US20110201387A1

US20110201387A1 - Real-time typing assistance

Info

Publication number: US20110201387A1
Application number: US12/753,744
Authority: US
Inventors: Timothy S. Paek; Itai Almog; Eric Norman Badger; Tirthankar Sengupta; Shawna Julie Davis; Matthew J. Bennett; Bryan W. Nealer
Original assignee: Microsoft Corp
Current assignee: Microsoft Technology Licensing LLC
Priority date: 2010-02-12
Filing date: 2010-04-02
Publication date: 2011-08-18
Also published as: US20110202836A1; US10698604B2; US9165257B2; US10156981B2; US10126936B2; US20140310213A1; CN102141889A; US20160103812A1; US8782556B2; US20170206002A1; US20190012076A1; CN102141889B; US20190087084A1; US20110202876A1; US9613015B2

Abstract

An apparatus and method are disclosed for providing feedback and guidance to touch screen device users to improve the text entry user experience and performance through the use of indicators such as feedback semaphores. Also disclosed are suggestion candidates, which allow a user to quickly select next words to add to text input data, or replacement words for words that have been designated as incorrect. According to one embodiment, a method comprises receiving text input data, providing an indicator for possible correction of the text input data, displaying suggestion candidates associated with alternative words for the data, receiving a single touch screen input selecting one of the suggestion candidates, and modifying the input data using the word associated with the selected suggestion candidate.

Description

CROSS-REFERENCE TO RELATED APPLICATIONS

This application claims the benefit of U.S. Provisional Application No. 61/304,341, filed Feb. 12, 2010, and entitled “TYPING ASSISTANCE FOR EDITING,” which is hereby incorporated herein by reference in its entirety.

FIELD

The present disclosure pertains to devices and methods for enhancing text entry using a touch screen device.

BACKGROUND

With the increasing popularity of mobile devices, including cellphone devices, handheld devices, handheld computers, smartphones, PDAs, etc., there is a need for improving the user interface experience by increasing user text input speed, reducing text entry errors, and improving the overall user experience.
Mobile devices with capacitive or resistive touch capabilities often utilize a touch screen keyboard, a hardware keyboard, speech recognition, handwriting recognition, or combination of the four, for entry of text input. Touch screen keyboards enable larger displays for videos, web pages, email, etc., without the requirement of a physical keyboard. Because touch screen keyboards are software-based, they can be easily adjusted for different languages, screen orientation, and key layouts. Furthermore, touch screen keyboards can be augmented with widgets for word prediction and disambiguation candidates.
Users of devices with touch screens, especially mobile devices, have varying abilities and styles of entering text. In particular, some users prefer to type large chunks of text input fairly rapidly, and do not to review and correct the entered text until complete phrases, sentences, or complete messages have been entered. Similarly, users entering text using speech recognition or handwriting recognition do not want to stop to review their text input until having entered completed phrases, sentences, or complete messages. Predictive typing assistance software such as T9 only offer word prediction candidates as users type. After users finish typing, they are usually left without any assistance, and must then struggle to edit text by placing cursors in-between characters in order to proof and correct text.
Therefore, there exists ample opportunity for improvement in technologies related to facilitating user input on electronic devices by providing more helpful and accurate assistance in the text correction process in order to accelerate user text entry and reduce user input error rates.

SUMMARY

An apparatus and method are disclosed for providing feedback and guidance to touch screen device users to improve the text entry user experience and performance.
One exemplary embodiment disclosed herein is a method comprising receiving first input data comprising one or more input words from a keyboard, providing one or more feedback semaphores, wherein the feedback semaphores are operable to alert a user using the keyboard that one or more suggestion candidates are available, automatically displaying the one or more suggestion candidates, wherein the suggestion candidates are each associated with an alternative word for one or more of the input words, automatically receiving a touch screen input selecting one of the suggestion candidates, and modifying the first input data using the alternative word associated with the selected suggestion candidate. In some examples, one or more alternative words for one of the input words are generated, the first input data is immediately auto-corrected are providing the feedback semaphores by automatically replacing the input word with one of the alternative words. In some examples, the first input data comprises a character designated as a delimiter and the providing feedback semaphores occurs based on a word probability for the input word and receiving the delimiter character. In some examples, the feedback semaphores can include one or more of the following: playing a sound, generating haptic feedback, highlighting one or more keys on the keyboard, highlighting a background area of the keyboard, and highlighting a suggestion candidates area. In some examples, the feedback semaphores include highlighting a delimiter key and receiving a single touch screen input for one of the highlighted delimiter keys. In some examples, the feedback semaphores are not provided until the input word is designated incorrect using a candidate source. In some examples, the feedback semaphores provided can be selected based on receiving typing speed data for the first input data. In some examples, the feedback semaphores provided can be selected based on detecting whether the user is typing on the keyboard with one or two hands.
Another exemplary embodiment disclosed herein is a computer-readable storage media storing computer-readable instruction that when executed by a computer cause the computer to perform a method comprising: receiving text input data comprising at least one word using a keyboard coupled to a touch screen, detecting that an event has occurred based on the text input data and based on the detected event, automatically displaying on the touch screen one or more suggestion candidates for the at least one word based on the text input data and one or more candidate sources, where each of the candidates is associated with at least one next word designated as likely to follow the at least one word, receiving a single touch screen input selecting one of the suggestion candidates, and, based on the single touch screen input, automatically modifying the text input data by adding the at least one next word associated with the selected suggestion candidate to the text input data. In some examples, the event is detected based on a measured keystroke latency. In some examples, the event is an auto-correction of the text input data. In some examples, the event is receiving a delimiter key in the text input data. In some examples, a method further comprises automatically displaying on the touch screen one or more replacement suggestion candidates for the at least one word based on the text input data and one or more candidate sources, where each of the replacement suggestion candidates is associated with at least one replacement word for the at least one word, and the displaying includes displaying the replacement suggestion candidates in a manner that distinguishes the replacement suggestion candidates from the suggestion candidates. In some examples, a method comprises immediately after the automatically adding the at least one word, displaying one or more suggestion candidates for the text input data, which includes the selected next word, and repeating the automatically modifying the text input data.
A further exemplary embodiment disclosed herein is a mobile device comprising one or more processing units operable to execute computer-executable instructions for text entry and correction, one or more memory units coupled to the processing units, one or more touch screens coupled to the mobile device configurable to have a text display area, a suggestion candidates area, and a touch screen keyboard area, wherein the text display area, the suggestion candidates area, and the touch screen keyboard area occupy distinct, non-overlapping areas of the touch screens, and wherein the one or more touch screens are operable to receive touch input over at least a portion of the touch screen keyboard area and the suggestion candidates area. A mobile device further comprises storage for storing the computer-executable instructions for text entry and correction using a text input module for receiving text input using the touch screen keyboard and displaying at least a portion of the text input in the text display area a candidate generation module for generating one or more suggestion candidates comprising one or more replacement suggestions and/or one or more next word suggestions for an input word of the text input an indicator generation module for producing one or more indicators that can notify a user that the suggestion candidates are available, a suggestion presentation module for presenting the suggestion candidates associated with the input word in the suggestion candidates area and receiving touch screen user input for selecting one of the presented suggestions wherein if the selected suggestion is a replacement suggestion, replacing the input word with the selected suggestion, and if the selected suggestion is a next word suggestion, adding the selected suggestion as a next word subsequent to the input word in the text input. In some examples, the indicator generation module and the suggestion presentation module are configured to not produce indicators and to not present suggestion candidates, respectively, until one of the input words is designated incorrect by a dictionary, a common speller application programming interface, or an input history data source. In some examples, the indicator generation module and the suggestion presentation module are configured to not produce indicators and to not present suggestion candidates, respectively, until a delimiter character is received immediately after the input word.
The described techniques and tools for solutions for improving text entry user experience and performance can be implemented separately, or in various combinations with each other. As will be described more fully below, the described techniques and tools can be implemented on hardware that includes software touch screen keyboards or hardware keyboards. As will be readily apparent to one of ordinary skill in the art, the disclosed technology can be implemented using various platforms coupled with a touch screen including, but not limited to, mobile devices (cellphones, smartphones, PDAs, handheld devices, handheld computers, PDAs, touch screen tablet devices), tablet or laptop computers, desktop computers, and home theater systems. As used herein, a touch screen includes a display coupled with touch sense capabilities (for example, displays using capacitive or resistive sensors).
The foregoing and other objects, features, and advantages will become more apparent from the following detailed description, which proceeds with reference to the accompanying figures.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG. 1 is a system diagram depicting an exemplary mobile device including a variety of optional hardware and software components.

FIG. 2 illustrates a generalized example of a suitable computing environment in which described embodiments, techniques, and technologies may be implemented.

FIG. 3 illustrates a generalized example of a suitable implementation environment for a text entry device connected to a computing cloud.

FIG. 4 illustrates a generalized example of a text entry device having a touch screen including a touch screen keyboard, a suggestion candidates area, and a text entry area.

FIGS. 5A-5C illustrate generalized examples of text entry devices, including a touch screen device with a touch screen keyboard, a mobile device, and a personal computer.

FIG. 6 is a flow chart of an exemplary implementation of a method for generating possible corrections for input data, providing indicators including feedback semaphores that indicate the availability of the possible corrections, and displaying suggestion candidates.

FIGS. 7A-7E depict a method of providing unexpected-key feedback, including providing feedback semaphores, when an unexpected key is pressed.

FIGS. 7F-7H depict three alternative designs for providing feedback semaphores.

FIG. 8 is a flow chart that outlines an exemplary implementation of the method shown in FIGS. 7A-7E.

FIGS. 9A-9D depict an exemplary implementation of a method of providing auto-correction notification, including providing feedback semaphores, when auto-correction of text input data occurs.

FIG. 10 is a flow chart that outlines an exemplary implementation of the method shown in FIGS. 9A-9D.

FIGS. 11A-11C depict an exemplary implementation of a method of providing likely next phrases, including using suggestion candidates after receiving a text input word and a delimiter.

FIGS. 11D-11F depict an exemplary implementation of a method of providing likely next words or phrases, including using suggestion candidates after receiving a text input word and a delimiter, and then providing an additional next word.

FIG. 12 is a flow chart that outlines an exemplary implementation of the methods shown in FIGS. 11A-11C and 11D-11F.

FIG. 13 is a chart representing events that occur during an exemplary implementation of a method that includes playing random click sounds as a user presses keys on a touch screen keyboard, as well as the use of keystroke latency to determine when to present suggestion candidates.

FIG. 14 is a flow chart that outlines an exemplary implementation of the methods shown in FIG. 13.

FIG. 15 illustrates a generalized example of a suitable implementation environment including a computing cloud and various connected devices in which described embodiments, techniques, and technologies can be implemented.

