KR100776275B1 - Touch-typable devices based on ambiguous codes - Google Patents

Abstract

The design of the typing device 6005, in particular the touch type typing device 6005 in which the ambiguity symbol 6001 is envisioned, shows many ergonomic problems. Here are the solutions to these problems. This development, given the design constraints such as size, shape, and processing capacity of the device, as well as the existing constraints such as problems with alphabetical or qualitative arrays, is mainly due to the use of ambiguity 6001 and touch typed typing. Teaches how to select an ambiguity from a class of practical best ambiguities. Touch typed devices are, for example, computers, telephones, pages, personal digital assistants, smart cards, television set top devices and information appliances.

Description

Touch-typable devices based on ambiguous codes}

The present invention relates to the design of a touch type typing device, the use of a touch type typing device used in computers and telecommunications, in particular to a touch type typing device based on an ambiguity and a substantially optimal ambiguity for a strongly touch type typing. .

Since typewriters were invented more than 100 years ago, keyboard engineering has been an active area of research and development, and as a result there have been many competitive designs. With the growth of personal computers and telecommunications, the number of keyboard designs has increased as designers have embraced a variety of constraints to take advantage of opportunities for new technologies to be announced. Nevertheless, significant modifications to the preceding keyboard design are not due to a variety of constraints and opportunities, but rather to keyboard designers not fully understanding the constraints inherent in the problem to be solved. It also lacked a common and effective way to optimize for these constraints. Thus, the current situation in the art is that there are an excessive number of partial solutions. These diseases are solved by the keyboard design method according to the present invention. In order to describe the various aspects of the present invention, optimization methods are applied to the design of each of the various device embodiments that are desired with practically optimal solutions to a given set of design constraints.
The present invention relates to a touch type typing device. Like typing on an instrument, touch typing is a manual technique that is difficult to learn. Once learned, it is difficult to change the acquired motor pattern. This difficulty is a significant constraint on keyboard design. The superior status of known Qwerty keyboards (and their close variants, eg Azerty keyboards in France) is to deeply plant and learn excessively the motive patterns associated with touch typing. Thus, the Qwerty keyboard was widely based, thus becoming a barrier to suppressing improved keyboards such as the Dvorak keyboard. Indeed, such vending machines were limited to limited users. Due to the large number of keys, the Qwerty keyboard is unsuitable for smaller typewriters that are handheld. The advent of these devices provided keyboard designers with a viable market niche. In a growing market segment, new designs will remain superior because of the inherent tendency of repetitive motive patterns to be fixed, even if more appropriate designs appear later. From this perspective, keyboard designers have considerable responsibility to ensure that future keyboard users do not impose suboptimal design.

delete

There are two main approaches in the prior art to reduce the number of inputs required to encode a given set of symbols. 1) Coding method-In this method, each symbol is encoded by activating a combination of input means. 2) ambiguous codes-in this way the combination of symbols is encoded by the respective input means. Coding schemes were not practically successful because they were difficult to learn, and few people wanted to invest the time required. Thus only the ambiguities, or the ambiguities combined with the coding scheme, are promising as a solution to this problem.

Therefore, it is an object of the present invention to provide a method for substantially optimal ambiguity design for a typing device.

The present invention also provides a method for the design of a strongly-touch-typable ambiguous code for a typing device.

The present invention also provides a keyboard suitable for touch-type typing on full-size and reduced-size keyboards.

The invention also makes it possible to send an alphanumeric (alphabetically and numerically) message from a regular telephone or pager of a transponder / receiver to other devices of its class without human intervention and at low cost.

The present invention also provides a personal digital assistant for typable touch.

The present invention also provides a keyboard that the driver can type without unnecessarily distracting the driver's attention.

The present invention also provides for a typeable communication device that is low in manufacturing cost and operates in a standard telephone communication system.

Still another object of the preferred embodiments of the present invention is to partially retain the frame of the conventional keyboard in the frame of the new keyboard so that the typing skills of the touch typed hitters trained on the conventional keyboard can be easily converted to the new keyboard.

It is a further object of the present invention to provide a general method of producing an ambiguity having a substantially minimum lookup error rate.

The present invention also provides a general method of generating an ambiguity code having a substantially minimum query error rate.

The present invention also provides an apparatus for reducing damage caused by typing.

The present invention also provides a handheld computer device that can be folded twice.

The present invention also provides a one handed keypad suitable for performing in a handheld computer.

The present invention also provides a one- and two-handed keypad suitable for performing on a handheld computer or desktop keypad.

The present invention also provides a Qwery-like keyboard.

The present invention also provides a coding keyboard that can be easily learned.

The present invention also provides a synergistic hybrid of a coding keyboard and an obscure keyboard.

The present invention also provides a touch type typing-oriented querying mechanism for an input device in which an ambiguous code is implemented.

The present invention also provides a touch type typing-oriented unambiguation mode for a typing device with an ambiguity implemented.

The present invention also provides a complex coding / ambiguity code that is fully compatible with standard telephone keyboards.

The present invention also provides an ergonomic assignment of symbols to modes.

The present invention also provides a substantially transparent touch typeable interface for a typewriter device composed of a touch screen.

The present invention also provides an optimization for a series of natural languages.

The invention also uses a device that can be typed using a one-handed device while reducing scanning time.

Objects other than these objects of the present invention will be described in detail below.

The detailed description of the preferred embodiments of the present invention will be described with reference to the following briefly described drawings.

Figure 1 shows an overview of the optimization for developing the typing device according to the present invention.

2 shows a flow chart for constructing a device based on an ambiguous code for a strong touch type batter.

3 shows a flow chart for satisfying at least one constraint and constructing an ambiguity optimized for these constraints.

FIG. 4 shows a particular embodiment flow chart for the method of FIG. 3 using a random search optimization method.

5 shows a lookup error probability distribution for randomly selected ambiguous codes for several selected keys.

6 shows the query query probability distribution for randomly selected ambiguous codes for several selected keys.

7 shows a flow chart for directed random walk optimization.

8 shows a flowchart for constructing an ambiguous code for a strong touch type batter.

9 shows a plot of lookup error rate versus key number for any selected substantially optimized ambiguous code.

10 shows a plot of query query error rate versus number of keys for any selected substantially optimized ambiguous codes.

11 shows a lookup error rate versus query lookup error rate for some substantially optimized ambiguities in a range of key numbers.

FIG. 12 shows a table relating, for several different optimization methods, the grade of strong touch typability and the number of keys required to achieve that grade.

13 shows a flowchart of a method of synthesizing an encoded symbol.

FIG. 14 summarizes the embodiments and the main features thereof in order to help the reader understand the unity of the invention, with respect to the many device embodiments necessary to clearly and clearly point out the broad scope and various aspects of the invention. Show the table.

FIG. 15 shows an embodiment of a smart card with 16 keys that contributes to encoding a character code.

16 shows an embodiment of a smart card with nine keys that contributes to encoding a character code.

17 shows the keyboard embedded in the handle.

18 shows a telephone with a sign substantially optimized for ten keys.

FIG. 19 shows reduced-ambiguity ambiguity ordered alphabetically as an example applied to a mobile phone.

FIG. 20 shows a qualitative-like keyboard optimized for a lookup error rate and a query query error rate in consideration of a character arrangement in each column of the quartie keyboard.

Fig. 21 shows another qualitative-like keyboard.

22 shows an ambiguous keyboard implemented in a standard numeric keypad arrangement.

Figure 23 shows an ergonomic touch type batter-oriented disambiguation mechanism.

24 illustrates a method of answering a query query in a touch type batter oriented manner.

25 shows an embodiment of a keyboard for one hand designed to preserve typing skills between one and two hand keyboards.

FIG. 26 shows an embodiment of a keyboard for two hands designed to preserve typing skills between one and two hand keyboards. In this case the two-handed keyboard is weighted for the maximum similarity of typing motions between the two cases.

FIG. 27 shows an embodiment of a keyboard for two hands designed to preserve typing skills between one- and two-handed keyboards. In this case the two-handed keyboard is given an equal weight between the two hands.

28 shows an integrated mouse / keyboard.

29 is a plan view of an unfolded state of the information device which can be folded twice;

30 is a bottom view of the non-folding information device that can be folded twice;

Fig. 31 shows the folded state of the double collapsible information device with the additional function exposed.

32 shows a twice folded state of a foldable information device with another additional feature exposed.

33 shows a separate state of a foldable information device, allowing two-handed batters.

34 shows a typical personal digital assistant with a touch screen.

35 shows a typical personal digital assistant with a potentially transparent keyboard.

36 (A), (B) and (C) show three modes for a 16-key keyboard.

37 shows a standard telephone layout.

38 shows a complex coding / unsigned keyboard implemented in a telephone.

FIG. 39 shows the distribution of lookup error rates and query query error rates for all complex coding / ambiguities of a particular structure, compared to the lookup error rate and query query error rate of a standard ambiguity code.

40 shows a flow chart for generating an ambiguous code for a multi-level strong touch typing.

FIG. 41 shows a flow chart for generating a particular embodiment of a multi-grade strong touch ambiguity code.

FIG. 42 shows a typing apparatus suitable for performing the multi-level ambiguity code of FIG. 41.

FIG. 43 shows the apparatus of FIG. 42, operative to display the first class of a multi class ambiguous sign.

44 shows the second class code of a multiple class ambiguity code.

FIG. 45 shows the apparatus of FIG. 42 operative to display a portion of the second class code of a multiple class ambiguity code.

FIG. 46 shows a sequence of operating states of the device of FIG. 42 when used in combination with a multi-grade ambiguous code to hit the word "think."

FIG. 47 shows the operating state of the device of FIG. 42 when used in conjunction with a multi-level ambiguity sign to strike the word "think", as in FIG. In this case, the operation of the visual cache to reduce scan time is shown.
FIG. 48 shows the operation of an unambiguous selector for the sole selection of the predicted character when there are two characters per key.
Fig. 49 shows a general ambiguity structure, which maps an encoded symbol sequence to a decoded symbol sequence.
Figure 50 shows a maximum-2 to maximum-1 ambiguity mapping a coded symbol sequence of maximum length 2 to a decoded symbol sequence of length 1.
51 shows a maximum-2 to maximum-1 ambiguity code, wherein the length-2 sequence consists of a combined coded symbol and a key-assigned coded symbol; The decryption symbols are assigned the same key as the encoding symbol with the index.
52 shows a maximum-2 to maximum-1 ambiguity code, wherein the length-2 sequence consists of a combined coded symbol and a key-assigned coded symbol; The decryption symbols are letters az, and the characters explicitly selected in the character set of each of the keys of the standard telephone keypad are c, e, h, l, n, s, t and y.
Fig. 53 shows the maximum touchable typeable area (shading) in the lookup error and query query error planes.

Definition and basic concepts

This chapter summarizes the definitions and concepts of words that will be used in the following description.

Language: Given a set of symbols, one can construct a sequence of symbols and assign probability to the sequence. Symbol sets, symbol sequences, and probabilities will be referred to here as language. For clarity of discussion and without limiting the scope of the present invention, the language we refer to is a natural language of letters such as English. Although specifically referring to symbols as letters or punctuations, those skilled in the art will recognize that the symbols in this discussion can be separate units of writing and transferred to standard symbols or princes, such as Chinese ideograms. It will be understood that it contains an created symbol, such as the name of a master known in the art .

Keyboard / Input Means: A keyboard is a component of a communication and / or computer device that converts physical movement by an operator into a symbolic sequence. The keyboard consists of at least one input means, which converts a subset of the physical movements to activate the keyboard into a subset of the symbol sequence.

Physical movements used in keyboard manipulation typically appear in the form of the motion of a finger and / or thumb or of a stylus held in the hand. This definition can be extended to other physical movements such as head, tongue or eye movements that can be used for symbol selection in keyboards. By this definition, a device comprising a keyboard will be referred to as a typable device.

Under the definition of "typewriter", it is understood that not only the physical device including the keyboard, but also the typewriter device is embedded and a complete communication system whose scope is defined according to the ambiguity plan dependence. In the case of a typing device where input symbols appear directly on the display, which is a physical part of the typing device, the range of the system is clear, which is defined by the physical perimeters of the device. In a more general case, for example, if a typing device includes a telephone handset that sends information to a central computer, and the central computer decodes or otherwise acts on textual information communicated from the handset, the "typewriter device" It should be understood that it includes a central computer arranged to operate in the required manner in accordance with the software made in the teaching aspect of the present invention.

It will be appreciated that each of the at least one input means configured in the keyboard can be variously physically manifested. An essential feature of the input means is to allow the operator to select a subset from a set of symbols that can be encoded by the keyboard. Based on this understanding, to enhance the understanding of the present specification, the word "key" will be frequently used interchangeably with the word "input means".

