, and minimization of short-
`term memory requirements (Shneiderman, 1998) will undoubtedly transfer to mobile applications.
`However, as shown in the previous section, the devices that the user might interact with are quite
`different than the desktop computers used in much of the interface design research to date. While
`further study is needed, it is likely that much of the specific research on effective screen design and
`information output cannot be generalized to mobile devices.
`Furthermore, context will factor heavily into the use of mobile applications and devices,
`which is something that was not as much (if any) of a concern with stationary desktop applications.
`Mobile tasks and technology use are significantly different than their stationary counterparts.
`People can now literally be anywhere at anytime and use a mobile application, which was not true
`with the traditional (wired) Web since a physical connection was needed to the Internet. Location
`will need to be factored into the usability of an application and a device, as will the dynamic nature
`of the environment within which it is used. Conceivably, a mother could be walking down a street in
`
`
`
`5
`
`SCEA Ex. 1044 Page 5
`
`

`
`an unfamiliar city trying to use a mobile application to find the location of an office for an
`appointment, while keeping track of her three children and processing all the other input coming
`from her environment. Interface design that may be well suited to a relatively stable office or home
`environment will not necessarily work well in the Amazon rain forest or in an automobile cruising
`down a highway.
`Let us now turn to some of the recent research that specifically addresses the design and
`usability of mobile applications and devices, first from the viewpoint of input interaction. One
`usability concern is how well users can perform tasks using the assortment of keypads and
`keyboards found on many wireless devices. Looking at keypad text entry performance, Silfverberg,
`MacKenzie, and Korhonsen (2000) created models to predict the entry rates for multi-press, two-
`key, and linguistic-based keypad text entry methods. Using empirical data, they estimated that
`expert users could achieve rates of up to 27 words per minute (wpm) using thumb (one-handed) or
`index-finger (two-handed) input with the multi-press and two-key methods. For the particular
`linguistic-based method that they investigated, they predicted speeds up to 46 wpm for expert
`users using two hands and their index finger. A study done by Weiss, Kevil, and Martin (2001) on a
`particular mobile phone found that users in general had difficulty in using its keypad. Some user
`frustration came from confusion as to which keys performed what functions, and how the keys
`were labeled. All subjects had difficulty in entering text. Difficulties in navigating through
`applications were also encountered, in part due to use of the keypad and in part due to the
`confusing structure of the applications tested.
`There have been many studies on soft keyboard performance. Those by Lewis, LaLomia,
`and Kennedy (1999) and MacKenzie and Zhang (1999) found that users could achieve speeds of
`up to 40 words per minute with a QWERTY layout on a soft keyboard, although speed varied with
`the devices used, the tasks performed, and the amount of practice. Alternate soft keyboard layouts
`can produce even higher text entry speeds than the QWERTY configuration, but usually after much
`experience with the alternate layout (e.g., MacKenzie and Zhang (1999)). A study by Zha and
`Sears (2001) showed that the size of a PDA soft keyboard did not affect data entry or error rates.
`Additional subjective ratings did not suggest that users preferred larger keyboards, which implies
`that soft keyboards could be successfully implemented on smaller devices, such as mobile phones.
`Looking at virtual keyboards, a study by Goldstein et al. (1999) found that their non-
`keyboard (i.e. glove) device resulted in fewer errors and higher subjective satisfaction than a soft
`keyboard and a miniature keyboard on mobile devices (although a full-size keyboard was still the
`most preferred). The Fukumoto and Tonomura (1997) FingerRing device was tested only with
`users producing chords (symbols) rather than individual characters on a QWERTY keyboard, so
`there is no way to compare use of their device to other keyboard types.
`If a stylus is used to write input on the screen of a mobile device (using gesture or
`handwriting recognition), performance is generally much poorer compared to using any type of
`keyboard. Studies such as MacKenzie and Chang (1999) found that data entry rates of up to 18
`words per minute (wpm) can be achieved using various gesture recognition systems. But these
`studies did not test performance using handheld devices. An exception to this is Lewis (1999),
`which reported speeds of up to 24 wpm on PDAs, but used simulated “perfect” handwriting
`recognition where any attempt at creating a letter was considered correct. More recently, Sears
`and Arora (2001) compared Jot and Graffiti using Pocket PC and Palm devices, respectively. They
`used tasks that they felt were more realistic than previous studies, and kept track of data entry
`times and error rates. Novice data entry rates of 7.37 wpm were obtained for Jot and 4.95 wpm for
`Graffiti. The recognition of “gestures” also covers stylus-made marks other than letters or numbers
`used for data input or commands. A survey of handheld device users completed by Long, Landay,
`and Rowe (1997) showed that users generally liked using gestures for device input, although they
`often found them difficult to remember and became frustrated when a device did not recognize
`what they wrote. More recent research such as Long, et al. (2000) looked at designing gestures
`that are easier for people to use and remember.
`
`
`
`6
`
`SCEA Ex. 1044 Page 6
`
`