FIGS. 16A-16D are charts depicting some experimental results obtained using the disclosed technologies.

DETAILED DESCRIPTION

I. General Considerations

This disclosure is set forth in the context of representative embodiments that are not intended to be limiting in any way.
As used in this application and in the claims, the singular forms “a,” “an,” and “the” include the plural forms unless the context clearly dictates otherwise. Additionally, the term “includes” means “comprises.” Further, the term “coupled” encompasses mechanical, electrical, as well as other practical ways of coupling or linking items together, and does not exclude the presence of intermediate elements between the coupled items.
The described things and methods described herein should not be construed as being limiting in any way. Instead, this disclosure is directed toward all novel and non-obvious features and aspects of the various disclosed embodiments, alone and in various combinations and sub-combinations with one another. The disclosed systems, methods, and apparatus are not limited to any specific aspect or feature or combinations thereof, nor do the disclosed things and methods require that any one or more specific advantages be present or problems be solved.
Although the operations of some of the disclosed methods are described in a particular, sequential order for convenient presentation, it should be understood that this manner of description encompasses rearrangement, unless a particular ordering is required by specific language set forth below. For example, operations described sequentially may in some cases be rearranged or performed concurrently. Moreover, for the sake of simplicity, the attached figures may not show the various ways in which the disclosed things and methods can be used in conjunction with other things and methods. Additionally, the description sometimes uses terms like “produce,” “generate,” “select,” “highlight,” and “provide” to describe the disclosed methods. These terms are high-level abstractions of the actual operations that are performed. The actual operations that correspond to these terms will vary depending on the particular implementation and are readily discernible by one of ordinary skill in the art.
Theories of operation, scientific principles or other theoretical descriptions presented herein in reference to the apparatus or methods of this disclosure have been provided for the purposes of better understanding and are not intended to be limiting in scope. The apparatus and methods in the appended claims are not limited to those apparatus and methods that function in the manner described by such theories of operation.
In the following description, certain terms may be used such as “up,” “down,” “upper,” “lower,” “horizontal,” “vertical,” “left,” “right,” “over,” “on,” “near,” and the like. These terms are used, where applicable, to provide some clarity of description when dealing with relative relationships. But, these terms are not intended to imply absolute relationships, positions, and/or orientations.
As used in this disclosure, the term “wait” may be used to describe the action a device takes while waiting for particular value or type of input before proceeding with a particular operation. This waiting should not be construed as limiting the device to only waiting for the particular type of input, rather, the device may receive other input or perform other actions concurrently with the waiting.
As used in this disclosure, the term “automatically” is used to describe actions that can proceed immediately, without receiving further user input. As used in this disclosure, the term “immediately” means that an action occurs within a short time period following a preceding action without needing to receive intervening user input. In some cases, there may be intervening actions performed between or concurrently with the preceding action and the action occurring “immediately,” for example, screen refresh or redraw, sound playback, etc.
As used in this disclosure, the term “incorrect” is used to describe a designation of a word or phrase as being incorrect. A word designated as incorrect can be automatically highlighted or auto-corrected, even though the word designated as incorrect by a correction module might actually be considered to be correct by the user. For example, a word can be designated as incorrect because it does not exist in a dictionary, CSAPI (common speller application programming interface), or IHDS (input history data source). Alternatively, a word can be designated as “incorrect” even though it exists in a dictionary, CSAPI, or IHDS, because of other checking rules implemented in a correction module or candidate generation module, or because of the context of the word within a phrase.
As used in this disclosure, the term “indicator” is used to describe output that is intended to capture the attention of a user of a text entry device, and can include visual, audio, and haptic feedback. Some examples techniques for providing visual indicators include highlighting an object on a display by changing the color of the object, shading of the object or background of an object, changing the color of one or more elements of an object (e.g., by coloring the letter “F” on the “F” key of a touch screen keyboard), flashing the display, or by drawing a new object (e.g., using underlines, balloons, or suggestion candidates) on the display. Some examples of audio indicators include playing back a pre-recorded or synthesized sound, including a randomly selected sound. Some examples of haptic indicators include vibrating a handheld device or area of a touch screen keyboard, as well as increasing or decreasing the amount of pressure sufficient to register a key press on a keyboard, including touch screen keyboards. An example of varying key press resistance to prevent typing errors may be found in Hoffmann, et al., “TypeRight: A Keyboard with Tactile Error Prevention,” Proceedings of the 27th Int'l Conf. on Human Factors in Computing Systems, pp. 2265-2268 (2009).
As used in this disclosure, the term “semaphore” is used to describe the use of one or more indicators to provide feedback and/or guidance signals on a text entry device user to help discover and avoid text entry errors. Generally speaking, “feedback semaphore” is used to describe an indicator or signal that is provided after an event is received (for example, an event could be an auto-correction or other action that a text entry device performs, or the receiving of a key press or touch screen touch input). In contrast, “guidance signal” is used to describe an indicator that is provided before the occurrence of an event. For example, a text entry device can predict which key a user is likely to press next and highlight the most likely key, thereby providing a guidance signal, so the user can more easily find the most likely key.
As used in this disclosure, the term “over” is used to describe the positioning of one or more objects (for example, a finger, thumb, or stylus) over, on, or near a location on a touch screen. In some embodiments, this object need not come into contact with the touch screen for the object's position to be determined. In other embodiments, the object described as “over” the touch screen may be in contact with the surface of the touch screen. In some embodiments, the object determined to be “over” a location of the touch screen may not actually be positioned directly over the touch screen location, but determined to be “over” the location on the touch screen, for example, by a position correction module of the text entry device or touch screen.
The disclosed technology includes various approaches to improving typing accuracy or typing speed when using devices having a touch screen by using suggestion candidates to augment other input devices. These suggestion candidates are typically represented in a candidates area, which need not be permanently reserved for that purpose, or can appear in varying location on the touch screen. After entering one or more words to form a text entry, the user can review the text entry by viewing the touch screen and deciding whether to select word(s) for “suggestions.” Although some examples disclosed herein describe “a word” or “a selected word,” it should be understood that in some examples, selecting a word can include but is not limited to selecting a single word with a single touch screen input, selecting multiple words of a phrase with a single touch screen input, or selecting multiple words of a phrase using touch screen input comprising plural single touch screen inputs. For example, auto-correction or unexpected-key feedback can be generated for a single word, or for a phrase comprising multiple words and spaces, but that are related in some way.
In some examples, after an indicator is provided, one or more suggestion candidates are displayed on a touch screen display. The suggestion candidates can be presented as “buttons” which include a word related to the word selected by the user. Suggestion candidates can be determined to be related to the selected word using a candidate generation module, which can use a dictionary, thesaurus, common speller application programming interface, input history data source, or other sources or methods to generate suggestion candidates. The candidate generation module can also determine the rank order in which suggestion candidates are presented. For example, the suggestion candidates can be presented from left to right, with the suggestion candidate determined to be the most likely presented furthest to the left, and the least likely suggestion candidate presented furthest to the right. The user reviews the suggestion candidates, and selects one of the candidates for replacement using a single touch screen input over the desired suggestion candidate on the touch screen.
As used in this disclosure, a single touch screen input refers to the input received when a user positions an object over the surface of a touch screen such that the touch screen device can determine the position of the object. In some embodiments, the object can be the user's finger or thumb. In other embodiments, the object can be a stylus or puck. In some embodiments, the single touch screen input is received after the user “taps” the touch screen over a word or suggestion candidates. In other embodiments, the single touch screen input is received when the user presses the screen with a finger, thumb, or stylus. Receiving a single touch screen input is sufficient to determine which suggestion candidate the user is indicating on the touch screen—no additional keyboard input, mouse input, trackball input, voice input, or additional touches are necessary. Using a single touch screen input to determine user selections simplifies the input process and allows for the fast correction of text entries without the need to use submenus, popup menus, or additional input devices.

II. Example Mobile Device

FIG. 1 is a system diagram depicting an exemplary mobile device 100 including a variety of optional hardware and software components, shown generally at 102. Any components 102 in the mobile device can communicate with any other component, although not all connections are shown, for ease of illustration. The mobile device can be any of a variety of computing devices (e.g., cell phone, smartphone, handheld computer, Personal Digital Assistant (PDA), etc.) and can allow wireless two-way communications with one or more mobile communications networks 104, such as a cellular or satellite network.
The illustrated mobile device 100 can include a controller or processor 110 (e.g., signal processor, microprocessor, ASIC, or other control and processing logic circuitry) for performing such tasks as signal coding, data processing, input/output processing, power control, and/or other functions. An operating system 112 can control the allocation and usage of the components 102 and support for one or more application programs 114. The application programs can include common mobile computing applications (e.g., email applications, calendars, contact managers, web browsers, messaging applications), or any other computing application.
The illustrated mobile device 100 can include memory 120. Memory 120 can include non-removable memory 122 and/or removable memory 124. The non-removable memory 122 can include RAM, ROM, flash memory, a hard disk, or other well-known memory storage technologies. The removable memory 124 can include flash memory or a Subscriber Identity Module (SIM) card, which is well known in GSM communication systems, or other well-known memory storage technologies, such as “smart cards.” The memory 120 can be used for storing data and/or code for running the operating system 112 and the application programs 114. Example data can include web pages, text, images, sound files, video data, or other data sets to be sent to and/or received from one or more network servers or other devices via one or more wired or wireless networks. The memory 120 can be used to store a subscriber identifier, such as an International Mobile Subscriber Identity (IMSI), and an equipment identifier, such as an International Mobile Equipment Identifier (IMEI). Such identifiers can be transmitted to a network server to identify users and equipment.
The memory 120 can also be used for the candidate sources 116, which are used for generating and suppressing auto-corrections and generation suggestion candidates. Candidate sources 116 can include but are not limited to: a system dictionary, a user dictionary, a common speller application programming interface (CSAPI), touch models, and an input history data source.
The mobile device 100 can support one or more input devices 130, such as a touch screen 132, microphone 134, camera 136, physical keyboard 138, and/or trackball 140 and one or more output devices 150, such as a speaker 152 and a display 154. Other possible output devices can include haptic output devices such as a piezoelectric transducer 156 or other suitable device. Some devices can serve more than one input/output function. For example, touch screen 132 and display 154 can be combined in a single input/output device.
A wireless modem 160 can be coupled to an antenna (not shown) and can support two-way communications between the processor 110 and external devices, as is well understood in the art. The modem 160 is shown generically and can include a cellular modem for communicating with the mobile communication network 104 and/or other radio-based modems (e.g., Bluetooth 164 or Wi-Fi 162). The wireless modem 160 is typically configured for communication with one or more cellular networks, such as a GSM network for data and voice communications within a single cellular network, between cellular networks, or between the mobile device and a public switched telephone network (PSTN).
The mobile device can further include at least one input/output port 180, a power supply 182, a satellite navigation system receiver 184, such as a Global Positioning System (GPS) receiver, an accelerometer 186, and/or a physical connector 190, which can be a USB port, IEEE 1394 (FireWire) port, and/or RS-232 port. The illustrated components 102 are not required or all-inclusive, as any components can deleted and other components can be added.