Typing: A process of sequentially selecting at least one input means to select a sequence of subsets of symbols from a set of symbols to be encoded by the keyboard. It may also be understood as a type of typing that interprets a collection of drawing-motions into a selection of a subset of a set of symbols using known handwriting recognition software.

Touch type typing: In the process of generating a symbol sequence on a keyboard, it is the process of using kinesthetic feedback mainly or mainly rather than visual or auditory feedback.

Strongly correlated symbols and symbol sequences: It is well known that different letters appear in words with different frequencies. For example, in the previous sentence, the letter e appeared 11 times, while the letter z did not appear at all. This is also true of two pairs of characters, triples of characters, and so on. This is related to the fact that words never occur at the same frequency. The three letter word the is very common in English, but the three letter word zap is not very common. This statistical irregularity can be used in the design of ambiguities. In fact, at least since the Quartie invention, statistical irregularities have been used in keyboard design.

In particular, attention is paid to symbols and symbol sequences whose distribution in a typical text sample is strongly correlated with substantially different symbols or symbol sequences, and such symbols will be referred to as strongly correlated symbols. For example, sometimes a symbol used to indicate the end of a sentence in English or another language. Can be a strongly correlated symbol since the distribution of sentence length is not random in typical text. In Hebrew, the symbol. Is used because some symbols are used for some characters that appear at the end of a word. Is also correlated with special character symbols, and the end of the sentence is correlated with the end of the word.

Reference statistics: The reference statistics of symbol sequences used to measure the correlation between symbols are typically estimated by reference corpus analysis. A reference battle is a huge collection of texts drawn to represent some form of language. As is well known to linguists, there are important and fundamental problems in constructing a corpora to represent the general features of the language, as opposed to the features of a particular class of text or a particular type of author. These problems are outside the scope of the present invention. Here we refer to the entire set of reference statistics collected from the British National corpus analysis, one of the largest existing language wars for English analysis. Choosing verbal warfare is an essential step in gathering results that enable us to compare various methods and embodiments. None of these specific choices should be construed as limiting the scope of the invention. In particular, the choice of verbal substitution of English texts is arbitrary. The same analysis may be performed for other written languages.

Encoding and Decoding: In the United States, the keys on a telephone keypad are sometimes labeled with letters as well as numbers, typically the key corresponding to the number 2 also corresponds to the letters a, b, c and the number 3 The keys correspond to standard character arrays of English characters, such as the letters d, e, and f.

Thus, the sequence of key presses associated with digit sequence 233 also corresponds to the character sequences add, bee and bed, which are English words, as well as to several meaningless character sequences such as cff. At this time, the sequence is considered to have meaning if it is in the reference list of meaningful sequences. Thus, all of these character sequences are associated with the same numerical sequence, whether they are meaningful or not. The keypress sequence 233 will be called encoding (encoding), and the sequence of add, bee, bed, eff and so on will be called decoding (decoding) of encoding 233. When no confusion occurs, "detox" can be used to mean "significant decryption." In this example, the set of symbols used in decoding, the alphabetic characters will be referred to as deciphering (decoding) symbols or simply as symbols if confusion does not occur. In this example, we will refer to the set of symbols used for encoding, the number, as the encoding symbol.

Ambiguous codes: Ambiguous codes are well known in the art. The standard telephone keypad used in the United States has twelve keys, ten of which encode numbers, and some of them, typically eight arranged in alphabetical order, encode three or four letters of the alphabet. This assignment creates an ambiguity that we can call standard ambiguities. This code is abc, def, ghi, jkl, mno, pqrs, tuv, wxyz.

Since several characters are encoded in each key, a disambiguation method should be used to determine which of the several characters the operator intends. In typical applications, for example in voice response systems, the intended character is found by comparing the input sequence and the stored response. If some of the stored responses correspond to the input sequence, the user must present the list of responses to the user from which to select. The order of expressing these choices is arbitrary or depends on the frequency with which each response is a correct response, with responses displayed in decreasing order of frequency.

Standard keyboards: There are three standard keyboards that are practically widely used: the Quartie keyboard and variations close to it, a twelve-key phone pad, typically a seventeen-key numeric keypad and its close variants. It is a unique advantage of the present invention to present a method of hitting keys useful for both standard phones and numeric keypads as well as the specially designed keypads proposed here.

Strong touch typability: A device with a fixed symbol assignment is a device in which the symbol assignment is substantially fixed to the key. Only in connection with these devices, the batter can develop physical reflexes to encode special symbols using special movement patterns. We can say that the typewriter has 1) a fixed array of symbols, 2) is based on ambiguities, and 3) in normal operation, that the touch-type typewriter can obtain text with an accuracy that can be accommodated using the typewriter. In other words, it will be mentioned that the rudder can be typed in a strong touch.

Strong touch typing is a matter of degree; This is a matter of touch type typing which depends on a number of factors, such as some associated with the individual typewriter, some associated with using the typewriter device, and some associated with the structure of the typewriter device itself. Even with a fixed touch typewriter, for example, a given typing device may be typed sufficiently strong for some typing tasks, but not for other tasks.

The accuracy of the text obtained is
Non-ambiguous means,
The situation of using the device, such as when driving or sitting at a desk,
The type of text to be typed, which partially determines the level of accuracy required;
Reference statistics,
The ability of the other person,
Individual preferences, and
Means that draw the user's attention to an unambiguous technique (eg, a speech synthesis technique that speaks a word or queries a query to the user) are more or less involved than a bell or flashing light. It can be understood that it depends on several factors, including distraction.

delete

delete

delete

delete

delete

delete

delete

Like temperature Strong touch typing is a matter of degree, but it is perfectly defined like temperature. Strong touch type typing of the typing device can be quantitatively measured for any user or group of users, once these various factors have been determined, using standard experimental protocols well known to those skilled in the art. Moreover, two aspects of strong touch typing can be measured directly in ambiguity: lookup errors and query lookup errors. Thus, the numerical value of the strong touch type magneticity can be determined solely in relation to the problematic ambiguity, without being directly related to the user.

Like temperature, the strong touch typing has a lower bound. It is clear that a device that requires user intervention after every word or after every three words to be unambiguous cannot be considered to be able to be typed in a strong touch to a hitter who is dedicated to the task. The low boundaries of strong touch typing can be expressed by the continuity of attention. If the user's attention substantially continues to focus on the operation of the unambiguous technique to obtain acceptable text, the device is not typed in a strong touch.

Substantially lower boundaries of strong touch typing are associated with users with average skills in the field of touch typing, which is higher than the theoretically described lower boundaries. In order to apply the numerical value to the progressive concept of strong touch typing, as well as conceptual accuracy and clarity, the numerical value is set to the strong touch typing with lookup errors and query query error values. This numerical specification helps to more clearly point out the differences between the inventive methods and apparatuses and all preceding methods and apparatuses.

The strong touch type ambiguity code is an ambiguity code based on the strong touch type type device.

Feedback device: In a device where the user is involved at various points in decoding symbol sequences generated using ambiguity, the user needs a sensory feedback scheme. Typically this feedback will be in the form of visual representations of the symbols, but the feedback can be in many forms, for example auditory, tactile or even olfactory.

Constraints: Keyboard designs with ambiguities satisfy many constraints. These constraints include: reduced lookup error rate, reduced query lookup error rate, selection of the number of keys to match the desired keyboard size, compatibility with existing vending machines, eg, qualitative vending machines, telephone keypads or numeric keypads, rules of partitioning structure. Conventional conservatives such as sex, anatomical fidelity, preservation of conventional operation, learnability, minimal mode-switching key usage, partitioning structure, one- and two-handed typing compatibility, and alphabetical arrangement. Other constraints include ergonomics of unambiguous techniques, ergonomics for encoding weakly correlated symbols, look-and-feel, and transmission and reception of communications systems using ambiguity symbols. Includes the availability of computer processing resources.

Lookup error: Measures the error performed by an unambiguous technique that systematically selects the most probable (meaning) interpretation from a set of possible interpretations of ambiguous sequences. Therefore, the lookup error rate of a sign is the sum of the reference probabilities for all possible decodings, not the most probable decoding of the ambiguity sequence. In the case of word-based deambiguity, these sequences begin and end with the symbol space. That is, words. Lookup errors are the probability that most-likely deciphering is incorrect. Lookup errors are expressed in terms of convenience, and lookup error rates are expressed in units of words per lookup error. The lookup error rate is the inverse of the lookup error probability.

Query error: all (significant) decryptions-this is the sum of the reference probabilities for all (significant) decryptions, not the only (significant) decryptions. This is the probability that a given word will have one or more meaningful interpretations, so a query query must be made to select which of the decryptions the user will use. The inverse of the query error is the query query error rate expressed in units of words per query query. The query query error rate is the average number of words entered between query queries.

Substantial optimality: A code will say that if one code is among the most suitable codes for a characteristic given the other constraints imposed on the code, the code is substantially optimal with respect to the characteristic. For example, a sign on twenty keys may have a lower lookup error rate value than a sign on two keys. However, the sign on two keys may be substantially optimal with respect to the lookup error rate given the constraint that the sign is on two keys. The substantially optimal restricted code will be defined as the substantially optimal sign at the same time for each collection of constraints. Such constraints include, but are not limited to, the number of keys, lookup error rate, and query query error rate. For these three constraints, the pairs of constraints are correlated. The lookup error rate tends to increase with the query query error rate, and both the lookup error rate and the query query error rate tend to increase as the number of keys decreases. When this given constraint is the only optimization constraint, the best possible value for a given constraint may be better than the best possible one that can be obtained when other constraints are also optimized. Therefore, the appropriate constraints for a given design should be determined and weighted for their importance as the first step in the optimization method taught by the present invention.

It will be emphasized that the optimality of the ambiguity cannot be discussed absolutely, but must be evaluated in terms of a set of reference statistics for the language being encoded. In fact, given any ambiguity, it is possible to construct a series of statistics and the codes are optimized for the constructed statistics.

Given a set of reference statistics, we can estimate the optimization of a given code from an experiment consisting of random code generation and will be discussed in more detail below.

Unambiguous method: The actual optimality for ambiguity is well defined only in relation to the chosen unambiguous method. A substantially optimal sign for one unambiguous method may not be substantially optimal for another unambiguous method.

At least two unambiguous methods are well known in the art. There are word-based unambiguous and context-based unambiguous. In word-based unambiguization, a list of words is used to select this, along with their probabilities, among other decodings of the encoding given in the ambiguities. For example, all words in the list that are meaningful interpretations of a given encoding will be compared so that the word with the highest probability is selected. Context-based deambiguation is similar except that the list contains text fragments of arbitrary size and has a probability for those paragraphs.

Both word-based unambiguous and context-based unambiguous are special cases of a more general framework, which we will call sequence-based unambiguous, where the sequence of encoding symbols and decoding symbols is Databases are associated with probabilities, and unambiguity is affected by this database reference. It is noted that the space symbol that defines a word boundary in a language such as English is no different from any other symbol for the purposes of this discussion. A sequential list and sequential probabilities that include space symbols can be defined, and thus can extend beyond word boundaries. Further, one may define sequences containing wildcard symbols, and define sequences of lists containing arbitrary sequences, which sequences may or may not correspond to words in a language. In this way arbitrarily complex language expressions can be established and used in the obfuscation method. For example, a syntactic semantic relationship between sequences can be introduced to resolve conflicts between possible interpretations by means of decodes for obscurely encoded sequences of code symbols. . Clearly, unless otherwise specified, the specification focuses on unambiguity based on known words. The method taught by the present invention does not rely on word-based deambiguation; It will be understood by those skilled in the art that other non-ambiguous methods can be used.

Partition: A partition of integer n is a set of integers, so the sum of the elements of that set is equal to n. In general, many divisions are possible for a given integer. For example, the integer 5 has a 3: 2 division and also has a 2: 2: 1 division. Algorithms for generating all divisions of integers are well known to those skilled in the art. Most prior art codes use possible even-as-possible partitioning. That is, in partitioning, the number of characters per key is equal to the extent possible in a given number of keys related to the number of characters to be encoded. This choice, which is further extended below, is a sensible choice for some constraints, and it can be sub-optimal for others.

There are two kinds of ambiguities in which exclusive rights are claimed. 1) a strong touch type batter, and 2) a substantially optimal ambiguity code. Ambiguity may be substantially optimal but cannot be typed in a strong touch, or may not be substantially optimal but may be typed in a strong touch, may not be substantially optimal, may not be typed in a strong touch, or It is practically optimal and can be typed in a strong touch.

The present disclosure begins by pointing to how to make an ambiguous sign in both of these types, and determining which kind the code belongs to. It also describes how to use these two types of codes to make typing devices and how these codes are used to solve the various design problems faced by typing device designers.