`
`There is also research that looks into assisting the user with the data input or command
`process. Dunlop and Crossan (1999) proposed a text entry method for mobile phones that
`anticipates words based on a dictionary of common words stored on the device. The method was
`tested using a PC-based emulation of a mobile phone. Results showed some success with and a
`general user preference for the new method, although more testing needs to be done. Masui
`(1999) developed a dictionary-based text entry method that uses the context of the phrase or
`document being typed. Given the current input limitations of mobile devices, usability might also be
`increased by changing the nature of the data or instructions required by the application. Versign is
`introducing a service called WebNum, which would substitute a telephone number or other numeric
`string for a standard Web address (e.g., www.neu.edu). Testing still needs to be completed on this
`method as well.
`Voice recognition technology continues to improve, but there is still the question of how well
`it works for different applications and tasks. De Vet and Buil (1999) listed some general findings
`from user studies on the use of voice control compared to entering text data on limited-key
`devices. User operations that favor voice control included 1) direct addressing of content (e.g.,
`calling out someone’s name), 2) menu navigation and option selection, and 3) setting a range (e.g.,
`the starting and stopping times on a VCR). The operation of scrolling through a long list favored the
`use of cursor keys rather than voice commands for people who were browsing.
`Now we look at some of the research concerning design and usability related to output
`interaction. Output technology has received a fair amount of attention from researchers, with much
`of the recent focus on small displays. The fundamental question here is, can users perform tasks
`as well using small displays rather than larger ones? This answer to this will, of course, vary based
`on the size of the display and the task being performed. A study by Jones et al. (1999) found that
`users in a “small screen” environment (simulated by setting monitor resolution to 640x480 pixels)
`were less effective in completing search and retrieval tasks than users with a “large screen”
`environment (1074x768 pixels).
`Reading text on small devices, especially the size found on many mobile phones, can be
`difficult. There are various options that can be considered for formatting text on small screens and
`providing navigation. Melchior (2001) developed a method called “wiping” that may make it easier
`for people to read text on small displays. The method adds a perceptual guide (the graying of text
`that will be removed from the screen) during scrolling that aids in refocusing the user’s attention
`after the paging of text. A study on wireless application protocol (WAP) interface usability was done
`by Chittaro and Cin (2001) using novice users. Each screen of material on a WAP device is known
`as a card. They evaluated 1) navigation among cards using links versus an action screen, and 2)
`single-choice lists using a list of links versus a selection screen. Results showed that users
`performed better using links and a list of links, and perceived greater difficulty in using the action
`screen and selection screen environments.
`Rapid serial visual presentation (RSVP), which serially presents one or more words at a
`time at a fixed place on a screen, is another option for presenting text on a small screen. There are
`many studies that investigate the use of RSVP, but overall the results seem inconclusive as to
`whether the method works better for text presentation than other methods. Bernard, Chaparro, and
`Russell (2000) compared RSVP against presenting three lines of text at a time and ten lines at
`time on a simulated small-screen interface. Overall reading comprehension levels were about the
`same for the RSVP and ten-line methods, which were marginally higher than the three-line
`method’s comprehension levels. Subjects were equally satisfied with each method of presentation,
`and did not seem to prefer one method to the others. However, they did prefer a slower text speed
`and thought that the RSVP method produced more eyestrain. Studies comparing RSVP to
`sentence-by-sentence presentation were performed by Rahman and Muter (1999). They
`concluded that RSVP was not liked by subjects but is as efficient (as measured by reading speed
`and reading comprehension) as sentence-by-sentence and full-page presentations.
`Variations of RSVP are also being investigated for use on small screen devices, and may
`provide better presentation alternatives. Adaptive RSVP allows the exposure time for each word or
`
`
`
`7
`
`SCEA Ex. 1044 Page 7
`
`

`
`group of words to vary, based on word length and familiarity. Sonified RSVP attaches appropriate
`sounds (such as earcons) to groups of text. Details on the development of these two concepts can
`be found in Goldstein, et al. (2001). The concept of RSVP has also been applied to Web browsing
`on small screen devices (De Bruin, Spence, and Chong, 2001). The idea behind this concept is to
`rapidly display navigation choices sequentially when space is limited, allowing users to see the
`range of alternatives (links) available without a lot of searching. Initial testing of an RSVP browser
`against a WAP browser showed RSVP browsing to be at least as effective as WAP browsing for
`experienced users.
`Other types of browsers for small screen devices are being developed and tested as well,
`all hoping to increase the usability and effectiveness of mobile devices for Web-based tasks. When
`viewing a Web page on a small screen, most current browsers show a subset of the original page
`(usually with minimal graphics) after processing it through a proxy server. An application called
`Power Browser was developed and tested by Buyukkokten, et al. (2000) against various other
`handheld device Web browsers. Their method presented Web pages as text-only summary views
`based on info