III. Example Computing Environment

FIG. 2 illustrates a generalized example of a suitable computing environment 200 in which described embodiments, techniques, and technologies may be implemented. For example, the computing environment 200 can implement unexpected-key feedback, auto-correction notification, invoking likely next phrases, random playback of click sounds, and waiting for a keystroke latency before presenting suggestion candidates, as described below.
The computing environment 200 is not intended to suggest any limitation as to scope of use or functionality of the technology, as the technology may be implemented in diverse general-purpose or special-purpose computing environments. For example, the disclosed technology may be implemented with other computer system configurations, including hand held devices, multiprocessor systems, microprocessor-based or programmable consumer electronics, network PCs, minicomputers, mainframe computers, and the like. The disclosed technology may also be practiced in distributed computing environments where tasks are performed by remote processing devices that are linked through a communications network. In a distributed computing environment, program modules may be located in both local and remote memory storage devices.
With reference to FIG. 2, the computing environment 200 includes at least one central processing unit 210 and memory 220. In FIG. 2, this most basic configuration 230 is included within a dashed line. The central processing unit 210 executes computer-executable instructions and may be a real or a virtual processor. In a multi-processing system, multiple processing units execute computer-executable instructions to increase processing power and as such, multiple processors can be running simultaneously. The memory 220 may be volatile memory (e.g., registers, cache, RAM), non-volatile memory (e.g., ROM, EEPROM, flash memory, etc.), or some combination of the two. The memory 220 stores software 280 and candidate sources 285 that can, for example, implement the technologies described herein. A computing environment may have additional features. For example, the computing environment 200 includes storage 240, one or more input devices 250, one or more output devices 260, one or more communication connections 270, and one or more touch screens 290. An interconnection mechanism (not shown) such as a bus, a controller, or a network, interconnects the components of the computing environment 200. Typically, operating system software (not shown) provides an operating environment for other software executing in the computing environment 200, and coordinates activities of the components of the computing environment 200.
The storage 240 may be removable or non-removable, and includes magnetic disks, magnetic tapes or cassettes, CD-ROMs, CD-RWs, DVDs, or any other medium which can be used to store information and that can be accessed within the computing environment 200. The storage 240 stores instructions for the software 280 and candidate sources 285, which can implement technologies described herein.
The input device(s) 250 may be a touch input device, such as a keyboard, keypad, mouse, pen, or trackball, a voice input device, a scanning device, or another device, that provides input to the computing environment 200. For audio, the input device(s) 250 may be a sound card or similar device that accepts audio input in analog or digital form, or a CD-ROM reader that provides audio samples to the computing environment 200. The output device(s) 260 may be a display, printer, speaker, CD-writer, or another device that provides output from the computing environment 200. The touch screen 290 can act as an input device (receiving touch screen input) and as an output device (displaying the text entry area, suggestion candidates area, and/or touch keyboard).
The communication connection(s) 270 enable communication over a communication medium (e.g., a connecting network) to another computing entity. The communication medium conveys information such as computer-executable instructions, compressed graphics information, or other data in a modulated data signal.
Computer-readable media are any available media that can be accessed within a computing environment 200. By way of example, and not limitation, with the computing environment 200, computer-readable media include memory 220, storage 240, communication media (not shown), and combinations of any of the above.
Computer-readable media are any available media that can be accessed within a computing environment 200. By way of example, and not limitation, with the computing environment 200, computer-readable media include memory 220 and/or storage 240. As should be readily understood, the term computer-readable storage media includes the media for data storage such as memory 220 and storage 240, and not transmission media such as modulated data signals.

IV. Example Text Entry Device

FIG. 3 illustrates a generalized example of a suitable implementation environment 300 of a text entry device 305 connected to a computing cloud 325. The text entry device 305 includes several modules stored on a computer-readable storage medium 310, including a text input module 330 for receiving text entry input, touch input module 332 for receiving touch screen input from a touch screen (not shown), and an output module 334 for providing output to a touch screen. The communication module 320 adapts the text entry device 305 so that it can communicate with service providers located in the cloud 325. The computer-readable storage medium 310 also includes an indicator generation module 340 for generating feedback semaphores and other indicators, a correction module 342 for checking and correcting text entries, and a candidate generation module 344 for generating suggestion candidates. Indicator generation module 340, correction module 342, and candidate generation module 344 can communicate with multiple modules to determine correction and suggestion candidates, including a grammar checking module 350, a system dictionary module 351, a user dictionary module 352, a CSAPI (Common Speller API) module 353, and an IHDS (input history data source) module 354. In some embodiments, one or all of these source modules 350-354 can be provided by a service provider in an alternate location 380 in the cloud 325.
FIG. 15 illustrates a generalized example of a suitable implementation environment 1500 in which described embodiments, techniques, and technologies may be implemented.
In example environment 1500, various types of services (e.g., computing services) are provided by a computing cloud 1510. For example, the cloud 1510 can comprise a collection of computing devices, which may be located centrally or distributed, that provide cloud-based services to various types of users and devices connected via a network such as the Internet. The implementation environment 1500 can be used in different ways to accomplish computing tasks. For example, some tasks (e.g., processing user input and presenting a user interface) can be performed on local computing devices (e.g., connected devices 1530-1532) while other tasks (e.g., storage of data to be used in subsequent processing, including candidate sources) can be performed in the cloud 1510.
In example environment 1500, the cloud 1510 provides services for connected devices 1530-1532 with a variety of screen capabilities. Connected device 1530 represents a device with a computer screen (e.g., a mid-size screen). For example, connected device 1530 could be a personal computer such as desktop computer, laptop, notebook, netbook, or the like. Connected device 1531 represents a device with a mobile device screen (e.g., a small size screen). For example, connected device 1531 could be a mobile phone, smart phone, personal digital assistant, tablet computer, and the like. Connected device 1532 represents a device with a large screen. For example, connected device 1532 could be a television screen (e.g., a smart television) or another device connected to a television (e.g., a set-top box or gaming console) or the like. One or more of the connected devices 1530-1532 can include touch screen capabilities. Devices without screen capabilities also can be used in example environment 1500. For example, the cloud 1510 can provide services for one or more computers (e.g., server computers) without displays.
Services can be provided by the cloud 1510 through service providers 1520, or through other providers of online services (not depicted). For example, cloud services can be customized to the screen size, display capability, and/or touch screen capability of a particular connected device (e.g., connected devices 1530-1532).
In example environment 1500, the cloud 1510 provides the technologies and solutions described herein to the various connected devices 1530-1532 using, at least in part, the service providers 1520. For example, the service providers 1520 can provide a centralized solution for various cloud-based services. The service providers 1520 can manage service subscriptions for users and/or devices (e.g., for the connected devices 1530-1532 and/or their respective users).
Although the operations of some of the disclosed methods are described in a particular, sequential order for convenient presentation, it should be understood that this manner of description encompasses rearrangement, unless a particular ordering is required by specific language set forth below. For example, operations described sequentially may in some cases be rearranged or performed concurrently. Moreover, for the sake of simplicity, the attached figures may not show the various ways in which the disclosed methods can be used in conjunction with other methods.
Any of the disclosed methods can be implemented as computer-executable instructions stored on one or more computer-readable media (e.g., non-transitory computer-readable media, such as one or more optical media discs, volatile memory components (such as DRAM or SRAM), or nonvolatile memory components (such as hard drives)) and executed on a computer (e.g., any commercially available computer, including smart phones or other mobile devices that include computing hardware). Any of the computer-executable instructions for implementing the disclosed techniques as well as any data created and used during implementation of the disclosed embodiments can be stored on one or more computer-readable media (e.g., non-transitory computer-readable media). The computer-executable instructions can be part of, for example, a dedicated software application or a software application that is accessed or downloaded via a web browser or other software application (such as a remote computing application). Such software can be executed, for example, on a single local computer (e.g., any suitable commercially available computer) or in a network environment (e.g., via the Internet, a wide-area network, a local-area network, a client-server network (such as a cloud computing network), or other such network) using one or more network computers.
For clarity, only certain selected aspects of the software-based implementations are described. Other details that are well known in the art are omitted. For example, it should be understood that the disclosed technology is not limited to any specific computer language or program. For instance, the disclosed technology can be implemented by software written in C++, Java, Perl, JavaScript, Adobe Flash, or any other suitable programming language. Likewise, the disclosed technology is not limited to any particular computer or type of hardware. Certain details of suitable computers and hardware are well known and need not be set forth in detail in this disclosure.
Furthermore, any of the software-based embodiments (comprising, for example, computer-executable instructions for causing a computer to perform any of the disclosed methods) can be uploaded, downloaded, or remotely accessed through a suitable communication means. Such suitable communication means include, for example, the Internet, the World Wide Web, an intranet, software applications, cable (including fiber optic cable), magnetic communications, electromagnetic communications (including RF, microwave, and infrared communications), electronic communications, or other such communication means.
The disclosed methods, apparatus, and systems should not be construed as limiting in any way. Instead, the present disclosure is directed toward all novel and nonobvious features and aspects of the various disclosed embodiments, alone and in various combinations and subcombinations with one another. The disclosed methods, apparatus, and systems are not limited to any specific aspect or feature or combination thereof, nor do the disclosed embodiments require that any one or more specific advantages be present or problems be solved.