The best mode for practicing the present invention depends on the design constraint constellation being optimized according to the teachings of the present invention. Thus, some specific situations that are actually relevant and useful will be selected to describe the scope of the methods and apparatus that the present invention teaches.

The scope of the machine that can be made by those skilled in the art in accordance with the teachings of the present invention extends significantly beyond what is specified in the preferred embodiments presented herein. Various extreme or special cases of design constraints are solved in the embodiments. In accordance with the teachings implemented in these cases, it will be apparent to those skilled in the art how to properly combine the features to solve intermediate or complex design problems.

One embodiment is optimized only for lookup error rates. This embodiment is implemented for devices with limited memory and computer processing power, such as smart cards. With such a device, computer resources cannot support complex query lookup techniques for user activation (interference) in an unambiguous process. Thus, such an apparatus applies one of the simplest possible unambiguous techniques, which systematically selects the most probable decoded symbol sequence corresponding to a given encoded symbol sequence.

Another embodiment is optimized only for query query error rates. This embodiment is implemented for use by a driver of a vehicle, for example a motor vehicle. Although it is possible for computer processing to support complex query queries techniques, the use of these techniques should be kept to a minimum in order to prevent the driver from distracting from driving.

The next embodiment is to provide a telephone keypad optimized for both lookup error rate and query query error rate, which can be compatible with the arrangement of the standard telephone keypad.

Another embodiment is optimized for conventional preservation constraints: alphabetical order (array) maintenance. The characters are assigned to standard touch-tone keypads in alphabetical order. It is possible to reduce the lookup error rate and query inquiry error rate by optimizing the partitioning while maintaining the alphabetical order of the conventional telephone keypad.

Optimization over partition shows an additional embodiment of preserving the prior qualitative keyboard layout as well as possible for vending machines with substantially optimal query lookup and lookup errors.

Another embodiment showing a keyboard design that corresponds as well as possible to a conventional design is a keyboard based on an ambiguous code preserved key layout and numeric label of a numeric keypad.

In many applications, there is a need for keyboards that can be ergonomically operated in both obscure and unambiguous fashion. For this purpose, it is desirable to choose an ambiguity sign over a number of keys in which the number of symbols to be coded is distributed, which corresponds to an almost equal distribution of the keys possible. Particularly preferred is the number of keys which are ½ of the symbol number. The importance of this preferred choice is shown by the following examples.

Another embodiment shows how the keyboard is optimized for cross-platform compatibility. In this embodiment, the two keyboards, one-handed keyboard and two-handed keyboard, convert quickly so that the touch-type typing movement used to manipulate one keyboard is switched uninterrupted to the touch-type typing movement used to manipulate the other keyboard. It is designed to be manipulated possibly. This keyboard has the additional advantages of having the potential to reduce hitter damage, as well as other objectives or advantages described in the detailed description.

The above-mentioned embodiments mean that using different keyboards has different kinds of optimizations, and since some users need keyboards for some other use, various different solutions coexist in a single device. This means that several mechanisms must be provided. A surprising solution to this problem is made possible by the small typing device which can be achieved with ambiguity, which is the twice-foldable personal digital assistant described in the embodiment.

The embodiments discussed so far include both hardware and software implementations. However, many of the objects of the present invention can be achieved entirely or primarily using software solutions. An example software solution will be described in detail that will show how the features of the hardware are specified to use the appropriate software to achieve the objects of the present invention.

The last embodiments are a combination of two different approaches for producing keyboards with fewer keys: a coding scheme and an ambiguity scheme to have a synergistic effect.

First, by the ergonomic construction of the coding pattern combined with the optimization of the lookup error rate and query query error rate, an ambiguity on n keys can be made to act like an ambiguity on m keys that are substantially larger than n keys. Shows that there is. In particular, when applying this method to standard ambiguities, the eight character keys of the standard ambiguities will have substantially optimal coding characteristics without coding over the 13 keys. Descriptions of how to extend the ambiguity creation discussed in reference to English can be extended to other languages. Specifically, this discussion is performed with respect to this embodiment, but the descriptions generally apply to all embodiments.

Second, combining the method of division and acquisition with the method presented in the preceding embodiment, the number of input means is further reduced, in which four input means are ambiguously encoded with 16 elements. It is used to manipulate it. The handheld device that implements this code is selected by the number 4 so that it can be manipulated using the fingers and thumbs of the hand holding the device.

Overview of the operation of the strong touch typewriter device: FIG. 2 shows an overview of the strong touch typewriter device operation based on ambiguity. Such an apparatus has an input means and generates a sequence of encoding symbols in accordance with user activation 140. 141. In step 142, a strong touch type batter code pair is encoded symbol sequence database paired with a decoding symbol sequence. It is used to map the encoding symbol sequence to the decoding symbol sequence with reference to. Decoded symbol sequences may optionally be output in electronic form for processing, transmission or storage on or off the screen for direct viewing by users of the device 143.

It should be noted that the ambiguity of the device of FIG. 2 can satisfy other constraints as well as strong touch typing.

Construction of substantially optimal codes: A methodological step of optimizing ambiguity codes for a series of constraints is described with reference to FIG. In summary, these steps are:

2000 Selection of a statistically correlated series of decoded symbols to be represented by an ambiguity, consisting of the following substeps:
Select a set of 2007 reference statistics,
Statistical Correlation Analysis of Symbols for Statistics Selected in Step 2008

delete

delete

2001 Selection of Deambiguous Methods.

2002 Select the number of encoding symbols.

2003 Selection of constraints on how codes should be practically optimal.

2004 Weighing the importance of the constraints selected in step 2003.

Selection of 2005 Optimization Methods.

Application of the optimization method selected in the 2006 stage 2005, thus creating a practically optimal ambiguity code.

Steps 2000 through 2003 may possibly be applied in any order, and applying one of these steps will affect the choices in the other steps. Detailed description of the application of these steps is given below.

Selecting a set of statistically correlated symbols that must be represented by an ambiguity symbol 2000. This step is a sub-step 2007 that selects a set of reference statistics and a sub-step 2008 that analyzes the statistical correlation of the symbols to the selected statistics. Consists of. The goal of these steps is to identify those symbols that can be expressed ambiguously. All unambiguous methods use the correlation between symbols to predict which sequence of decoding symbols should be associated with the sequence of certain encoding symbols. If the decoded symbols are distributed randomly across all the texts to be encoded, it cannot be represented by an ambiguity sign. This is because no prediction can be made for randomly distributed symbols. In general, for the natural language, the symbols used to encode the language (eg, letters in English, phonetic letters in Chinese) are statistically sufficiently correlated, thus providing an effective ambiguity for these symbols. The sign can be designed. There may also be other symbols, such as punctuation symbols, that are statistically significantly correlated with the letters or phonetic characters used to record each other and language as letters. The details of steps 2007 and 2008 depend on the natural language being expressed. Statistical correlation analysis of symbols used in written natural language is a well-known technique for linguists.

Steps for selecting an unambiguous method 2001 . As already mentioned, there are at least two well-known methods of unambiguization, context-based unambiguity and word-based unambiguity. These methods all use a statistical context of symbols to predict which sequence of decoding symbols is determined corresponding to the sequence of a given encoding symbol. Even with higher levels of information about languages such as syntax and semantics, both context-based and word-based methods can increase. It is an object of the present invention to construct an ambiguity code that optimally selects a decoding sequence corresponding to each coding sequence in relation to the selected unambiguous method. Therefore, the details of the chosen unambiguous method can affect the details of the design of the ambiguity being designed. Other unambiguous methods will also be discussed, but this method will be described for non-ambiguous selection based on words as an unambiguous method.

Selecting the number of encoded symbols 2002 . Choosing the number of coded symbols is very important for the design of typewriters based on ambiguities. This choice is made taking into account various factors, including the size of the typing device and an acceptable level of ambiguity. These factors and their interplay will be best described with specific examples, which are discussed later in this specification.

2003. Selecting constraints for which the sign is substantially optimized. One major aspect of the present invention is the discovery and definition of various constraints that determine the quality of typing devices based on ambiguity. These constraints include strong touch typing, lookup error rate, query lookup error rate, anatomical fitness, conventional gesture preservation, preservation of conventional placement structure, partition structure, cross platform compatibility, regularity of placement, scan rate. Depending on the application, one or more constraints may be related to the typing device design.

Step 2004 adds importance to the constraints selected in step 2004 . When one or more constraints relate to the typing device design, a weight for the importance of the constraints must be determined. It is very rare for constraints to be optimized separately, such as when the same optimization for a given constraint optimizes for other constraints.

Steps to select an optimization method 2005 . Two optimization methods, random selection and managed random behavior, will be discussed in more detail below. Generally, random selection is easier to perform, but managed random behavior produces better codes. These two methods represent a large class of methods that may be suitable for a given typing device design. In some cases, for example, in the coding / ambiguity apparatus considered below, the number of examined codes is small enough to be able to thoroughly check all of them.

To apply the optimization method is selected in step 2005. In substantially the step of generating the optimal ambiguous code 2006 along. Notwithstanding the optimization method selected in step 2005, a technique must be used to apply the method for generating an optimal ambiguity code substantially. In particular, when an optimal condition is needed for various constraints at the same time, it is desirable to consider the constraints individually first, so that the quality of the code that can be ultimately achieved can be estimated. This prediction can be very useful in adjusting the optimization process and will be described in more detail.

Random search: The basic method for finding codes with good characteristics is to randomly select the symbols, test their properties, and choose the one with the best characteristics. With exhaustive enumeration, where all the symbols in the candidate set are tested, it is generally not a viable option because the number of symbols is too large to test within a reasonable amount of computer time.

The random survey provides a benchmark that can measure the usefulness of other methods for sign selection. Suppose you are given a set of constraints and their weights. By generating additional ambiguities at random, one can estimate the practical optimum of these constraints and the first ambiguity for these weights. If the number of random attempts is small, it is possible to find a sign of the same or better value as the first sign for a given constraint, in which case the first sign is not substantially optimal.

On the other hand, if a substantially large number of trials are required to produce a code having a value equal to or better than the first code, and if a good code exists, then the first code is substantially optimal.

Referring to FIG. 4, a method for denying that the candidate ambiguity is substantially optimal is described in detail.

The steps outlined are as follows:

3000: Determine a set of related constraints that defines a set of suitable codes including the candidate code.

3001: Determining a set of constraints in which the candidate code may be substantially optimal.

3002: Randomly select a subset of signs from the set determined in step 3001.

3003: For each of the constraints determined in step 3001, evaluate each of the symbols selected in step 3002.

3004: For the constraint selected in step 3001, compare the value of the candidate sign with the values obtained in step 3003. If any values obtained in step 3003 are more optimal than the values of the candidate code, the assumption that the candidate code is substantially optimal can be negated.

Details of these steps:

Step 3000. Determine a set of related constraints that define a set of candidate symbols. The set for which the actual optimality of the candidate code is to be evaluated must be properly defined. Possible related constraints are: allowance of specified arrays such as number of encoding symbols, subdivisions, alphabetical arrangement. Each of these constraints limits a set of codes that can be properly compared with the candidate code.

Step 3001 , determine a set of constraints in which the candidate code may be substantially optimal. Constraints that may relate to the analysis of candidate codes are: lookup error rate, query lookup error rate, acceptance of specified arrangements such as alphabetical arrangement, acceptance of regular placement, and anatomical suitability.

Once steps 3000 and 3001 are performed, a distribution of sign characteristics for a set is defined, and this distribution can be sampled arbitrarily. 5 shows an example where the sets of codes are defined to have 1) possible even division, and 2) number of encoded symbols specified by 7, 9, 11, and 13. This definition completes step 3000. Then, it is determined that the lookup error is the only relevant constraint. This decision completes step 3001. In addition, these steps determine the distribution of shapes that can be determined by random sampling. Steps 3002 and 3003. In Fig. 5, 5000 symbols are selected from each distribution, completing step 3002, and each lookup error Complete step 3003 as measured. The data is expressed as percent lookup error (inverse of the lookup error rate) versus the number of symbols with a given percent lookup error. It shows that the sphere distribution peaks increase as the number of keys increases. If the process is repeated with a query query error instead of a lookup error, the data of FIG. 6 can be obtained.

To illustrate step 3004, a candidate code to be tested for substantial optimality is selected. This code is the 14 key code pn gt cr zk wj a e hi so ud xf ym vl qb proposed by [1]. The lookup error of the code in relation to the reference statistics is 105 words / lookup error. Performing as above, it is possible to find, on average, seven random trials of 14 key codes with possible equal divisions and with lookup errors equal to or better than those of the candidate codes. If the above procedure is repeated except that the query query error rate is used instead of the lookup error rate as a relevant measure, the code with a query query error rate that is better than the candidate code (4 words / query query) is averaged 3 times in random attempts. It will be discovered. Therefore, the ambiguity code of [1] is not practically optimal for lookup error rate or query inquiry error rate. In practice, most of the codes for the query query error rate are better than the given codes.

Experience has shown that if a sign is not explicitly optimized for a constraint, it is not practically optimal when measured against reasonable linguistic statistics.

Directed random walk

Managed random behavior is an iterative optimization method, where in each step, the best preceding code is used as a seed to generate a new code in which one or more new codes are better than the best preceding code. Thus, while repeating this process, better codes are found. This processing procedure will be explained intuitively at first and then more formally.

In the context of ambiguity optimization for one or more constraints, we assume that there is no detailed knowledge about the structure of the space to be searched. Without this knowledge, there is no insight into which direction to go for the best search. Thus, the safest processing procedure is to take as few steps as possible in as many directions as possible and not move until the steps are evaluated and compared in each of these as many directions. Since small steps can accumulate and lead the searcher to a dead end, such a search should always be reinforced with a 'restart' that will allow us to escape from unpromising or blocked roads.

More formally, the problem is to proceed with the desired code in the minimum step through the space of the ambiguities. According to the present invention, substantially minimal steps in the ambiguity space correspond to a single pairwise permutation that assigns decoded symbols to coded symbols. In each step of the optimization method, it is to test as many pair permutations as possible, preferably to test all possible pair permutations. The step is completed by selecting a pair permutation that best improves the characteristic being optimized. If it does not improve most, one of the pair permutations is chosen randomly.

Referring to Figure 7, the steps of the method are as follows:

4000: Select a starting code from the set of candidate codes.

4001: Generation of new signs from the start code by perturbation of the start code, preferably by a pair permutation of assigning a symbol to the key, preferably by all possible pair permutations.

4002: Measure the characteristics of the codes thus generated.

4003: Check whether a stopping criterion has been reached, such as a criterion for improvement limitations.

4004: If the stop criterion is reached, output the best sign.

4005: If the stop criterion is not reached, check if the set of codes generated by the perturbation of the current starting code includes a code that is better than the current best code.

4006: If there is a better code in the current set of codes than the current best one, select the best of these as the current best new code.

4007: Select new start code. If step 4005 is yes, select the best starting code from the current set as the new starting code, otherwise randomly select a new starting code from the current set. When complete, return to step 4001.

When only one constraint is to be optimized, selecting the best sign from the set of candidate codes is a simple matter of selecting the sign with the most optimal value of that constraint. However, when several constraints are optimized at the same time, there is only a partial ordering of the constraint values, so it is not clear how to choose between these values to best advance the optimization procedure.

One way to perform concurrent optimization is to optimize each variable independently. In this case, simultaneous optimization cannot be achieved in the case of collision, which is a general case, and should be determined by weighting the importance of each constraint. Relative weighting is part of the design constraints.

Construction of Strong Touch Typeable Codes : To facilitate the explanation of how to generate and use strong touch typeable ambiguities, we will define three classes that become increasingly strict.

Grade A: This type of touch typing is : 1) 20 words per minute and distracted every 15 seconds, that is, the query query rate averages one query query every 5 words, and 2) a 2 percent lookup error. That is, the lookup error rate is exemplified by an informal, generous typewriter that is characterized by one error every 50 words or 2 and 1/2 minute typing.

Grade B: This type of touch typing is : 1) 20 words per minute and distracted every 30 seconds, that is, the query query rate is one query query every 10 words, and 2) a 1 percent lookup error. That is, the lookup error rate is exemplified by a slightly formal, less generous typewriter that is characterized by one error every 100 words or 5 minutes typing.

Grade C: This type of touch typing is : 1) 40 words per minute and distracted every 30 seconds, that is, the query lookup rate averages one query lookup every 20 words, and 2) a 0.5 percent lookup error. That is, the lookup error rate is exemplified by the skilled hitter specified by one error every 200 words or every 5 minutes typing.

Referring to FIG. 8, a method of constructing a strong touch type batter code consisting of the following steps will be described.

5000: Determine quantitative values of acceptable lookup error rate and query query error rate.

5001: Select an ambiguous optimization method.

5002: Determine the minimum number of keys needed to be able to obtain the lookup error rate and query query error rate values determined in step 5000 using the keys and the optimization method selected in step 5001.

5003: Determine the maximum number of keys that can be given when the design of the target typing device is given.

5004: Determine if constraints are compatible. If the number determined in step 5003 is greater than or equal to the number determined in step 5002, the constraints are compatible, otherwise they are not compatible.

5005: As determined in step 5004, if the constraints are compatible, apply the optimization procedure selected in step 5001 to build an appropriate strong touch typing bat code. If incompatible, the procedure fails.

The details of the method are as follows:

Step 5000: Determine quantitative values of the acceptable lookup error rate and query query error rate. This can be achieved, for example, by selecting a strong touch type magnetic grade, as described above, by simply testing a person or a group or simply preselecting the desired values for lookup error rate and query lookup error rate.

Step 5001: Choose an ambiguous optimization method. As mentioned above, with regard to the construction of substantially optimal ambiguities, two methods have been described: randomization and managed randomization. Discretionary surveys are less potent than controlled randomization, but if the number of keys allowed is large enough and the desired strong touch type magnetic rating is low, the randomization will be sufficient. A weaker method, only code selection with a single random view, may be sufficient in some situations. In order to examine this in more quantitative detail, some experimental results will be described with reference to FIGS. 9, 10 and 11.

In this experiment, 5000 ambiguities with possible even divisions are randomly selected from each of the sets of ambiguities for 2-20 keys. In addition, for each number of keys up to 2-20, the optimization process was performed by applying a managed random action under each of the three conditions, which, by applying the target value method, Optimization for lookup error rate, 2) optimization for query query rate only, and 3) optimization for lookup error rate and query query rate. From the lookup error rate and query query error rate values calculated for randomly selected codes, the following statistics were calculated: best value, worst value, mean value, median value. All of these statistics are shown in FIG. 9 for the lookup error rate, and in FIG. 10 for the query query error rate, in addition to the results of the optimization process, in the only constraint that the lookup error rate and query query error rate are each optimized. The result of the progress of optimizing the lookup error rate and query query error rate is shown in FIG. From all of this data, you can decide which optimization method to use. It is more preferable to use the most powerful method, but a less powerful method may be sufficient. For example, if constraints are not strict enough, then even a randomly chosen sign can meet these specific constraints. This will be discussed further below.

delete

Step 5002: Determine the minimum number of keys needed to be able to obtain the lookup error rate and query query error rate values determined in step 5000 using the keys and the optimization method selected in step 5001.

Regarding the above test results and the grade of the strong touch type, the table can be drawn giving the minimum number of required keys and the three types of optimization relations mentioned for each of the three types of strong touch type. This table is shown in FIG.

Step 5003: Determine the maximum number of keys that can be allowed when the design of the target typing device is given. In general, ambiguous symbols will be used in small devices, and the number of keys will generally be a compromise between the size of the keys and the total typing device size. In some cases the number of keys may be determined according to a convention (for example, the convention of using twelve keys on a telephone keypad).

Step 5004: Determine whether the constraints are compatible. If the number determined in step 5003 is greater than or equal to the number determined in step 5002, the constraints are compatible, otherwise they are not compatible.

As will be seen more clearly in the proposed device embodiments detailed description below, the number of keys that can be accepted in the typing device may depend on a number of factors, and more precisely or less strictly by these factors. Can be determined.

Step 5005: As determined in step 5004, if the constraints are compatible, apply the optimization procedure selected in step 5001 to build an appropriate strong touch typing bat code. If incompatible, the procedure fails.

If the procedure fails, at least one of the following occurs:

A stronger optimization method is chosen.

The device design is modified to account for more keys.

Accepts lower hard touch typing.

The device is abandoned.

Smart Card for 9-16 Character Keys: A smart card is a substantially credit card-sized device that includes computer components such as a processor, memory, and interface circuits. The preceding smart card may also include a keyboard and a display. They are currently used for security and banking applications, but there are many other applications. This embodiment exemplifies a smart card sized device provision method and application range expansion method having a touch type keyboard. As a simple example, in banking and security applications, users of current smart cards must remember a series of numbers that are their device passwords. However, using a smart card for typing, it is a secret phrase in a natural language that is easy to remember but relatively long, and can be used in place of a short password that is difficult to remember. Examples of smart card sized devices to which the teachings of this embodiment may apply include personal digital assistants manufactured by Franklin Corporation and sold under the REX trademark.

With the current technology, due to the small size of the smart card, complicated and power consuming communication components for the transmission of data entered into the card keyboard are not possible. Therefore, this smart-card embodiment proposes a low-cost machine for sending messages ergonomically and efficiently using standard touchtones and standard touchtone generators.

Most phones have twelve keys, each associated with a touchtone, which activates each key to give a distinct touchtone to the handset. However, the global DTMF standard provides 16 touch tones, and the DTMF tone generator installed in most phones can generate these 16 tones. By using additional tones, each of the 16 keys can be assigned to a touch tone and used to encode alphanumeric symbol sequences. If the others are the same, the greater the number of keys, the lower the sign ambiguity associated with these keys. The teaching of this embodiment is therefore to substantially use all sixteen touch tones to encode alphanumeric sequences. In this way, information communication devices having a low ambiguity code can be manufactured by applying low cost components that can be easily applied.

Another object of this embodiment is:

Provides a touch keyboard for smart card-sized devices.

A method for simulating a set of encoded symbols that is larger than a set of encoded symbols that may be physically generated by the device.

Provide example device where lookup error is the main constraint.

Provides a keyboard / visual display device combination suitable for smart-card sized devices.

Provides an unambiguous technique that can be operated using very limited computer memory.

One or more unambiguous techniques can be operated, each of which provides an example of a system suitable for local computer processing power. In this case, the first decryption technique is used to provide feedback to the user of the transmitting end, and the second decryption technique is used at the receiving end.

The manner in which these objects are achieved by the present embodiment will be described in detail.

The strong touch typewriter smart card device is small so that substantially only a small number of full-size keys can be placed thereon. If any area of the card is reserved for the visual display device, then the keys area is further reduced. Between the requirement for substantially full-size keys for touch type typing and the requirement for a large number of keys to enable low ambiguity codes, a desirable compromise on the number of keys is 9-16 keys. Is in the range of. Two possible arrangements for the arrangement of keys with numbers in this range are shown in FIGS. 15 and 16. The assignment and functionality of the keys and the relationship between these and other components of the smart card will be discussed in detail below.

Synthesis of Encoding Symbols Using Coding Sequence With Meaningful Decodings: The set of decoding symbols is divided into two subsets with reference to FIG. 1) a core set composed of symbols associated with encoded symbols having a one-to-one relationship with a physical input means, and 2) an auxiliary set composed of symbols associated with encoded symbols having a one-to-one relationship with a physical input means. 100). The method of synthesizing the coded symbols consists of the following steps:

101: Set a first potential ambiguity that associates subsets of the core set with coded symbols. The coded symbol has a one-to-one relationship with a physical signal, and the physical signal is generated by the touch type device to physically express the coded symbol.

102: Identify a short sequence of encoded symbols in which possible decoding of the coded symbols sequence may not form part of a meaningful decoding.

103: Set a second potential ambiguity code as the relationship between the subsets of the sub-set of decoded symbols with the short sequences of encoded symbols identified in step 102.

For example, suppose that 16 coded symbols are associated with 16 DTMF tones so that the tones physically represent coded symbols. The tones will be displayed as (0, 1, 2, 3, 4, 5, 6, 7, 8, 9, *, #, A, B, C, D). We will take the letters [a-z] as the core set of symbols and associate them with a coded symbol that can be physically represented via a first ambiguity code as follows. (0, aw), (1, bi), (2, cx), (3, d), (4, ej), (5, fo), (6, g), (7, hv), (8 , ky), (9, l), (*, mu), (#, n), (A, pz), (B, qr), (C, s), (D, t). Here, the first element of each pair gives an encoding symbol, and the second element gives decoding symbols associated with the encoding symbol. The auxiliary set of decryption symbols will be a single set of symbol 'spaces'. Thus, we synthesize one encoded symbol to represent one decrypted symbol 'space'. Any candidate sequence is A8A, which corresponds to the following decoding sequences (pkp pkz zkp zkz pyp pyz zyp zyz). Since none of these decoded sequences form part of the words in the meaningful sequential reference list, the encoding sequence A8A is a suitable sequence for representing the elements of the subset, and we use pairs to represent the 'space' symbols. A8A, space). And the 'space' symbol is associated with the input means, and when the input means are activated, the input means emit a sequence of tones associated with the A8A. On the receiving end, the decryption means will convert the sequence A8A into a 'space' symbol. It is clear to the user whether a given input means is associated with a unique coded symbol or a synthesized coded symbol that can be physically represented. Any large subset of the decryption symbols can be expressed in this way. Given the above specification, a programmer with ordinary skill in the art can write software that automatically generates a first (ambiguous) code for the core set of coded symbols given a desired number of synthesized coded symbols and a set of reference statistics. It's easy to make.