V. Example Touch Screen Text Entry Device

FIG. 4 depicts an exemplary embodiment 400 of a text entry device 401 having a touch screen 402. The touch screen 402 includes a display area for a touch screen keyboard 420, a suggestion candidates area 450, and a text display area 403. The text display area 403 is shown displaying a phrase “smsing while dr” 404. In some embodiments, only a portion of the text input is displayed because of, for example, screen size limitations. A carat (cursor) 406 is also shown in the text entry area 403. In some embodiments, the carat 406 can be placed at various positions in the text entry area using a single touch screen input. In some embodiments, the touch screen 402 has a substantially planar surface, and the display capability can be implemented using LED, LCD, electronic ink, DLP, Plasma, CRT, or other suitable display technology.
The text entry device 401 has a touch screen 402 that displays a touch screen keyboard 420 having several keys 424, 426, 428, 430, 432, 434, etc. Some of the keys, including the backspace key 430, return key 432, and space key 434 are also designated as delimiter keys. As shown, the touch screen keyboard displays the keys 424, 426, 428, 430, 432, 434, etc. as images on the touch screen 402, as the user's finger 442 is touching the key “s” 428. The touch screen can include capacitive, resistive, inductive, or other suitable technologies for determining the position of one or more touch inputs detected over the surface of the keyboard and converting this touch input into text input. In some embodiments, the touch input is created using a stylus or puck, while in other embodiments the touch input can be created using a finger or thumb. In other embodiments, the touch screen keyboard 420 can be implemented as a hardware keyboard including mechanical keys.
The suggestion candidates area 450 is depicted in FIG. 4 showing three suggestion candidates, including “drive” 455, “drug,” “draft,” “drop,” and “driving.” As shown, the suggestion candidates area 450 includes widgets directly above the keyboard area containing word prediction and disambiguation candidates. Placing the suggestion candidates area 450 close to the touch screen keyboard 420 can improve typing speed and reduce errors by allowing users to maintain their focus of attention near the keyboard area while correcting a phrase 410 in the text entry area 403. In other embodiments, the suggestion candidates area 450 can be placed nearby the phrase 404 in the text entry area 403.
The location of the text entry area 403, the keyboard 420, etc., can be varied based on the particular implementation and design.
FIGS. 5A-5C illustrate additional exemplary implementations 500, 530, 560 of the disclosed technology. FIG. 5A depicts an exemplary implementation 500 of a text entry device, which includes a touch screen 506 displaying a window 510, including a designated text entry area displaying the phrase “smsing while driving” 512. The touch screen 506 is also displaying a candidates area 520, and a touch screen keyboard 524. A finger 522 is shown selecting the candidate “driving” in the candidates area 520.
FIG. 5B depicts an exemplary implementation 530, which includes a mobile device 531 having a touch screen 532 with a text entry area 534 (displaying the phrase “smsing while driving” 536), a suggestion candidates area 542, and a hardware keyboard 540.
FIG. 5C depicts an exemplary implementation 570 of a text entry device, which includes a personal computer (PC) device having a display screen 571 and a keyboard 580. The display screen 571 is shown displaying the phrase “smsing while dr” 575 in a window 574. Also shown is a suggestion candidates area 586 displaying candidates “drive,” “draft,” “drop,” and “driving.” In some examples, the text entry device is not limited to receiving text input using a touch screen keyboard, but can also use hardware keyboards, handwriting recognition, or speech recognition to receive text input.
Methods and apparatus for performing handwriting recognition can include but are not limited to those based on: Bayesian networks, neural nets, hidden Markov models, or k-nearest-neighbor approaches. Methods and apparatus for performing speech recognition can include but are not limited to those based on a dynamic time warping approach or hidden Markov models.

VI. Example Indicators, Signals, and Feedback Semaphores

Feedback and guidance signals are examples of indicators that can be implemented in many ways. For example, mobile tactile feedback, such as piezoelectric tactile displays that responds to a stylus, or a stylus that produces tactile feedback, can be employed. For touch screen keyboards in particular, artificial tactile feedback via vibration actuators can improve text entry performance in both static environments and mobile environments (for example, typing in a moving vehicle such as an automobile or train). In some examples, keyboard events can be defined including a fingertip-over event to signal when a fingertip has touched a key, a fingertip-click event to signal that a key has been registered, a fingertip-slip event to signal when the fingertip moved over the edge of a key, and a home-key event for indicating where the home keys are located on the touch screen keyboard. For example, in an English-language keyboard, the home keys of “F” and “J,” one for each finger, can be associated with haptic feedback that is provided to indicate a user's fingers are over the home keys. Using these events can improve tactile touch screen keyboard accuracies obtained to close to those obtained using a physical keyboard. Examples of using haptic feedback with a touch screen keyboard may be found in: Brewster, et al., “Tactile feedback for mobile interactions,” Proc. of SIGCHI Conf. on Human Factors in Computing Systems, pp. 159-162. (2007); Brown & Brewster, “Multidimensional Tactons for Non-visual Information Display in Mobile Devices,” Proc. of MobileHCI Conf. 2004, pp. 231-238 (2004); and Hoggan et al., “Investigating the Effectiveness of Tactile Feedback for Mobile Touchscreens,” Proc. of the 26th Int'l Conf. on Human Factors in Computing Systems, pp. 1573-1582 (2008).
Although employing different keyboard layouts can assist text entry on touch screen keyboards, these layouts have not yet become widespread. It is desirable to create semaphores for a touch screen keyboard that can be marketed widely.
In addition or instead of semaphores created using tactile feedback, similar results can be achieved by providing visual and auditory cues with a standard touch screen keyboard. While some users complain that their fingertips obfuscate visual feedback provided using a soft keyboard, visual and auditory feedback for the three keyboard events discussed above can be employed. Some examples of visual/audio feedback includes: a tooltip balloon that appears and disappears at the fingertip-over and fingertip-slip events, respectively, and an audible click is played at fingertip-click events.
Obstacles to providing effective visual and auditory cues in mobile settings include small screen size, outside noise, social restrictions and other circumstantial demands. Moreover, it has not been empirically established that artificial tactile feedback alone can impart enough information to realize full parity with physical keyboards. Furthermore, some users find tactile feedback to be annoying. Because individual differences may account for users' preference of one modality over another, touch screen keyboards employing a variety of tactile, visual and auditory semaphores, and combinations thereof, can be used.
Certain examples of the disclosed technology include touch screen keyboards using “typing semaphores.” As described further below, typing semaphores can provide multimodal feedback and/or guidance signals that help users discover and avoid typing errors. Multimodal feedback refers to the use of visual, auditory, and/or haptic (touch output) communication that are received be a text entry device user to indicate information such as potential errors, the occurrence of events such as auto-correction, or the availability of suggestion candidates.
A conventional semaphore is a system for conveying information by means of hand-held flags or lights. Widely adopted by the maritime world in the 1800s, semaphores are still employed today to safely guide airplanes onto naval vessels. Typing semaphores can be viewed as multimodal signals that guide users to enter text correctly—by analogy, a kind of flight control for typing. State of the art touch screen keyboards make it too easy for users to “crash,” expecting them to edit their text thereafter, which can be mentally disruptive, time-consuming, and taxing.
The purpose of typing semaphores is to improve text entry user experience and performance as users type (e.g., in real-time) or otherwise provide input data using both feedback signals to alert user to possible typing errors when they occur, and guidance signals to prevent future errors.

VII. Example Application Using Indicators and Suggestion Candidates

FIG. 6 is a flow chart 600 that outlines a method of generating possible corrections for input data, providing indicators, including feedback semaphores that indicate the availability of the possible corrections, and displaying suggestion candidates. After displaying the suggestion candidates, an input is received selecting one of the suggestion candidates, and the input data is modified with one or more words associated with the selected suggestion candidate.
At process block 610, input data comprising one or more input word(s) is received from a source such as a touch screen keyboard. In some embodiments, the first input data includes text characters, text words, position data for key presses on a touch screen keyboard, typing speed data, correction data, and/or touch screen orientation data.
At process block 620, one or more feedback semaphores are provided to indicate that one or more suggestion candidates are available for one or more of the input words. The feedback semaphores can include various indicators or signals to alert a user while maintaining focus in the keyboard area by, for example, highlighting one or more keys of a keyboard, playing back a sound using a speaker, generating haptic feedback (e.g., vibrating a handheld device), or displaying a tooltip balloon.
At process block 630, one or more suggestion candidates are provided to a user (e.g., using a touch screen display). Each of the suggestion candidates is associated with an alternative word for one or more of the input words. In some examples, the suggestion candidates are associated with alternative words for only one of the input words, while in other examples, suggestion candidates are associated with alternative words for more than one of the input words. An exemplary display of a suggestion candidate includes displaying a button using a touch screen display, where the button includes the text of an associated alternative word within the boundary of the button.
At process block 640, an input selecting one of the suggestion candidates is received. For example, the selection can received using a single touch screen input created by a user pressing a finger over a button associated with the desired touch screen candidate on a touch screen display.
At process block 650, the input data is modified using the alternative word associated with the selected suggestion candidate. In some examples, the alternative word is used to replace the input word in the input data. In other examples, the alternative word is added to the input data preceding or subsequent to the input word. In some examples, the “alternative” word is a word that was previously auto-corrected, and the alternative word is therefore used to effectively undo a word inserted in the input data using an auto-correction routine. Thus, as described above, a quick and effective way of providing suggestion candidates and indicators of their availability is provided that allows users to quickly modify input data using the suggestion candidates without moving their focus of attention away from a touch screen keyboard.