Provision of an unambiguous technique that can be operated using very limited computer memory: Due to the limited processing power and memory capacity of the current generation of smart cards, low ambiguity is significant in ambiguity design. There is little computer processing power allowed for non-decryption devices on such cards.

For any ambiguity, most of the unambiguous efforts on the user's side as well as on the computer hardware and software side are in choosing which alternative decodings should be chosen for ambiguity encoding. Given the limited computer processing power of the smart card, the query lookup for alternate decryptions can be entirely eliminated. In the absence of query lookups, only the most probable decoding for each encoding needs to be stored, because when each encoding sequence is accepted, only the only possible decoding will be output by the unambiguous technique. With this simplification, in particular a compact database can be obtained, for example a simple suffix schematic form.

Since querying is eliminated, sufficient visual feedback can only be provided using a simple low power consumption display, e.g., a single-line traveling banner display relative to the type of display used in the handheld calculator or digital clock. Can be.

We will refer to this unambiguous method in which only the most probable decoding sequences are stored and output as simple lookup disambiguation. Simple lookup unambiguation can work effectively with ambiguities that can be typed in a significant strong touch. Thus, the surprising result of strong touch type typing codes is that they effectively operate a very simplified de-ambiguization technique.

A substantially optimized ambiguity code of 16-keys suitable for the application of this embodiment is a code having a lookup error rate of 4043 words / lookup error, and 68 word / query lookup error rates aw bi cx d ej fo g hv ky l mu n pz qr st, which is shown in array 51 on the 16-character-key smart card of FIG. 15. The figure also shows the display means 50 for outputting the decodes of the coded symbols input via the keyboard and the auxiliary input means 52 which can be activated with the thumb and can encode various further symbols and mode conversions, This will be explained in more detail with respect to other embodiments. Of particular note here, the screen output means 50 is arranged such that the character input means 51 and the auxiliary input means 52, which are operated by the thumb, are arranged in a position which is comfortable for operation by one hand (right hand, in the figure). At the same time, if the size of the smart card is given, it is desirable that the screen be arranged to allow as many keys as possible and to be comfortable and fully visible within the frame made of the thumb and index finger activating the input means 51, 52. This arrangement solves the problem of allowing the touch-type keyboard and the large screen output means to coexist functionally on the smart card.

Since query lookup is not allowed on the sending side in this embodiment, the code was chosen by the relevant random behavior optimization using a lookup error as the only constraint of the optimization. It is noted that lookup errors using this code will occur on average about once every 16 pages of typed text. Thus, this code is suitable for actually communicating long messages correctly even in the absence of query lookup techniques. If it is desirable to sacrifice some lookup optimization on the sending side to reduce the query lookup rate at the receiving end, for both query and lookup errors, examples of other optimized codes are lookup error rate 2670 words / lookup error and query lookup error rate 101 words. / W with query aw bu cx d eV pz go hv im ky l NQ prst. Choosing an optimized code for both lookup error rate and query query error rate is appropriate in at least two circumstances, which may be useful if: 1) if the smart card is actually strong enough to maintain a query query technique and / or 2) a query based query query. This is the case when the encryption technique is used at the receiving side of the communication initiated by the smart card. In the latter case, for example, if the user composes a message using a smart card, sends it to another computer via a telephone line, and then executes a second unpassive pass using a more powerful decryption technique. . Indeed, the second unambiguous transmission need not be executed by the person who wrote the message, but may be executed by the second person, for example the first person's secretary.

In any case, the look-up error rate of this second sign is still extremely low compared to the rate at which a very skilled hitter, for example, makes an error of one error approximately every 100 words. A hitter who hits 20 words per minute answers one query every three minutes for the first 16-character-key code and one query every five minutes for the second 16-character-key code. By any reasonable measure, it is believed that these 16-character-key codes can be typed in a strong touch in both cases. When performing the same optimization with respect to 9-character-keys, for example, lookup error rate 116 words / lookup error and query lookup error rate 4.4 words / query lookup and optimized for only lookup error rate akw bnq cly dhx epv fim gr jot suz This code is shown in the arrangement on the 9-character-key smart card of FIG. By optimizing both the lookup error rate and the query lookup with managed random behavior, you can also build codes such as am bnz cfi dhx evw gjr kos luy pqt with a lookup error rate 109 words / lookup error rate and 6.2 query / query lookup. Can be. It is noted that when a simple lookup error deambiguation is used and a query query cannot be performed, the strong touch type typing can be discussed with respect to the lookup error rate. Evaluate whether lookup errors occur at a sufficiently low rate to produce acceptable text. Even with these 9-key codes, the look-up error rate can only be compared to the rate at which a skilled hitter generates a typing error, so these codes can be considered to be typed in a strong touch manner. Moreover, since smart cards are used for writing short messages, for writing e-mails, for communicating with Pippi, etc., the criteria for the accuracy of the text may be lower than for the final text, for example. In this consideration, the definition of 9-16 keys has been reached as the preferred range of this embodiment. More than sixteen keys are difficult to fit into a smart card, but actually have the advantage of full-size keys. On the other hand, the ambiguity of less than nine keys will not be able to be typed in a strong touch on a simple unambiguous technique that is compatible with the limited computer processing power of a smart card.

Feedback to Others : Smart cards with simple lookup errors unambiguous can be manipulated by a competent touch type, and require no feedback from the smart card during communication and / or potentially allow the other to enter the smart card via the telephone line. In the form of synthesized words based on the symbols entered into the device, feedback can be received from the receiving communication device. However, when there is sufficient computer processing resources mounted on the card to maintain the feedback, it is desirable to provide feedback directly from the card.

It is pointed out that the computational resources required to provide useful feedback are much less than the computational resources required for simple lookup error unambiguization. Even if an unambiguous database and software are not present on the card and use only the basic electronics system well known to those skilled in the art, any particular type can be sent directly to the screen output in response to each key press, The most likely character (according to reference statistics) associated with the key. For example, using the symbols aw bi cx d ej fo g hv ky l mu n pz qr s t mentioned above, the resulting text is typically readable (by humans). For example, the first line of a Gettysburg speech made using the 1-block (one-character) statistic reads as follows:

oour score and sehen kears ago our oathers irought oorth on this continent, a nea nation, conceihed in liiertk, and dedicated to the proposition that all uen are created erual.

This level of accuracy already provides enough guidance to the batter about the text you type on the smart card. This example shows that an unambiguous technique can be achieved with extremely small amounts of memory. Here, the only memory required is the memory required to store the 16 characters to be displayed in response to the activation of the 16 keys. This approach can be weighed in terms of the computer processing resources required. With more memory contiguous, 2-, 3-, and higher block probabilities can be stored and used as the basis for unambiguization based on well-known contexts, thus showing the user text with increased accuracy. .

Context-based deambiguation is well known in the art but has proven to be not practical until now. The following example shows the reason for this: Unambiguous based context is not powerful enough to effectively unambiguous code that is too ambiguous. In this example, context-based unambiguization can be typed with a sufficiently strong touch and linked with sufficiently unambiguous code, enabling effective de-ambiguization with context-based methods. The prior art has avoided context-based deambiguity and preferred word-based deambiguity. However, by applying the teachings of the present invention, context-based deambiguation is made operational and viable for practical use.

The above example also shows that 1) word-based de-ambiguization is not necessary to practice the teachings of the present invention, 2) microprocessors are not required to implement the teachings of the present invention, and 3) potentially one or more unambiguous techniques are ambiguous. It can be used in the same communication system based on sign. Word-based decryption methods or other de-ambiguization methods may be used on the receiving side of the smart card while local smart cards use simple context-based decryption to provide feedback to the user of the card.

As the size of blocks used for context-based deambiguation increases, so does the accuracy of the text produced. However, the amount of storage required at any block size approaches the amount of storage needed for word-based unambiguization, and word-based unambiguity generally provides better results than context-based unambiguity, so if word-based unambiguity is maintained, Word-based deambiguation will generally be preferred if there is enough memory to do so.

Other Applications: If there is memory available in addition to that required for unambiguous database and software storage in smart cards, the potential applications of this device are significantly increased. For example, a user may have only a few bytes of additional memory to invoke the appropriate voice answering system for phone book information, request a phone number by typing a name or other identity information on the smart card keyboard, and other powerful machines. Received phone numbers can be stored in user memory for download.

Minimize inquiry query errors-Typing device for transportation. This embodiment relates to a case where a query query error is a major constraint in the design of a typing device. Trying to answer a query query is distracting in typing, so reducing the query query rate is generally worth it. However, for some applications, it is of paramount importance to reduce query query rates.

The query will be output on the visual display on most typing devices that implement ambiguities. When the user's field of view is urgently fixed elsewhere, for example for the sake of safety when the user drives the car, the field of view should be minimized to draw attention to the evaluation of the query. Even if the inquiry is generated by auditory means, it is very important to minimize distraction of the driver's attention. Moreover, both hands of the user are generally focused on the steering wheel while driving the car, so do not let go of the steering wheel as much as possible to operate the typing device. This object is achieved by inserting the input means of the typing device directly into the handle.

Referring to FIG. 17, it is noted that any number of input means can be inserted into the handle 200. In order to allow the finger to grip the handle better, many handles are serrated on the inside or back surface of the handle. For this handle it is natural for the first plurality of input means 201 to be associated with each of the plurality of teeth. When the driver grabs the handle, four of the first input means 201 are in contact with the fingers of each hand. The area of the handle that is in contact with the driver's hand may change from moment to moment, for example when the driver rotates the handle at a large angle. At any moment, which set of eight keys is in contact with the driver's hand can be recognized by position sensing means, for example in combination with a high pressure sensitive key and a simple electronic circuit system which is clear to the person skilled in the art. The second plurality of input means 202 can be disposed along the outer or upper surface of the handle, so that one of the input means is in contact with the thumb of each hand when the driver grabs the handle. Again, this may, at any moment, detect which of the second input means is in contact with the thumb of the driver with the appropriate position detection means.

For the handle with the above-mentioned keyboard inserted, the eight keys of the keys associated with the first input means in contact with the fingers of both hands and the two state change keys associated with the second input means in contact with the thumb of the driver It is natural to choose.

Ambiguous Code Selection: Applying the managed random behavior taught by the present invention for practically optimal coding and optimizing only for query query error rates, for example, construct the following code on 8 character keys: Lookup error rate 70.2 words / Lookup error and query lookup error rate 4.1 words / query aksz bcev dfi gmo hqt jnw luy prx . Throughout this specification, these ratios are calculated using simple words based unambiguity as the unambiguous method for reference statistics. It is natural that the query query error rate of this code is so high that it cannot be regarded as being able to be typed in a strong touch. The hitter / driver hitting 20 words / minute is distracted by querying queries approximately every 12 seconds and is too frequent to be compatible with safe driving. On the other hand, even an experienced hitter would not be able to hit 20 words per minute while driving, potentially making the relationship between query inquiry error rate and typewriter speed acceptable to strong touch typing.

In addition to actually selecting the optimal code, there are some additional strategies to reduce query lookup rates, which will be used in combination. These are,

To increase the total number of keys by increasing the number of keys that can be activated by each finger. This is, for example, coding that adds a different key sequence to the handle or allows each key to be in multiple positions equally (multipositionable) or simultaneously presses two or more keys to encode a different subset of encoding symbols. It can be understood that it can be implemented using the method.

delete

Eliminating a query query when less likely deciphers are much less likely than most likely deciphers. The parameter that controls how close the probability between the less probable deciphers and the most likely deciphers that will elicit a query query is the parameter whose value can be selected by the user. This technique may be valuable in any embodiment where the query query error rate is a related constraint.

Using complex coding / ambiguity as described in detail below,

This involves using a more powerful method of unambiguization than simple word-based unambiguization.

Phone Keypad Compatible with Existing Phone Keypads: In this embodiment, the limit of the number of keys on the keypad is of paramount importance, since it must generally be compatible with current telephone equipment having 12 keys. In this embodiment, it is necessary to leave two keys for non-letter symbols such as space, back space, period, and transfer-end symbol. Thus, 26 characters should be split up to as many as 10 keys. In this embodiment, a minimum lookup error rate and a minimum query query error rate are desired. Using the optimization method taught by the present invention and using the most even partitioning of the 10 keys, the lookup error rate 138 words / lookup error and query lookup error rate 9.3 words / query lookup are optimized simultaneously for the lookup and query lookup error rates. You can find signs like amq be cdu fiy gpx hl jsv krz nw ot . This is the standard ambiguity lookup error rate of 29 words / lookup error and query lookup error rate. It can be compared to 2.2 words / queries, and is more than four times better than the standard ambiguities. The 10-key code, optimized for the lookup error rate and query query error rate, is shown in the arrangement of the telephone keypad of FIG.