VIII. Example Unexpected Key Feedback

Because touch screen keyboards lack tactile feedback, it is difficult for users to tell when they hit their desired keys. Consequently, they have to pay more attention to the keyboard area than when using hardware keyboards. FIGS. 7A-7E depict a method 700 for providing visual, audio, and tactile feedback to a user that has pressed an expected key on a mobile device 702. FIG. 8 is a flow chart 800 corresponding to the method 700 depicted in FIGS. 7A-7E.
As shown in FIG. 7A, a mobile device 702 comprises a text display area 704, suggestion candidates area 708, and touch screen keyboard 710. The text display area 704 is shown displaying a phrase “smsing whil” 706, with a carat (cursor) 712 immediately after the phrase as the user enters it on the keyboard 710.
FIG. 7B depicts the mobile device 702 providing the user with unexpected key feedback semaphores as an unexpected key, “r” 726, is pressed with a user's finger 722. The types of feedback semaphores shown include: a tooltip balloon 728 displaying the unexpected key, an audio “clunk clunk” sound 740 (played back over a speaker coupled to the mobile device), haptic feedback (e.g., a vibration created by an oscillating transducer coupled to the mobile device) indicated by squiggle lines 755 and 756, suggestion candidates “whip” 730, “whilst” 731, and “while” 732, and a highlighted key indicator shown as shading of the unexpected “r” key 726. The unexpected “r” key 726 is designated as unexpected based on the current context of the text input data shown on the text display area 704 (e.g., the word “whilr” does not exist in the English language). Here, the suggestion candidates 730-732 are actually based on the data input phrase “smsing whil” 706 before the “r” was received, but as shown in FIG. 7C, will be updated based on the unexpected key input.
FIG. 7C depicts the mobile device 702 shortly after providing one or more feedback semaphores as discussed above. As shown, the phrase 742 has been updated with the unexpected key “r,” and the carat 712 is placed immediately after the phrase. Further, suggestion candidates “while” 750, “white” 751, and “wholesale” 752 have been provided based on the phrase 742. In addition, designated delimiter keys including a space bar 758 and an enter key 759 have been highlighted to indicate that the highlighted suggestion candidate will replace the word “whilr” if one of the delimiter keys is pressed next. Alternatively, the user can press one of the suggestion candidates 750-752 in order to select the corresponding word shown on the candidate button to replace the word “whilr.”
FIG. 7D depicts the mobile device 702 as a user uses a finger 723 to provide a single touch screen input over a delimiter key (the space bar 758).
FIG. 7E depicts the mobile device 702 after receiving the delimiter key. As shown, the phrase 782 has been updated to replace the word “whilr” with the word “while,” which was the word associated with the highlighted suggestion candidate 750. The carat 712 has also been advanced one space past the phrase 782.
The manner of providing feedback semaphores used can be selected based on their effectiveness in alerting users without annoying them. For example, when tooltip balloons (e.g., tooltip balloon 728) appear above the keys of the keyboard 710, users do not always see these visual cues, especially if they are typing quickly. It is desirable to use feedback semaphores that can alert users to unexpected key presses so that they can keep their focus on the keyboard 710, or immediately switch their focus of attention to the candidates area 708. A key is designated “unexpected” when the letters entered so far (e.g., up to the previous word boundary) do not match the prefix of any word in a candidate source (e.g., a dictionary, CSAPI, or input history data source).
FIGS. 7F-7H depict different designs 760, 770, and 780 of highlighting a touch screen that can be used to provide unexpected-key feedback semaphores to a user. In FIGS. 7F-7H, a mobile device 702 is shown as the user types the “r” key after entering the phrase “smsing whil” 762. As shown in the design 760 of FIG. 7F, a shaded tooltip balloon 766 and both the background and keys of the touch screen keyboard 764 are highlighted to provide an unexpected key feedback semaphore. Similarly, the design 770 of FIG. 7G depicts the mobile device 702, but with a shaded tooltip balloon 776 and only the background of the touch screen keyboard 774 highlighted. Finally, the design 780 of FIG. 7H depicts the mobile device 702, but with a shaded tooltip balloon 786 and only the border of the background of the touch screen keyboard 784 highlighted. The highlighting and shading shown in FIGS. 7A-7H can be based on using a pattern and/or color(s), and can slowly fade back to the original key colors over time to present a more pleasing visual appearance.
Usability studies have indicated that some users found the design 760 shown in FIG. 7F, including a shaded keyboard 764, shaded “r” key, and shaded tooltip balloon 766, “too disruptive.” Although the design 770 of FIG. 7G, including shading for only the background of the keyboard 774, shaded “r” key, and shaded tooltip balloon 776, was found to be less disruptive, users still found this design more distracting than the design shown in FIG. 7H, which includes only a shaded border around the keyboard 784, shaded “r” key, and shaded tooltip balloon 786. Thus, by performing usability studies, desirable combinations of visual and auditory feedback that are agreeable to users can be determined.
FIG. 8 is a flow chart 800 further detailing the method 700 shown in FIGS. 7A-7E. FIG. 8 depicts process blocks for receiving input data, evaluating the data using one or more candidate sources to determine unexpected input, and based on the evaluating, displaying feedback semaphores and suggestion candidates allowing a word to be replaced with a selected suggestion candidate.
At process block 810, input data is received from a source such as a touch screen keyboard. In some embodiments, the first input data includes text characters, text words, position data for key presses on a touch screen keyboard, typing speed data, correction data, and/or touch screen orientation data.
At process block 820, the input data is analyzed and compared against one or more of the following candidate sources: a system dictionary; a user dictionary; a common speller application programming interface (CSAPI); or an input history data source that can include previously generated input history data for one or more users including word probability data, key probability data, edit distance data, and touch model probability data. In some examples, the analysis and comparison is repeated continuously for each character of text input, while in other examples, the analysis and comparison may only occur after an event. Examples of events include receiving a delimiter key (e.g., a space, backspace, enter, ESC, or other designated key(s)) or determining a latency in user typing speed that exceeds a threshold value.
At process block 830, the analysis and comparison data from process block 820 is evaluated to determine if any unexpected input data was received. If no unexpected input data is received (e.g., all the words of the text input data can be found in one or more of the candidate sources used at process block 820), the method proceeds to process block 810 in order to continue receiving input data. If unexpected input data is detected, the method proceeds to process block 840. In some embodiments, a word can be designated as unexpected input data even though the word may exist in one of the candidate sources, for example, using a grammar checker module. An unusual word such as “whilst” may still be designated as unexpected input data because other similar words, for example, “while” have a higher probability of correctness, based on word probability data, key probability data, edit distance data, and touch model probability data. In some examples, the feedback semaphores for unusual words designated as unexpected may be modified in comparison to feedback semaphores used for words that do not exist in any of the candidate sources at all.
At process block 840, an indicator such as a feedback semaphore is provided in order to alert a user that suggestion candidates are available. As discussed above, the feedback semaphores provided can include visual displays, including the highlighting of touch screen keyboard keys or keyboard background; audio playback, such as a distinctive “clunk clunk” sound (which indicates to the user that something may be incorrect in the input data), or haptic feedback.
At process block 850, suggestion candidates are provided, using for example, a candidates area on a touch screen display. Each suggestion candidate is associated with an alternative word that will be used to modify the input data if that particular suggestion candidate is selected. In some examples, one of the candidates can by displayed differently (or highlighted differently) than other candidates to indicate that it is the most likely suggestion candidate. This distinguishing display not only provides a visual cue to the user, but can also be used to indicate that the input data will be modified with the suggestion candidate if other designated keys are pressed, for example, a delimiter key such as the space key, punctuation keys, or the return or enter keys.
At process block 860, the selection (or absence of a selection) of one of the suggestion candidates is detected. Suggestion candidates can be selected several ways, including receiving a single touch screen input over the suggestion candidate on the touch screen display, or by receiving a keystroke for a designated delimiter key. If no suggestion candidate is selected, the method proceeds to process block 810 in order to receive more input data. If a suggestion candidate selection is detected, the method proceeds to process block 870.
At process block 870, the input data is modified using an alternative word associated with the selected suggestion candidate. In some examples, the alternative word is used to replace the input word in the input data. In other examples, the alternative word is added to the input data preceding or subsequent to the input word. In some examples, the “alternative” word is a word that was previously auto-corrected, and the alternative word is therefore used to effectively undo a word inserted in the input data using an auto-correction routine. Thus, as described above, a quick and effective way of providing suggestion candidates and indicators of their availability is provided that allows users to quickly modify input data using the suggestion candidates without moving their focus of attention away from a touch screen keyboard.

IX. Example Auto-Correction Feedback

Users often keep their focus of attention on the keyboard area when typing on a touch screen keyboard, which often causes users to fail to see auto-corrections that may be replacing legitimate words, such as names and technical terms that do not exist in the dictionary. This can lead to tremendous frustration, especially if users do not notice the text replacements until much later.
FIGS. 9A-9D depict a method 900 of providing the user with an auto-correction feedback semaphore. FIG. 10 is a flowchart 1000 corresponding to the method 900 depicted in FIGS. 9A-9D.
As shown in FIG. 9A, a mobile device 902 comprises a text display area 904, suggestion candidates area 910, and touch screen keyboard 908. The display area 904 is shown displaying a phrase “smsing” 920 after a user has typed the unknown word “smsing,” but before typing a delimiter, so a carat 916 is displayed immediately after the phrase. Shown in the suggestion candidates area 910 are the suggestion candidates “ending” 912, “facing” 913, and “smashing” 914.
FIG. 9B depicts the mobile device 902 as the user presses the space bar 932 with a finger 934, which causes the space bar to immediately be highlighted red, the word “smsing” in the phrase 920 to be replaced with a replacement word “ending,” 922 a “swish” sound 950 to be played over a speaker, and the mobile device 902 to vibrate, as shown by squiggle lines 955 and 956. The carat 916 is advanced by a space, and the mobile device also displays the suggestion candidates “smsing” 940, “facing” 941, and “smashing” 942 in the candidates area. The candidate “smsing” 940 is the word that was just replaced by auto-correction, and the candidate 940 is colored green to indicate that it is the replaced word.
FIG. 9C depicts the mobile device 902 after the user has decided to undo the auto-correction. To do so, the user uses a finger 962 to make a single touch screen input over the suggestion candidate “smsing” 966, which will undo the auto-correction.
FIG. 9D depicts the mobile device 902 after the auto-correction has been undone. As shown, the word “ending” has been replaced with the original word “smsing” 982, the suggestion candidates area 910 is cleared, and the carat 916 remains a single space past the replaced word.
FIG. 10 is a flow chart 1000 further detailing the method 900 shown in FIGS. 9A-9D. At process block 1010, a mobile device receives input data from a source such as a touch screen keyboard. In some embodiments, the first input data includes text characters, text words, position data for key presses on a touch screen keyboard, typing speed data, correction data, and/or touch screen orientation data.
At process block 1020, the input data is analyzed to determine if a designated delimiter key (e.g., a space key or an enter key) has been received in the input data. If not, then the method proceeds process block 1010 to receive more input data. If a delimiter key was received in the input data, the method proceeds to process block 1030.
At process block 1030, the last word of the input data is compared against one or more candidate sources, and an auto-correction is applied to replace a word designated as incorrect with a replacement word from a candidate source.
At process block 1040, one or more auto-correction feedback semaphores are displayed or played back, in a manner that is likely to keep the user's focus of attention on the keyboard. For example, a “swish” sound can be played, haptic feedback provided, and red highlighting displayed on the delimiter key that was pressed by the user. The feedback semaphores are displayed in a manner that does not require the user to shift their focus of attention away from the keyboard, and therefore the user is able to more easily recognize that an auto-correction has occurred, as well as to easily view suggestion candidates that will be presented at process block 1050.
At process block 1050, one or more suggestion candidates are presented to the user, including the original word that was auto-corrected. One of the suggestion candidates is also highlighted using, for example, shading or a different color. The highlighted suggestion candidate is designated as the most likely substitution that the user will select. In some examples, the highlighted suggestion candidate is original word that was auto-corrected. In other examples, the highlighted suggestion candidate is a word from a candidate source designated to be the most likely substitution. The method proceeds to process block 1060 after receiving touch screen input.
At process block 1060, the method determines whether the user selected a delimiter key. If a delimiter key was pressed, the method proceeds to process block 1070, where the word inserted by auto-correction at process block 1030 is itself replaced with the word shown as the highlighted suggestion candidate, and the method then proceeds to process block 1010, where more input data is received. If a delimiter key was not pressed, the method proceeds to process block 1080.
At process block 1080, the method determines whether the user selected a suggestion candidate (e.g., by providing a single touch screen input over the suggestion candidate). If a suggestion candidate is selected, the method proceeds to process block 1090, where the word inserted by auto-correction at process block 1030 is itself replaced with the word associated with the selected suggestion candidate. If a suggestion candidate is not pressed, the method proceeds to process block 1010 to process the selected key (e.g., by adding the selected key to the input data) and continue receiving input data.

X. Example Invoking Next Phrases

FIGS. 11A-11F depict exemplary methods 1100 and 1150 of providing a user with predicted next phrases after receiving a completed word of text input using suggestion candidates. In some examples, users can select two or more suggestion candidates sequentially, thereby improving typing speed performance. FIG. 12 is a flow chart 1200 illustrating an exemplary implementation of a method corresponding to the methods 1100 and 1150 shown in FIGS. 11A-11F.
As shown in FIG. 11A, a mobile device 1102 has a text display area 1104, a suggestion candidates area 1110, and a keyboard 1108. The mobile device 1102 is shown after a user has typed one word of the phrase “happy” 1106, and the carat 1116 is positioned immediately after the word “happy.”
FIG. 11B depicts the mobile device 1102 after the user has typed a delimiter key (e.g., the space bar). As shown, three suggestion candidates “birthday” 1120, “new” 1121, and “go lucky” 1122 are displayed in the candidates area 1110. The user is shown using a finger 1130 to select the first suggestion candidate 1120.
FIG. 11C depicts the mobile device 1102 after adding an additional next word “birthday” 1138 after the word “happy” in the phrase 1106 based on the selected suggestion candidate and positioned the carat 1116 one space away from the end of the word 1138.
FIG. 11D depicts an exemplary method 1150 including an alternative selection of a suggestion candidate by a user compared to that shown in FIGS. 11A-11C. FIG. 11D shows a mobile device 1152 entering a delimiter character immediately after typing the word “happy” 1154. The carat 1156 is positioned one space past the word. A user is shown selecting the suggestion candidate “new” 1162 with a finger 1165 in the suggestion candidates area 1160.
FIG. 11E depicts the mobile device 1152 after the user has selected the suggestion candidates “new” 1162. The word “new” 1170 is added to the text input after the word “happy,” another search of one or more candidate sources is performed for the phrase “happy new,” and several suggested next words for the phrase “happy new” are presented as suggestion candidates: “year” 1182, “baby” 1183, and “day” 1184. The user is shown selecting the suggestion candidate “year” 1182 with a finger 1175.
FIG. 11F depicts the mobile device 1152 after adding the word “year” 1190 associated with the selected suggestion candidate 1182 to the text input, to form the phrase “happy new year.” Another search of the one or more candidates sources is performed, but no matches are found, so the suggestion candidates area 1160 is cleared, and the carat 1156 is placed one space from the word “year” 1190.
FIG. 12 is a flow chart 1200 further detailing the exemplary methods 1100 and 1150 shown in FIGS. 11A-11F. At process block 1210, input data is received, and the method proceeds to process block 1220, where the presence or absence of a delimiter character in the input data detected. If no new delimiter key is received in the input data, the method proceeds to process block 1210 to receive additional input data. If a delimiter key is received, then the method proceeds to process block 1230 to generate suggestion candidates for the input data. In some examples, only the last word typed is used to search for suggestion candidates, while in other examples, additional words in the input data can be included in the search.
At process block 1240, the method determines if any suggestion candidates were generated. If no suggestion candidates are generated, the method proceeds to process block 1210 to receive additional input data. If one or more candidates were generated at process block 1230, the method proceeds to process block 1250. At process block 1250, the suggestion candidates are presented to the user using a touch screen display coupled to the mobile device.
At process block 1260, a suggestion candidate selection (or absence of a selection) is detected. Suggestion candidates can be selected several ways, including receiving a single touch screen input over the suggestion candidate on the touch screen display, or by receiving a keystroke for a designated delimiter key. If no suggestion candidate is selected, the method proceeds to process block 1210 in order to receive more input data. If a suggestion candidate selection is detected, the method proceeds to process block 1270.
At process block 1270, the input data is modified using an alternative word associated with the selected suggestion candidate. In some examples, the alternative word is used to replace the input word in the input data. In other examples, the alternative word is added to the input data preceding or subsequent to the input word. In some examples, the “alternative” word is a word that was previously auto-corrected, and the alternative word is therefore used to effectively undo a word inserted in the input data using an auto-correction routine. After modifying the input data, the method immediately proceeds to process block 1230, where more suggestion candidates are generated based on the modified input data. In this manner, a user can select multiple suggestion candidates from a changing set of suggestion candidates, thereby improving typing speed and accuracy, especially for highly frequent combinations of words and phrases.

XI. Example Use of Typing Speed and Simulating Click Sounds

FIG. 13 is a chart 1300 representing some events that occur during a method that includes the use of random click sounds played back as the user presses keys on a touch screen keyboard, as well as the use of keystroke latency to determine when to present suggestion candidate to a touch screen keyboard user. FIG. 14 is a flow chart 1400 that corresponds to the methods shown in FIG. 13.
Shown in FIG. 13 is a time axis 1330 that indicates when events occur or are received by a mobile device. The user enters several successive key presses on the touch screen: “b” 1340, “a” 1341, “n” 1342, and “a” 1343 in relatively rapid succession. At the same time, the mobile device randomly selects and plays a “click” sound that simulates the sound that a key press makes when typing on a hardware keyboard. For example, the mobile device can randomly select one of the five different click sounds labeled #1-#5. As shown, the click sounds 1310-1314, etc., are not always the same for a given key that is pressed. For example, the first keypress “a” 1341 is followed by playback of click sound # 2 1311, while the second keypress “a” 1344 is followed by playback of click sound # 3 1314. Similarly, the two keypresses shown for “n” playback sound # 3 1312 the first keypress 1342 and sound # 5 1313 for the second keypress 1344. By providing playback of random click sounds that sound like a hardware keyboard, the user receives more naturalist audio feedback that more closely simulates typing on a hardware keyboard.
Also shown in FIG. 13 is the presentation of suggestion candidates based on the user's keystroke latency. The mobile device monitors a user's keystroke latency as words are typed, and if the user pauses for a length of time that exceeds a pre-determined threshold, suggestion candidates are presented for the input data of the most recent word input. The techniques for presenting the suggestion candidates are similar to those discussed above. In some examples, suggestion candidates are only presented when the partial data input is determined to be incomplete, i.e., when the partial input data does not appear in a candidate source such as a dictionary or input history data. In other examples, suggestion candidates are presented after a pre-determined latency if no delimiter has been received for the word. In some examples, the user can define the latency threshold value, while others examples use a pre-defined threshold value.
As shown, the user types the keys “b” 1340, “a” 1341, “n” 1342, and “a” 1343 in relatively rapid succession, followed by a relatively longer pause. After the latency threshold has been reached based on the delay after the keypress 1343, the mobile device presents several suggestion candidates for the partial input data “bana,” includes candidates “banana” 1350, “bananas” 1351, and “banal” 1352. As shown, the user decides to ignore the suggestion candidates and continues typing “na<space>s” followed by another pause after keypress “s” 1345. Another time period exceeds the latency threshold value after the keypress “s” 1345, and different suggestion candidates 1360-1363 are generated. Thus, the suggestion candidate generation module can consider not only the partial input word “s,” but also the previous word input “banana.” (e.g., “banana s” is the input being considered by the candidate generation module). Thus, very accurate suggestion candidates can be provided even though the user has only entered a single letter “s” of the latest input word. As shown, the user selects the candidate “slug” 1362 by pressing the “slug” suggestion candidate 1370 on the touch screen display, and the input data is modified to “banana slug.”
FIG. 14. is a flow chart 1400 further detailing the method shown in FIG. 13. At process block 1410, input data is received from a source such as a touch screen keyboard. In some embodiments, the first input data includes text characters, text words, position data for key presses on a touch screen keyboard, typing speed data, correction data, and/or touch screen orientation data.
At process block 1420, a random click sound is selected from a collection of two or more sounds and played back to the user using, e.g., a speaker. As discussed above regarding FIG. 13, the click sounds are not associated with any specific key, and any particular key press on the same key can have multiple random sounds played back.
At process block 1430, the latency between the most recently received keystroke, and last keystroke received immediately preceding that keystroke, is measured. At process block 1440, the measured latency is compared to a pre-determined threshold value. In some examples, this latency threshold can be selected by the user. In other examples, the threshold can come from a default value. In other examples, the threshold can be selected based on a user's previous or average typing speed. If the measured latency does not exceed the threshold, the method proceeds to process block 1410, and more input data is received. If the measured latency does exceed the threshold, the method proceeds to process block 1450.
In some examples, the keystroke latency or typing speed data is also used to select which types of feedback semaphore(s) are presented. For example, if a user is typing relatively rapidly, only one or two more subtle semaphores are provided (e.g., only delimiter key highlighting can be used for fast typists). If a user is typing relatively slowly, more semaphores, or semaphores that use stronger visual, audio, and/or haptic cues are provided (e.g., louder audio, use of haptic force, and/or brighter or different-colored highlighting can be used for slower typists). Alternatively, or in addition, typing speed data, touch point data, device orientation, etc., can also be used to detect whether a user is typing using one or two hands. Thus, different types of feedback semaphores can be provided based on whether the user is typing using one or two hands.
At process block 1450, the input data is analyzed and suggestion candidates are generated and presented for the input data in a similar fashion to those techniques for generating suggestion candidates discussed above. In some examples, candidates are only generated if the input data is partial and cannot be found in a candidate generation source. In other examples, candidates are generated unless the last key pressed is a delimiter key. In some examples, only partial input data is considered, while other examples can analyze other words in the input data (e.g., the word immediately preceding or the word immediately subsequent to the current input word). After the candidates are generated, they are presented to the user using, e.g., a candidates area on a touch screen device.
At process block 1460, the selection (or absence of selection) of one of the suggestion candidates is detected. If a suggestion candidate selection is detected, the method proceeds to process block 1470, where the partial input data is modified by replacing the input word with the word associated with the selected suggestion candidate, and the method proceeds to process block 1410 to receive more input data. If a suggestion candidate is not detected as selected, then the method proceeds to process block 1410 to receive more data.