In addition, the 10-key code, US Glover Patent No. 5,818,437 and US King Patent No. It can be compared with the nine key signs in 5,953,541, respectively. The first afg bkn jlo mqr ew dhi sux ptv cyz of the symbols has a lookup error rate of 86.5 words / lookup error and query inquiry error rate of 3.9 words / query, and the second rpq adf nbz olx ewv img cyk thj su has a lookup error rate of 115 words / lookup Error and query lookup error rate 5.2 It has word / query lookup. Both of these codes are considerably inferior to the ten key codes designed in this embodiment. U.S. Patent No. 5,818,437 and No. Since both 5,953,541 are not mentioned about the construction of cited ambiguous codes and are not statistics available for optimizing these codes (if they are actually optimized), further conclusions are made regarding the practical optimality of the codes. Can't get off

Another useful comparison is for nine key codes optimized according to the teachings of the present invention for both lookup error rate and query query error rate. For example, a code having a lookup error rate 109 words / lookup error rate and a query query error rate 6.2 words / query lookup is configured as am bnz cfi dhx evw gjr kos luy pqt . Comparing these results, the improvement due to the teachings of this embodiment uses two or more keys to improve the two reasons, 1) lookup error rate and query query error rate, and 2) optimization for both lookup error rate and query query error rate. You can see that. If the teachings of this embodiment are extended to 11 and 12 key codes, for example, 11 key codes with lookup error rate 215 words / lookup error rate and query query error rate 10.1 words / query query avy bn cl dhx ew fip gjo kr mu qt sz and lookup error rate 313 word / lookup error and query inquiry error 13.2 12 key symbols with word / query lookup aw bn cky dhq ef go ip jr lz mx sv tu . Thus, at the expense of the use of the * and # keys for the encoding of non-letter symbols, the lookup error rate can be greatly improved, the query lookup rate can be significantly improved, and standard phone compatible keyboards (class B) can be strongly touched. To get to enough. Whether these improvements are more important than reducing the ability to encode non-letter symbols using the * and # keys can be determined in relation to the intended use of the built device. Non-literal symbols may be encoded using a sequence of encoded symbols, as described in connection with the specified smart card embodiment above. If the * and # keys are available for encoding non-letter symbols, then the ergonomic system to some extent considers the convention of using the # symbol as the transfer-end symbol. # Is encoded to make the end-of-word symbol, ## is encoded to make the sentence-end symbol, and ### is encoded to be the transfer-end symbol. In this way, the complexity of encoding a symbol can be changed in inverse proportion to the probability of the symbol. Depending on the application, sequences of * symbols can be used to encode non-letter symbols other than backspace, @ (for email), and / or can be used as mode conversion symbols.

Telephone keypad in alphabetical order . This embodiment proposes a solution to a very limited keyboard design problem in which the number of keys is determined, the arrangement of the keys, and the symbol arrangement of these keys is determined. This problem can be achieved by: 1) preserving the familiar alphabetical arrangement of standard ambiguous codes as well as possible, and 2) being compatible with existing standard telephone keypads, and 3) improving the lookup and query query error rates compared to standard ambiguous codes. Occurs in design. These constraints place restrictions on the choice of the number of keys based on the ambiguity. For example, an ambiguous code for characters may be selected to be assigned to the ten keys of the telephone keypad, leaving the * and # keys available for encoding non-alphabetic symbols. Also, while standard ambiguities use equal partitioning where possible, while still choosing other partitions that are not even, given constraints can still be considered.

Given the constraints of the alphabetical arrangement, each division into 26 elements arranged in ten groups corresponds to a unique ambiguity sign. Given enough computer processing time, it would be possible to evaluate each lookup and query query error of these codes and select the best one. Another more effective method is to apply the optimization method taught by the present invention to this constrained optimization problem. The present invention teaches that one should define the minimum initial step in the set of possible codes, without information suggesting application of any more complex initial step. In this sense, an ambiguous sign is an array list of ten groups of characters, all characters appearing in exactly one group, and within and across groups, the characters appearing in alphabetical order. One example is ab cd ef gh ij kl mn opqr stuv wxyz . Thus there are nine gaps separating the groups. The initial step is to move one character over one interval. For example, if we choose the second interval, we move the letter c to the left in one initial move to sign abc d ef gh ij kl mn opqr stuv wxyz or, to move the letter b to the right to sign a bcd ef gh ij kl You can create mn opqr stuv wxyz . Given a particular sign, all possible signs that can be obtained in one initial move from that particular sign can simply be generated. Given these observations and any of the above managed behaviors, it will be apparent to those skilled in the art how to apply the optimization method by the teachings of the present invention in this sense. Applying this method, for example, a code having a lookup error rate of 65 words / lookup error rate and query inquiry error rate of 5.8 words / query lookup ab cd ef gh ijklm no pqr s tu vwxyz You can find it. This code is shown in the preferred arrangement, on the telephone keypad of FIG. The error rate of this code shall be compared with that of a standard ambiguity code with a lookup error rate of 29 words / lookup error and query query error rate of 2.2 words / query. Thus, the lookup error rate is more than doubled and nearly tripled in query lookup error rate without losing easy-to-scan alphabetical order or compatibility with existing telephone equipment. Note that the description of selecting eleven or twelve keys encoding a character symbol in the telephone embodiment can also be applied in the present embodiment: using optimization for partitioning to substantially accommodate alphabetical order constraints. It is possible to generate substantially optimal codes for 11 and 12 keys.

The optimization method for partitioning is not limited to this embodiment: for example, this can be applied to the smart card embodiment discussed above or to a quartet-like keyboard discussed below, on the arrangement of 9-16 keys with character symbols. Can generate optimal codes with alphabetical order.

Quartie-like keyboards: 1) compatible with standard keyboards; 2) the approach used in the previous embodiment to develop keyboards optimized for various constraints is similar to: 1) similar to standard quartie keyboards; Can be used to develop keyboards optimized for various constraints. As in the previous embodiment, the order in which the symbols are assigned to the keys will be maintained to maintain the layout structure of the standard keyboard as much as possible, while minimizing the lookup and query query error rates as much as possible while maintaining possible even or possibly uneven partitioning. This embodiment generally maintains the characters in a row of the same key given in the question tie arrangement, as in the case of a conventional question tie keyboard in which letter assignment keys are reduced from top to bottom.

There is a quaternary-like keyboard layout with three columns of letter keys and variable rows from one to ten. It is obvious that the lookup error rate and query query error rate are very high for only one row, that is, only three keys. There is only one possible ambiguity that corresponds to the symbolic order of the quartet keyboard. This code is a qwertyuiop asdfghjkl zxcvbnm possible qualitative pseudo-like code with a lookup error rate of 2.8 words / lookup error and query rate of 1.1 words / query. This low quality code is unacceptable for any important application. As the number of rows increases, more and more ambiguous signs can be found, and strong typing can be done beyond any number of boundary rows. At the same time, as the number of rows increases, so does the size of the device needed to include a keyboard that remains substantially full-size. Therefore, the design of qualitative-like keyboards should be a compromise between sign ambiguity and keyboard size. For example, if using a full-sized key and wanting to make a quaternary-like keyboard that is substantially the same size as a small calculator, seven rows can be used, as shown in FIG. Possible non-uniform strong touch type typing codes that are practically optimal for the lookup error rate and the query lookup error rate are qwe rt yu iop as dfg hjk l zxc with a lookup error every 668 words and a query lookup every 35.5 words. vb nm and can be typed clearly in a strong touch with respect to the use of multiple classes of typewriters and keyboards. In FIG. 20 these symbols are shown in a preferred arrangement. A typewriter having a keyboard as shown in the drawing may be suitable for writing an e-mail or a memo. Those who are familiar with standard querty keyboards can easily type this with minimal or no learning, and even if it is made full-size, it fits easily in your pocket.

It is noted that the cost of insisting on the quel tie habit is high for the lookup error rate and the query inquiry error rate. If 17 keys can be used for character symbols and characters can be randomly arranged in 17 keys, lookup error rate 7483 word / lookup error and query lookup error rate 290 wrt bu gi ov p af sd ej ky You can find signs like hz cx n mq . This is equivalent to one lookup error every 30 pages of typed text, and less than one query query per page of typed text. It is difficult to think that such a keyboard cannot be typed by a strong touch method.

Referring to FIG. 21, this symbol may be arranged such that the 18 characters are placed at the quail tie position or at a position very close to the quail tie position. In this arrangement, in order to maximize the batting gesture habits in the quartet keyboard, the finger is placed in the optimized quartie-like keyboard groove row so that the index finger of the left hand is the (Space) key and the right index finger is the (ej) key. Should be placed on the table. In particular, by placing the space key and e in the groove row, this arrangement is not only in terms of the preservation of the gesture but also in the groove row weight and the weight of the strongest fingers. It is a step forward. By appropriately assigning symbols to keys, any ambiguity can be optimally matched to a qualitative (or other conventional) keyboard.

It is noted that by making some changes from the uniform quaternary array, very substantial results have been obtained in terms of the qualitative gesture and the functional preservation of the array. By shifting the columns slightly from each other, as in the standard quartier layout, many or most of the finger gestures required to manipulate the various keys of such a quartet-like keyboard are the same as or similar to the finger gestures required to manipulate a standard quartier keyboard. Do. This accounts for the tradeoffs between constraints to preserve conventional arrangements and constraints necessary to maintain conventional functionality.

Labeling for Optimizing Cross-Platform Compatibility
Between the various competitive constraints that must be optimized and the various user populations and their needs, between user-optimized qualitative-like keyboards and other corresponding keyboards optimized for other constraints, such as other optimized or lookup error rates and query query error rates. Such a device must be implemented in practice so that it can be selected from. This choice would be possible if the software could switch key markers. This object may be achieved if each key is provided with screen means capable of displaying at least one symbol at a time. Such display means may be, for example, a diode array or a liquid crystal display for emitting light, so that the key markings can be changed by software, thus eliminating the need for a fixed marking on the keys.

Those skilled in the art will appreciate that the keyboard design method used for this embodiment may be applied to maintain or partially maintain other conventional keyboard designs, such as Azerty keyboards used in France.

Numeric Keypad-like Keyboard: The purpose of this embodiment is to create an ambiguous keyboard useful for most computer users without changing existing hardware at minimal cost. These advantages include in particular the advantages of one-handed typing and the potential for compatibility with ambiguous keyboards designed for handheld devices. Standard 101-key vending machines for workstations and personal computers typically have a numeric keypad on the right side of the keyboard, which is deployed in a quartier arrangement, and, in the case of a laptop computer, is included as a subset in the quartier arrangement. It usually includes a series of arrow keys near the numeric keypad or other input means effective for moving the cursor.

Referring now to FIG. 22, there is shown an ambiguity optimized for a normal numeric keypad layout 600 with means 601 for moving the available cursor. The numeric keypad 600 has 17 keys of various sizes in this example. Some or all of these keys may be used for punctuation or other symbols, depending on other design constraints, and these other design constraints may be assigned letters. It can influence the selection of the number of keys, the distribution of other symbols about letters and modes, and the like. Essential features of this embodiment are as follows:

Assign ambiguities to multiple keys on the numeric keypad

Optional use of a thumb-operated auxiliary input for mode switching.

The assignment of ambiguities assigned to multiple keys of ambiguity keypads can be achieved in software; Those skilled in the art will understand that no special purpose hardware is required. However, if you want to label the keys so assigned with ambiguity elements, you will need to modify the keyboard markers slightly. In this case, to give a specific example of the use of ambiguities, an ambiguous encoder is selected that assigns characters to 17 keys in a manner that minimizes the lookup error rate and query lookup error rate for the standard corpus. The code af bu cx d ej gi hz ky l mq n ov p r s t w shown in FIG. 22 has a lookup error rate of 7483 words / lookup error and a query inquiry error rate of 290 words / query. Such symbols have already been discussed above. The signs given here are arranged to preserve alphabetical order. The given sign is not optimized for alphabetical order; Only the lookup error rate and query query error rate were optimized. This procedure is almost the same as the keyboard configuration of the question tie-order: first, optimization for the lookup error rate and / or query look-up error rate, and reassignment to obtain the desired character order. According to the teachings of the present invention, it will be possible to simultaneously optimize for lookup error rates, query query error rates, alphabetical order and / or other constraints.

Reference is again made to FIG. 22 to illustrate the use of a thumb operated auxiliary input means for mode switching. For this description, it is assumed that the auxiliary input means consists of four keys: up 602, down 603, left 604 and right 605 arrow keys. These functions are typically implemented as four keys that can be pressed, but sometimes they can be implemented as touch pads, joysticks, or any other device capable of generating a number of different signals as functions manipulated by a user.

In Fig. 22, a number of keys are represented by symbols, in this case numerals, which are not symbols of ambiguities. These other symbols can be obtained by pressing a particular one of the four keys of the auxiliary input means. Possible mode assignments from the secondary keypad to the input means are:

602 Shift key for uppercase

603 (bottom) Numerical / Shoelace status

Left symbol on the 604 (Left) key

605 (Right) Right Sign Above Key

It is understood that 1) it is possible to assign modes and / or symbols differently to an auxiliary keypad in accordance with the teachings of the present invention, and 2) to a further complex series of auxiliary input means, that additional modes and symbols can be assigned to the auxiliary input means. . Assigning a symbol to a mode will be discussed in more detail in conjunction with another embodiment. This discussion can be applied to other embodiments, including this embodiment.

Objects and Benefits of Thirteen-Character Key Codes: Several related embodiments in accordance with the teachings of the present invention are directed to keyboards in which the number of keys used for strongly correlated symbols is substantially half of the number of strongly correlated symbols. Take advantage of the surprising benefits of ambiguity optimization. In particular, if you take a set of strongly correlated symbols as the character set [az] used in English, the preferred number of keys is 13. The surprising benefits of the 13-character key sign for English are:

Steel touch type,

Ergonomic, touch-typed, unambiguous text input.

Inquiry that can be ergonomically typed by touch.

Compatibility with standard keyboard layouts (Quartie keypad, numeric keypad, and telephone keypad),

Provides preservation of typing skills from one hand to two hand typing,

Integrated mouse / keyboard provided.

Provides techniques to reduce typing damage.

Other objects and advantages will be apparent from the following detailed description, which is described in turn.

Strong Touch Typing: Again referring to FIG. 11 and FIG. 12, it can be seen that the ambiguity on the 13 keys with respect to word-based unambiguity can be typed in a strong touch, even for experienced hitters. have. Using the above-mentioned managed random behavior, for example, lookup error rate 515 words / lookup error rate and query inquiry error rate 21 words / query lookup, (class C) code with strong autonomous typing aw bn ck du ef go You can find hv ip js ly mx qt rz . This symbol can be seen in the preferred arrangement in FIG. As mentioned above, the query query rate can be further reduced by adapting a parameter that controls how important the query query rate is to require the attention of the user. In the extreme case where the query query is completely stopped, the lookup error adjusts the touch type typing. For this code, a lookup error will occur once every two pages of typed text, in other words, this rate is much less than the typing error rate of a very skilled hitter. The 13-key code is therefore suitable for multiple touch typing tasks for different users.

Unambiguous text input that can be typed ergonomically and in touch: with arbitrary ambiguity and unambiguous techniques, it is possible to enter information in an unambiguous way-for example to add information to an unambiguous database. It is desirable to provide a technique. Ergonomics that allow for unambiguous text input techniques for all uses of ambiguity are desirable, with ergonomics being easy to manipulate. With respect to the ambiguity typed by the touch method, it is preferable to operate the keyboard in a text input mode which is not ambiguous by the touch type type.

One strategy used in the prior art to achieve unambiguous text input using a small number of keys is to apply a chording scheme. To satisfy the intent to reduce the number of keys to the maximum, coding keyboard designers have been consistently off from providing a sufficient number of keys so that the coding pattern can be simplified. By the elementary coupling argument, if the complexity of the chord is never more than two, that is, if it is not necessary to operate two or more input means at the same time to encode the symbol unambiguously, the number of keys Cannot be less than ½ of the number of encoded keys. In contrast to the prior art, the present invention teaches to provide a number of keys less than ½ of the number of symbols to be coded, provided a simple technique for unambiguous encoding of symbols is provided. In particular, the present invention teaches the provision of at least 13 keys and at least one mode conversion key that represents the letter [a-z]. The mode conversion key, when combined with one of the character-assignment keys, functions to uniquely and unambiguously encode one of the characters associated with the character-assignment key.

In the obscure keyboard, some keys represent multiple symbols in a single keystroke. In order to use the same keyboard in an unambiguous mode, at least one other key press that extracts each individual symbol associated with that key, which is probably the key press of the same key, must be combined with the single key press. In ergonomic terms, the combination of key presses is preferably as simple as possible, and for touch-type typing, it is more preferred that substantially the same combination of key presses is used for the unambiguous input of all symbols. These two constraints are most satisfied if 1) the same number of symbols are associated with each mock key, and (possibly evenly) 2) the number of symbols on each mock key is small. Taken together, these constraints mean that the desired number of keys for the ambiguity to be typed in touch is half the number of symbols that are vaguely expressed. Thus, in the case where the symbol represented is 26 characters of the alphabet, the constraint means that 13 keys are preferred when there are no other constraints.

Through this observation, one method that can be used to manufacture an ergonomic and touch typeable keyboard having an ambiguous text input mode that can be ergonomically and touch typed is described with reference to FIG. 23. It is described in connection with the case of vaguely representing English alphabet letters having 13 keys. In this figure, each subset of keys used to obscure characters 700 encodes exactly two characters. The typing device further includes a mode key 701 and means for placing the device in an ambiguous or unambiguous text input mode. Means for placing the typing device in an ambiguous or unambiguous text input state are software means for detecting which of these modes are necessary depending on the context at a given moment, or double tapping on this mode switch key 701, long or short of any key. It may be a special pattern input to other input means such as pressing, or other means known in the art such that a predetermined key corresponds to various coded symbols. In an unambiguous text input mode, activating the 701 key results in one of the two symbols associated with one of the plurality of keys 700 being encoded when one key in the plurality of keys 700 is activated substantially simultaneously with the 701 key. The symbol is selected. It is desirable to consider a pair of symbols on each key, which consists of a left symbol and a right symbol in a plurality of keys 700 and is marked with keys having a left symbol on the left and a right symbol on the right. One of these, the left one, is then associated with the activation of the 701 key without losing generality, whereby an unambiguous text input can be achieved. In other words, when one key in multiple keys 700 is activated in combination with the 701 key, the left symbol is unambiguously selected, and if the same key in multiple keys 700 is activated without 701 being substantially simultaneously activated, the right The symbol is chosen unambiguously. If the keyboard is designed in combination with additional modes, this same unambiguous text input method can be used for symbols in other modes.

Inquiry directed toward touch typing: Even for optimal ambiguity combined with developable artificial intelligence for unrestricted computer processing and unambiguity, some ambiguities may require human intervention for complete unambiguation. Can be generated from

Indeed, touch typewriters type without looking at the keyboard, keeping their eyes on the text being created or the copy being typed. Thus, for touch typed hitters, different interpretations of the ambiguity sequence are carried out in such a way that: 1) they do not take their eyes off the text output screen, and 2) they can answer the query query in a simple and stereotypical manner. It is desirable to have a query query for. This object can be achieved by listing the candidate words and maintaining a key to make the on-screen ambiguity stand out. The user scrolls through the list of candidate words using the scroll key, and the word in the scroll box is considered to be selected as soon as a key other than the scroll key is pressed.

With reference to Figs. 23 and 24, the software focusing on querying oriented towards touch typing and the visual display it controls are described in detail. In the first step 800, query lookups are detected, that is, the unambiguous technique finds that one or more meaningful decoding sequences in the database correspond to the input encoding sequences. Starting in the query mode is to induce a means of drawing the user's attention to the decryption under the displayed query. This means can be visual means such as frame box 702 shown in FIG. The possible decryptions are then arranged according to the likelihood of the decryptions (step 802). Then (step 804) the most probable decryption is output on the screen, in the context with the previously entered text, in the place indicated by the cautionary means. The software is then ready to detect whether it is an input from the scroll key or from another key (step 806). If any input is received from any other key, the means of attracting the user's attention is removed (step 808), the decryption is added to the previously entered text (step 810), and enters the ambiguous text input mode again. On the other hand, if an input is detected from the scroll key in step 806, a test is made to see if there are any meaningful decryptions in the database (step 812). If present, the current decryption is replaced by the most probable decryption (step 814) and returns to step 806. If there is no most likely next decryption, the typing device enters an unambiguous text input state as described above (step 816). If the decryption sequence is entered unambiguously, the means of attracting the user's attention is removed (step 808), and the unambiguous input decryption is added to the existing text (step 810) and enters the ambiguity text input mode again ( Step 818).

Although this method for showing other alternatives is proposed with respect to query-oriented query-oriented typing, it will be understood that the same method can be applied to other formats. For example, when "detoxes" are words with a relevant meaning, the database is thesaurus, or when "detoxes" are several possible translations that translate a word into a foreign language, the probability of the deciphers is automatically translated. Can be supplied by the program.

Constrained design preservation across platforms: Mouse / Keyboard Consider a user with a small typewriter with a one-touch typewriter keyboard, such as a personal digital assistant. Usually, the user may also be typing on his desktop computer with a two-handed board. If the two devices can be typed in an effective touch manner to the user, the motive patterns used for the one- and two-handed keyboards should be as similar as possible. The shorter the interval between two keyboard uses, the greater the requirement to preserve typing skills. In order to elicit the purpose of preserving typing technology across platforms and the means to achieve them, embodiments will be described in detail that one- and two-handed keyboards can be used to change rapidly.

This embodiment relates to a one-handed keyboard suitable for inputting data into a form such as a spreadsheet or a web based format. This may be useful for interacting with a program, such as a game or drawing program, that 1) requires a quick change between typing and cursor movement and / or 2) an appropriate set of symbols to enter may vary depending on where the screen cursor is located. will be. When applying the standard computer layout of a quartier keyboard and mouse, the user must release his hand from the vending machine to move the mouse. Changing the use of these mice and keyboards in tasks that require continuous typing and mouse manipulation—for example, in the form of design marks or HTML forms displayed on a computer screen—can be very laborious. In this embodiment, the one-hand keyboard is mounted on a frame that can move over the surface to perform the functions of the mouse as well as the keyboard. Since users may want to use two-handed keyboards primarily for text input, and may want to change the use of one-handed keyboards and the use of two-handed keyboards, two-key keyboards can be used to move the touch-type typing technology from two-handed keyboards to one-handed keyboards without interruption. It is desirable to make the arrangement of the keys on the field as similar as possible.

Referring now to FIGS. 25, 28, and 26, it can be seen that the object can be achieved by ambiguity selection, ie the movements of the fingers and thumb of one hand (in this case the right hand), Whether typing on a hand or two-hand keyboard, the ambiguity may be placed on the one-hand keyboard so that the typing actions of the other hand are similar to the movements selected for one-hand typing.

The design strategy is as follows:

Choose a 13-key code that has a minimum lookup error rate and query query error rate.
Select the physical layout structure for the 13-key. A preferred layout would be a layout with five keys in the top row, five keys in the middle (groove) row and three in the bottom row.
Choose which hand will operate the one-hand keyboard.
Assign the keys associated with the hand selected in the one hand arrangement and in the previous step: where the weight for the home row is maximized. The specific gravity for the strongest fingers is maximized. The specific gravity for the top row is higher than the specific gravity for the bottom row.

delete

delete

delete

Thereafter, each key operated with the left hand is paired with a key operated with the right hand in order to obtain an arrangement structure for the two-hand keyboard. Preferably, with respect to the plane of symmetry that intersects the center of the keyboard, the members of the pair are arranged symmetrically and proceed from bottom to top of the keyboard.

Of the two symbols associated with each of the original 13-keys, one symbol is associated with one of each pair of keys selected in the previous step. This can be done in a manner favorable to a one-hand keyboard or in a manner favorable to a two-hand keyboard associated with it.

If favorable to one-handed keyboards: Hold the higher probability letters on the selected hand side of the two-hand keyboard, and place the lower probability letters on the opposite side of the keyboard.

If favorable for two-handed keyboards: For each of the 13-keys, choose whether to place a lower or higher probability letter on the given side of the two-handed keyboard so that the sum probability in each half of the keyboard is as equal as possible.

Starting with a 13-key code and a right-handed arc selected to describe the teachings of this embodiment, the keyboard layout structure described in FIG. 25 can be constructed. When this one-hand keyboard is expressed as a two-hand keyboard, the layout structure results are shown in FIG. In this keyboard, the right hand strikes approximately 84 percent of these letters, while the left hand strikes approximately 16 percent of these letters. This asymmetry is desirable in that most key presses will be performed in exactly the same way, using either one or two handed keyboards.