XII. First Experimental Results

A. Introduction
Experiments were conducted in order to measure the text entry performance improvement in a controlled experiment using several semaphore designs on naïve participants in a usability study.
One of the semaphores studied is called mobile key-trail feedback. To help users keep track of the keys they just typed in the keyboard area, the on-screen keyboard utilizes a feedback technique where pressed keys briefly light up and gradually fade away.
Another semaphore studied is a guidance semaphore called key-prediction guidance. An example of key-prediction guidance is highlighting or bolding the next likely letter on a keyboard.
The other two semaphores studied were the feedback semaphores unexpected-key feedback and auto-correction feedback.
B. Experiment Methodology
Four female and seven male participants from a metropolitan area were recruited by a professional contracting service. Three owned touch screen phones, five had owned a QWERTY phone at some point in their lives, and three owned only 12-key numeric phones. The participants came from various occupational backgrounds from housewife to IT professional, and were within an age range of 19-39. All participants were compensated for their time.
Participants were shown a short phrase on a desktop computer screen from the well-known phrase set described in MacKenzie & Soukoreff, “Phrase Sets for Evaluating Text Entry Techniques,” Extended Abstracts of CHI 2003, pp. 754-755 (2003) (hereinafter “MacKenzie I”). The set contains 500 short English phrases with no punctuation, varying from 16-43 characters with a high letter frequency correlation with an English corpus. The phrase set was supplemented with news headline phrases culled from the Internet containing words not found in the test dictionary (e.g., “smsing while driving is risky,” “obama is inaugurated”). Four phrases and one supplemental phrase were randomly selected for each of sixteen conditions: one for each semaphore (4-choose-1), plus combinations thereof (4-choose-2,4-choose-3, and 4-choose-4). These combinations were tested in order to examine if any of the semaphores conflicted with each other, and to investigate how people felt about having a semaphore be absent in another condition. The order of the conditions was not counter-balanced. Using the “think-aloud” protocol described in Lewis & Rieman, Task-Centered User Interface Design: A Practical Introduction (1993) (available via anonymous ftp at: ftp.cs.colorado.edu), participants were asked in each of the sixteen conditions to type in the phrases “as quickly as possible,” but instructed that they should pause and verbalize any new thoughts they had about the semaphores at any moment. During the study, specific questions were asked about each semaphore, and at the end of the study, participants were asked to rank-order any semaphores they would leave on by default.
C. Results
Because this was a usability study aimed at refining the user experience and design of the semaphores, the visual and auditory parameters of some semaphores were continually adjusted based on user feedback. As such, it is difficult to accurately interpret raw statistics. However, the numbers do convey general trends reported here.
1. Unexpected-Key Feedback
Seven of the eleven participants stated that they would leave the unexpected-key feedback semaphore turned on by default, even when using the more “distracting” versions shown in FIGS. 7F-7H. Five of the eleven listed this semaphore as their top choice, stating that they “depended” on it. In particular, they noted how it would alert them to the candidates area where they would almost certainly find their desired word as a choice. In fact, many participants claimed that this was their “strategy” for typing as quickly and as accurately as possible. Interestingly, after the settling on a visual design, half the participants said they did not perceive the visual cues and only relied on the auditory cue, and half said that they would turn off the auditory cue, as they considered the visual cue informative enough.
2. Auto-Correction Feedback
All the eleven participants loved the auto-correction semaphore and said that they “depended” on knowing when words were being auto-corrected. Furthermore, several participants noted that they could even predict when certain words would be auto-corrected (e.g., “obama”). This helped them to be prepared to select their replaced word in the candidate area, as shown in, for example, FIG. 9C. Many participants asked to have the replaced word be automatically added to the dictionary when selected from the candidate area.
3. Key-Trail Feedback
Only five of the eleven participants stated that they would leave the semaphore on by default, with two of the five claiming that they “depended” on it, and three of the five claiming that it was “cool eye candy.” Informally, it was observed that those two subjects had very slow baseline typing speeds. Of the six who did not choose to leave the semaphore on, two really disliked it (although this may have been due to the fading rate, which were later adjusted).
4. Key-Prediction Guidance
Only five of the eleven participants stated that they would leave the semaphore on by default, with three of the five listing it as their top choice. Of the five, only one claimed he “depended” on it, particularly for one-handed use, with the rest stating that they would leave it on “just in case [they] were unsure of how to spell a long word.”
5. Combinations
Initially, two participants disliked the combination of the key-prediction feedback with the key-trail feedback, but that was when the keys were entirely colored. Once the color blue was applied only to the letters, no complaints were received. No other combinations seemed to bother the participants.
6. Discussion
Given the uniformly positive response for the auto-correction feedback, it was decided not to test this semaphore further. Indeed, for the subsequent experiment, this semaphore was left on all the time. The diversity of preferences for the other semaphores suggests that users should be given the option of turning specific semaphores on and off in a control panel. With the unexpected-key feedback, some people seemed to be more tuned to the audio cue and some to the visual, suggesting that giving users fine-grained control over which cue is on would be valuable.
In watching and listening to the usability participants, it was observed that the semaphores seemed to have more utility for people who typed slower than those who were accustomed to typing quickly on touch screen keyboards. Interestingly, all of the touch screen phone owners stated that they would leave the unexpected-key feedback on by default, with two of the three listing it as their top choice.

XI. Second Experimental Results

A. Introduction
Given the feedback received from the usability study discussed above, it was decided to compare the text entry performance of only the unexpected-key feedback and the key-prediction guidance semaphores. The auto-correction feedback semaphore was incorporated into the baseline for two reasons: (1) the usability results for auto-correction feedback were compelling, and (2) to see how well it would work with the other two semaphores for timed text entry. Key-trail feedback was not tested in the second experiment. Although the key-trail feedback received usability responses that were as ambivalent as the key-prediction feedback, key-prediction feedback was tested instead because of related previous research as discussed in MacKenzie & Zhang, “Eye Typing Using Word and Letter Prediction and a Fixation Algorithm,” Proc. of the ACM Symposium on Eye Tracking Research and Applications—ETRA 2008, pp. 55-58 (2008) (hereinafter “MacKenzie II”) and Magnien, et al., “Mobile Text Input with Soft Keyboards: Optimization by Means of Visual Clues,” Proc. of MobileHCI, pp. 337-341 (2004) (hereinafter “Magnien”).
B. Methodology
Eighteen participants (nine males and nine females) between the ages of 21 and 39 were recruited using the same professional contracting service as in Experiment One. Participants again hailed from a wide variety of occupational backgrounds. All participants were compensated for their time. Five owned touch screen phones at some time in their life, nine owned QWERTY phones at some time in their life, and seven owned 12-key numeric phones only (note that these are not exclusive categories). During recruiting, all participants answered that they were familiar with the QWERTY layout and could type on a normal-size keyboard without frequently looking at the keys.
The MacKenzie and Soukoreff phrase set was again utilized, except in order to make sure that participants had a chance to hit every letter on the keyboard, a script to select the shortest sequences of phrases was used that covered the entire alphabet from A to Z. Supplemental phrases with words not found in the test dictionary were not included, to attempt to reproduce previous results for the key-prediction guidance semaphore on a mobile QWERTY touch screen keyboard. For each condition, subjects received eight practice and twenty stimuli items.
All participants were first taught the basics of using the touch screen keyboard on a mobile device, in particular, a prototype before-market phone with a 3.5 inch resistive screen having 800×480 WVGA resolution. A target phrase was displayed on a desktop computer screen and participants were asked to memorize it. They had as much time as they needed to memorize the phrase. Participants were asked to memorize the phrases in order to mimic the real experience of entering intended text. When participants felt they were “ready,” their task was to type the phrase into the mobile device “as quickly and as accurately as possible.” The phrase was left displayed on the computer screen, because in previous experiments with the same task, some naïve participants (who were not all university students) experienced difficulties with memorization under timed conditions. Timing began as soon as they entered the first letter of the phrase and ended when they hit the “Enter” button twice. The entire experiment took slightly under two hours.
The primary independent variable used was Semaphore, consisting of Baseline, Unexpected-Key Feedback and Key-Prediction Guidance. As dependent measures, time to enter text, accuracy, and the efficiency measure: keystrokes-per-character or KSPC as described in MacKenzie & Tanaka-Ishii, Text Entry Systems: Mobility, Accessibility, Universality (Morgan Kaufmann Publishers 2007) (hereinafter “MacKenzie III”) were examined.
The number of times users pressed the backspace button was used as an additional measure. Note that users were not allowed to place the cursor onto their typed text for editing since not all participants were proficient at this task. In short, a simple, within-subject experiment was conducted where all participants encountered the three Semaphore conditions in counter-balanced order.
C. Results
1. Time to Enter Text
Given the previously reported success of key-prediction, it was hypothesized that the key-prediction guidance semaphore would significantly reduce the average time participants spent entering in each phrase. It was also predicted that the unexpected-key feedback semaphore would reduce time to enter text as well since several usability participants had already mentioned using this feedback as a strategy for selecting word prediction choices in the candidates area. Indeed, using repeated measures analysis of variance (ANOVA), a significant main effect for Semaphore was observed (F_2,1057=4.97, p<0.01). However, Tukey post-hoc pair-wise comparisons only revealed significant differences between the Baseline and Unexpected-Key Feedback (p<0.01) and between the Unexpected-Key Feedback and the Key-Prediction Guidance (p<0.05), but not between the Baseline and Key-Prediction Guidance. FIG. 16A is a chart 1600 that shows the average time spent per phrase for the Baseline, Key-Prediction Guidance and the Unexpected-Key Feedback conditions, which were 26.58, 25.81, and 23.95 respectively.
One possible reason for not observing a significant difference in time to enter text for the Key-Prediction Guidance over the Baseline is that participants were all familiar with the QWERTY layout, unlike the dynamically changing keyboard task described in Magnien. Hence, participants did not experience a need to limit the visual search space.
2. Accuracy
Accuracy was measured in terms of whether or not the participant ultimately typed in the correct phrase. Finer-grained measures of correctness were also looked at, such as the Minimal String Distance Error Rate described in MacKenzie III, which computes the distance between two strings in terms of the lowest number of edit operations to turn one string into the other. However, these measures could not be used since some of the errors made by users were those related to pressing a word prediction or disambiguation candidate, which swaps the typed text completely. In this case, one keystroke engenders a lot of errors. Hence, finer-grained measures could be unreliable.
It was hypothesized that both semaphores would result in higher accuracy since the Unexpected-Key Feedback is immediately informing users of errors and the Key-Prediction Guidance is leading users to the correct spelling. However, in analyzing the data, a main effect for Semaphore was not observed. All three conditions obtained accuracies between 0.12 and 0.14 with fairly wide standard deviations. This deviates from previous results described in MacKenzie II in which Key-Prediction Guidance reduced error rates for eye-typing, although it is hard to compare across modalities.
3. KSPC
It was also hypothesized that the two semaphores would reduce keystrokes-per-character. Indeed, a significant main effect was observed for Semaphore (F_2,1057=6.74, p<0.001), with significant Tukey post-hoc differences between the Baseline and the Unexpected-Key Feedback (p<0.05) and between the Baseline and the Key-Prediction Guidance (p<0.05). FIG. 16B shows a chart 1610 displaying the average keystrokes-per-character for the Baseline, Key-Prediction Guidance and the Unexpected-Key Feedback conditions, which were 1.22, 1.10, and 1.09 respectively.
Assuming that participants accurately hit only the letter they are supposed to and not any of the word prediction candidates, the baseline KSPC for a QWERTY layout keyboard is theoretically 1 according to MacKenzie III. Note that all three conditions had a KSPC above 1, indicating that typing on a small 3.5 inch resistive screen with fat fingers is not easy, even with the word prediction and disambiguation candidates.
4. Number of Backspaces
The number of backspaces was measured as a way to tease apart intentionally corrected mistakes from KSPC, which encompasses that as well as keystrokes for selecting word prediction candidates. It was hypothesized that Unexpected-Key Feedback would reduce the number of backspaces by alerting users to a mistake before they enter more characters to the end of that mistake, and that Key-Prediction Guidance would also reduce the number of backspaces by leading users to correct spelling. Indeed, a main effect was observed for Semaphore (F_2,1057=7.26, p<0.01) but with just a significant Tukey post-hoc difference between the Baseline and Unexpected-Key Feedback (p<0.001). FIG. 16C is a chart 1620 that shows the average number of backspaces for the Baseline, Key-Prediction Guidance and the Unexpected-Key Feedback conditions, which were 3.54, 2.79, and 2.29 respectively.
5. One Versus Two Hands
During the process of running the experiment, an interesting trend was noticed; those participants who chose to perform the entire study with two thumbs tended to make less use of the word prediction and disambiguation candidates. Because participants had to maintain the same way of holding the device throughout the experiment, incoming subjects were asked thereafter to use either two hands (i.e., two thumbs) or one hand for input. No participant held the device with one hand and typed with the thumb for the entire experiment, though during practice some participants certainly tried out that position. All one-handed participants used one hand to hold the device and another to enter text with their index finger.
While it would have been ideal to run a within-subjects, 3×2 factorial design experiment with the OneVsTwoHands added as another independent variable, because fair amount of data had already been collected, between-subjects analysis of the results was conducted, making sure that both one-handed and two-handed users were represented in all possible orderings of the three conditions studied. Hence, additional statistical analyses treating OneVsTwoHands as an additional factor in a univariate ANOVA was performed.
For time to enter text, a significant main effect for OneVsTwoHands was observed (F_1,1057=5.66, p<0.05), as well as a significant interaction effect with Semaphore (F_2,1057=3.68, p<0.05). FIG. 16D is a chart 1630 that shows a breakdown of the previous time taken to enter text result in terms of the number of hands. When using two hands, participants do not seem to be reducing time to enter text in either the Key-Prediction Guidance or the Unexpected-Key Feedback conditions, as shown by the bars 1632, 1633, and 1634 for Baseline, Key-Prediction Guidance, and Unexpected Key Feedback, respectively. However, when participants used one hand, the differences with the Baseline become more accentuated, as shown by the bars 1636, 1637, and 1638 for Baseline, Key-Prediction Guidance, and Unexpected Key Feedback, respectively. Although significant main effect for OneVsTwoHands on KSPC (F_1,1057=2.77, p<0.01) and number of backspaces (F_1,1057=5.46, p<0.05) was observed as well, the same kind of interaction effect with Semaphore was not observed for any dependent variable other than time to enter text. Hence, for KSPC and number of backspaces, the significant post-hoc differences observed previously recur, regardless of the number of hands used.
6. Final Questionnaire
After the experiment, users were asked to pick their favorite condition and to rank-order which of the two semaphores they would leave on by default. This was done to confirm qualitative findings from the usability study. Thirteen of the eighteen participants found the Unexpected-Key Feedback condition to be the most favorable. No one picked the Baseline. With respect to rank-ordering, eight of the eighteen participants said they would leave on the key-prediction feedback semaphore and nine of the eighteen said they would leave on the unexpected-key feedback semaphore. The fact that only nine, and not thirteen, participants said they would leave it on implies that although some people found the semaphore useful, it is disruptive enough to them that they would turn it off by default and turn it on as needed.
D. Discussion and Design Implications
From conducting the usability study, it was observed that visual cues for semaphores cannot grab too much attention. If they do, users generally dislike them. Toning down the semaphores allows users who want to use the visual cues, to mentally grab hold of them, and those who do not, to ignore them. As such, it is desirable that semaphores be designed to be very visually subtle.
Thus, these experimental results indicate that both the unexpected-key semaphore and the key-prediction guidance semaphore have the ability to improve text entry performance.

XIII. Example Alternatives and Combinations

Any of the methods described herein can be performed via one or more computer-readable media (e.g., storage or other tangible media) comprising (e.g., having or storing) computer-executable instructions for performing (e.g., causing a computing device to perform) such methods. Operation can be fully automatic, semi-automatic, or involve manual intervention.
Having described and illustrated the principles of our innovations in the detailed description and accompanying drawings, it will be recognized that the various embodiments can be modified in arrangement and detail without departing from such principles. It should be understood that the programs, processes, or methods described herein are not related or limited to any particular type of computing environment, unless indicated otherwise. Various types of general purpose or specialized computing environments may be used with or perform operations in accordance with the teachings described herein. Elements of embodiments shown in software may be implemented in hardware and vice versa.
In view of the many possible embodiments to which the principles of our invention may be applied, we claim as our invention all such embodiments as may come within the scope of the following claims and equivalents thereto.

Claims

1. A method, comprising:

receiving first input data comprising one or more input words from a keyboard;

providing one or more feedback semaphores, wherein the feedback semaphores are operable to alert a user using the keyboard that one or more suggestion candidates are available;

automatically displaying the one or more suggestion candidates, wherein the suggestion candidates are each associated with an alternative word for one or more of the input words;

automatically receiving a touch screen input selecting one of the suggestion candidates; and

modifying the first input data using the alternative word associated with the selected suggestion candidate.

2. The method of claim 1, further comprising:

based on the first input data, generating one or more alternative words for one or more of the input words;

auto-correcting the first input data by automatically replacing the input word with one of the alternative words; and

wherein the providing the feedback semaphores occurs immediately after the auto-correcting.

3. The method of claim 2, wherein:

the first input data comprises a character designated as a delimiter for the input word; and

the providing the feedback semaphores occurs based on receiving the delimiter character in the first input data and on word probability data associated with the input word.

4. The method of claim 1, wherein the feedback semaphores include one or more of the following: playing a sound, generating haptic feedback, highlighting one or more keys on the keyboard, highlighting a background area of the keyboard, and highlighting a suggestion candidates area.

5. The method of claim 1, wherein:

the feedback semaphores include highlighting one or more delimiter keys on the keyboard; and

the receiving a single touch screen input comprises receiving a key press for one of the highlighted delimiter keys.

6. The method of claim 1, wherein the feedback semaphores are not provided until the input word is designated incorrect using a dictionary, a common speller application programming interface, or an input history data source.

7. The method of claim 6, wherein:

the providing the feedback semaphores occurs based on receiving the delimiter character in the first input data.

8. The method of claim 1, wherein the providing the feedback semaphores does not occur until a delimiter key for the input word is received in the first input data.

9. The method of claim 1, wherein:

the keyboard is a touch screen keyboard; and

the receiving text input data further comprises playing a randomly selected keypress sound for a character of the first input data, wherein the keypress sound is selected from a group including at least two or more keypress sounds.

10. The method of claim 1, further comprising:

receiving typing speed data for at least a portion of the first input data; and

based on the typing speed data, selecting one or more feedback semaphores from a group including at least two or more of the following: a tooltip balloon, audio feedback, haptic feedback, highlighting one or more delimiter keys, highlighting the keyboard, and wherein the one or more feedback semaphores are the selected feedback semaphores.

11. The method of claim 1, further comprising:

detecting whether the user is using the keyboard with one hand or two hands, and;

based on the detecting, selecting one or more feedback semaphores from a group including at least two or more of the following: a tooltip balloon, audio feedback, haptic feedback, highlighting one or more delimiter keys, highlighting the keyboard, and wherein the one or more feedback semaphores are the selected feedback semaphores.

12. A computer-readable storage media storing computer-readable instructions that when executed by a computer cause the computer to perform the method of claim 1.

13. A computer-readable storage media storing computer-readable instruction that when executed by a computer cause the computer to perform a method, the method comprising:

receiving text input data comprising at least one word using a keyboard coupled to a touch screen;

detecting that an event has occurred based on the text input data and based on the detected event, automatically displaying on the touch screen one or more suggestion candidates for the at least one word based on the text input data and one or more candidate sources, wherein each of the candidates is associated with at least one next word designated as likely to follow the at least one word;

receiving a single touch screen input selecting one of the suggestion candidates; and

based on the single touch screen input, automatically modifying the text input data by adding the at least one next word associated with the selected suggestion candidate to the text input data.

14. The computer-readable storage media of claim 13, wherein the event is detected based on a measured keystroke latency exceeding a threshold value.

15. The computer-readable storage media of claim 13, wherein the event is an auto-correction of the text input data or receiving a delimiter key in the text input data.

16. The computer-readable storage media of claim 13, further comprising automatically displaying on the touch screen one or more replacement suggestion candidates for the at least one word based on the text input data and one or more candidate sources, wherein each of the replacement suggestion candidates is associated with at least one replacement word for the at least one word; and

wherein the displaying includes displaying the replacement suggestion candidates in a manner that distinguishes the replacement suggestion candidates from the suggestion candidates.

17. The computer-readable storage media of claim 13, further comprising:

immediately after the automatically adding the at least one word, displaying one or more suggestion candidates for the text input data, which includes the selected next word; and

repeating the automatically modifying the text input data.

18. A mobile device, comprising:

one or more processing units operable to execute computer-executable instructions for text entry and correction;

one or more memory units coupled to the processing units;

one or more touch screens coupled to the mobile device configurable to have a text display area, a suggestion candidates area, and a touch screen keyboard area, wherein the text display area, the suggestion candidates area, and the touch screen keyboard area occupy distinct, non-overlapping areas of the touch screens, and wherein the one or more touch screens are operable to receive touch input over at least a portion of the touch screen keyboard area and the suggestion candidates area;

storage for storing the computer-executable instructions for text entry and correction using:

a text input module for receiving text input using the touch screen keyboard and displaying at least a portion of the text input in the text display area;

a candidate generation module for generating one or more suggestion candidates comprising one or more replacement suggestions and/or one or more next word suggestions for an input word of the text input;

an indicator generation module for producing one or more indicators that can notify a user that the suggestion candidates are available;

a suggestion presentation module for presenting the suggestion candidates associated with the input word in the suggestion candidates area and receiving touch screen user input for selecting one of the presented suggestions, wherein:

if the selected suggestion is a replacement suggestion, replacing the input word with the selected suggestion; and

if the selected suggestion is a next word suggestion, adding the selected suggestion as a next word subsequent to the input word in the text input.

19. The mobile device of claim 18, wherein the indicator generation module and the suggestion presentation module are configured to not produce indicators and to not present suggestion candidates, respectively, until one of the input words is designated incorrect by a dictionary, a common speller application programming interface, or an input history data source.

20. The mobile device of claim 18, wherein the indicator generation module and the suggestion presentation module are configured to not produce indicators and to not present suggestion candidates, respectively, until a delimiter character is received immediately after the input word.