Conversely, if most typing consists of two-handed keyboards and only occasionally one-handed keyboards, it may be better to have as much specific weight as possible on the two hands used to operate the two-handed keyboards. This object can be achieved with another arrangement structure of the two-hand keyboard shown in FIG. It will be understood that the typing behavior used to type in these two-handed versions of the keyboard is actually the same as the typing movement used in one-handed keyboards, whether you type one-handed keyboards with your right hand or with your left hand.

For the 13-key ambiguity, it is understood that there are 2 ¹³ ways of pairing the left and right hand keys of the two-hand keyboard in homology. This number is small enough so that in each paired set, the symmetry of the weights assigned to both hands can be assessed and the appropriate allocation chosen according to which intermediate value is desired, depending on whether you want the most symmetric weight or the most asymmetric weight. Can be.

While referring to the detailed view of the one-hand keyboard of FIG. 28, the keyboard is provided with a plurality of keys 300, an input means 301 capable of operating with a thumb, a mouse key 302, a palm grip 303, and a display 304. You can see how to achieve this. The keyboard may further comprise communication means for communicating symbol selections by the keyboard means to a computer. The communication means may simply be a wireless communication means such as a wired communication means or an infrared communication means. The keyboard is movably supported by a support means such as a desktop so that it can be moved on the support means using the pressure of the palm of the hand operating the keyboard. The keyboard preferably has palm grip means for the palm so that the pressure is effective for keyboard movement. The means for the palm of the hand for actuating the keyboard is preferably formed by means of a shape in which less pressure is effective to move the keyboard in the desired direction. For example, the shape can be an indentation shape in which the palm can be reliably placed. By moving the keyboard in this way, the fingers of the hand that moves the keyboard while moving are free to move in an effective way to operate the keys. This keyboard can thus be used in situations such as playing computer games, i.e., situations in which movement must be input simultaneously with symbol sequences such as text.

It is noted that this device functions as a mouse, but has little physical similarity to the mouse. Its shape factor is determined by the anatomy of the hand in a comfortable position to type in touch. Thus, the device must be considerably larger than a standard mouse, and the means for moving the device are substantially different. In order to communicate the movement of the keyboard 305 to the computer, the keyboard is, for example, a trackball familiar to those skilled in the art. Movement sensing means such as is provided. Preferably, the keyboard 305 is equipped with a biasing means such as a spring, so as to lift the keyboard from the support means when the hand weight on the keyboard decreases, thereby smoothly moving. In contrast, when a substantially complete hand weight is applied to the keyboard, the keyboard is relatively stably fixed to the support means, thereby making typing easier.

delete

As such, the two-handed keyboard will not be interrupted by the need for moving the mouse and one-handed keyboards for fast typing / mouse movement changes and can be applied to typing long episodes.

Visual representation of the keyboard: When using a single device that supports more than one ambiguity or more than one mode, it may be helpful for the user to have a representation that associates the current keys with symbols printed on the display device at any given moment. . This expression may be valuable for any typewriter, but is particularly useful for touch typewriters since all or some keys are hidden from view (typically by the operator's fingers) during use. Thus the display means integrated in the keys is of limited utility to the touch typed hitter. The most useful visual representation is that the physical layout of the keyboard is represented on a visual display. Such a device 304 is shown in FIG. 28, but can be incorporated into many embodiments described herein.

Reduction of typing damage: Damage from typing (repetitive stress syndrome) plagues many keyboard users. Many keyboards are designed in an attempt to reduce the repetitive movement stress of typing. It has long been recognized that the most effective way to reduce typing damage is for the typewriter to take regular breaks from typing. However, it is hardly practical because hitters are often under pressure to complete their typing tasks. The one-hand keyboard just described suggests a solution to this problem. Although the one-handed version of this embodiment describes the use with the right hand, it is obvious that the same design methods can be introduced in the one-handed keyboard for left-handed use. Each of the keyboards can encode all of the same symbols. Therapeutic typing device equipped with both a right hand keyboard and a left hand keyboard, such as the left hand and the right hand mouse / keyboard, can be operated by changing to the left or right hand. A user who wants to reduce repetitive stress damage with this pair of keyboards can type using one of those keyboards for a period of time—for example, for 15 minutes—and switch to the other keyboard for the next period. This gives the user a break on each hand without reducing typing productivity. If desired, the typing device may be provided with a locking device that can alternately lock one or the other of the keyboards to force the use of the change. When this treatment is complete, it is understood that the user can return to the two-handed version of the keyboard without relearning the required typing skills.

Collapsible PDA: In the smart card embodiment, the keyboard is manipulated with one or two handed fingers, the thumb being ambiguous for use in operating additional input means, especially mode switching input means, on these areas and on the side of the finger operation of the keyboard. It can be seen that it is convenient and ergonomic to arrange the text screen of the typing device based on the code. In this embodiment, when not grounded, folded once. When folded twice, it relates to a rather large range of other type devices using the same concept as the folding concept for designing a double folded information device that can be ergonomically performed differently.

The double folding design is a surprising result of ambiguous sign. A typing device made by the method taught by the present invention is capable of 1) effective for coding natural language, 2) using substantially full-size keys, and 3) enabling keyboards that are small enough to fit in a pocket or small hand bag. It is noted that This embodiment is based on manufacturing a handheld computing device from elementary units that are substantially the same size as a keyboard designed for ambiguity and can be arranged in various ways according to the immediate needs of the user. . The basic units may be connected to one another in various forms, collapsible and / or detachable from each other. In this way, computing devices can serve as laptop computers, personal digital assistants, phones, game consoles, and the like.

Referring first to FIG. 29, it specifies in detail four components of substantially the same size that are adapted to perform a particular function and interconnected in a collapsible and / or separable manner. 29 shows the device in an ungrounded state. Thus, one of the parts 900 has a first side that serves as the first visual display output, 901 has a first side that serves as the first keyboard, and 902 indicates that it has a first side that serves as the second keyboard. The keyboard arrangement on the first and second keyboards may be of other choices, but a 13-character-key keyboard is preferred. The last part 903 has a first face that functions as a mode switching thumb switch pair, used in combination with the first and second keyboards. The first keyboard means to work with your right hand. It will be apparent to those skilled in the art that there are similar configurations that can be typed with the left hand, which can be obtained by simple reassignment and repositioning of the four parts. In fact, a two-handed keyboard can be obtained by separating and reassigning four units, as shown in FIG. 33.

30 is a bottom view of a computer that can be folded twice without being grounded. Part 904 is telephone keypad 905 and corresponding second visual display 906. Part 907 is the third visual display and part 908 is the third keyboard. Parts 904, 905, 906, and 907 form second surfaces of parts 900, 901, 902, and 903, respectively.

Folding the computer along line 908 shown in FIGS. 29 and 30 yields the arrangement of FIG. 31. In this arrangement the third keyboard is exposed for typing, and the third visual display is used as the corresponding display. The keyboard layout here is a 12-key keyboard, but other choices are possible. This configuration may be used when the user cannot or does not want to open the computer completely due to time and space constraints. It can also be used to provide different functionality than a fully deployed computer, for example game functionality.

Finally, the computer can be folded along the fold line 909 shown in FIG. 31 to provide the double folding configuration shown in FIG. This configuration is typically the configuration of a computer that can be maintained for movement; In this configuration the device is small enough to fit in a pocket. Moreover, in this double folding configuration, telephony functionality is exposed for use. For many users, this may be the most commonly used device configuration. Although other choices are possible, it is noted that the telephone keypad with the ambiguity of the preceding embodiment is proposed.

In summary: Ambiguity allows the design of portable communications and computing devices that can function interchangeably as telephones, personal digital assistants, and laptop computers.

It will be understood by those skilled in the art that if each of the basic units including keyboard units is made of a touch screen, the various configurations and uses of the device will be further increased. But what can be lost is tactile feedback on keyboards made from standard keys that can be pressed. Various modifications are possible consistent with the teachings of the present invention.

Software Embodiments for a Typewriter Device Made of a Touch Screen The present invention includes embodiments of software as well as hardware. In particular, the methods of the present invention can be used to design typing techniques for touch screen devices, such as a personal digital assistant series manufactured by 3Com Company and sold under the PALM PILOT trademark and other trade names. Although the other methods described herein can be applied to any type of device with a touch screen, the focus will be on PALM PILOT grade devices, including handheld computers that can run multiple applications for illustration.

Referring to FIG. 34, a device of the PALM PILOT class is composed of: a touch screen 1000, a touch-sensing area 1001 used for inputting type through handwriting recognition software. The touch-sensing area may be a sub area of the touch screen, or may be implemented separately.

One of the essential and surprising features of this embodiment is the design of a completely new user interface for an information device, using a touch-type keyboard in a touchscreen device, which allows the keyboard to compete for limited screen space and applications. There is no need. The same touch screen area can be used for applications and keyboards.

If the keyboard has a character array fixed to the key, the keyboard does not need to be shown to the user to be manipulated. The user's finger "knows" where the key is without visual referencing. The keyboard can thus be used to enter data into an application currently displayed on the touch screen. If the keyboard can be typed in a strong touch, the keyboard can be used to produce high quality text, even if any screen space is not used for query inquiry feedback to the user.

With reference to FIGS. 34 and 35, some of the special features of the PALM PILOT class apparatus described in this embodiment are:

Touch screen 1000 ability, showing different keyboard layout structure easily.

Touchscreen ability to show different levels of intensity and / or multiple colored images.

Place area 1001 for inputting typeface at a distance away from the touchscreen screen or in a non-center area of the touchscreen

Use of personal digital assistants that run a variety of programs that compete with the keyboard for space on the touchscreen, such as schedule programs or address book programs.

The ability of the touchscreen to easily show different keyboard layout structures is used in this embodiment to allow a given input means to represent different symbols or groups of symbols, depending on the "mode" of the keyboard at a given time. When the touch screen is used to implement the keyboard, each input means is associated with a specific area of the touch screen. The dual function of the touchscreen as a visual display and a number of mechanically active input means is used to give different inputs and different labels to each input means depending on the mode. However, it will be understood that each of the mechanical keys is provided with a display device, so that the same effect can be obtained with the mechanical keys of the traditional pressable structure. In this way, the mode switching method specified herein in connection with the device constituting the touch screen can be applied to a device composed of mechanical keys, such as other devices embodied herein.

Mode Selection: Once the number of keys has been determined, one strategy of keyboard design to increase the number of symbols that can be coded is to increase the number of keys with the mode switch key. Pressing a mode conversion key converts a symbol encoded by a number of other keys. A typical example of this is a shift key on a standard typewriter fingerboard that changes the symbols encoded by the letter keys from lowercase to uppercase. In principle, uppercase letters can be encoded in a set of keys separated from the key that encodes the lowercase letters, and if uppercase letters occur at the same frequency as lowercase letters in communication, this may be an applicable option. Also, in principle, having uppercase and lowercase letters accessible in other modes does not mean that uppercase letters should be assigned to the same key as the corresponding lowercase letter. The same key assignment is actually chosen because there is a customary, conceptual and statistically close relationship between lowercase and uppercase letters.

Thus, there are three principles of assigning a symbol to a mode and a symbol to a key within a mode: faithfulness to statistical relationships between symbols, faithfulness to conventional relations, and faithfulness to conceptual relationships. In the design of typing devices using ambiguity, the problem with mode design is that a number of keys already bear the burden of encoding one or more character symbols on each key, and the number of keys available for symbol encoding is generally severely limited. Is especially sensitive. However, the same methods applied above to generate the ambiguities for character symbols can also be applied to non-character symbols if the non-character symbols have a strong correlation.

The collection of symbols encoded by the keyboard can be divided into subsets corresponding to the modes. Depending on how many operations are required in the user part to obtain each mode and / or how often the symbols in each mode are used, the modes may be at least partially arranged. Therefore, the primary, secondary, and tertiary modes can be defined in increasing order of the number of operations required to obtain each mode and / or decreasing probability of the symbols in the mode.

Character symbols are preferably placed in the primary mode. A sensitive problem in design relates to methods of assigning non-letter symbols to modes and arranging the spatial arrangement of each mode.

The first statistical measure to consider is the probability of non-letter symbols. Some non-letter symbols, such as punctuation or numbers, are essential for communication and can occur at frequencies comparable to or exceeding the frequency of letter symbols. These punctuation marks are candidates included in the primary or secondary symbol set in any keyboard design. Considered next is the correlation arising from the non-character acting with other non-literal symbols. Some non-letter symbols have a customary and conceptual relationship with other non-letter symbols, for example, the symbols (left parenthesis) and (right parenthesis) are related to each other when they act together to express meaning. . The symbol. Is related to the symbol, because the two symbols express similar meanings (phrases and sentence termination). These examples are examples of general relationships common to most uses of the language, including languages such as English. There are other more local relationships that may be mentioned in the design of keyboards for specific purposes, such as: and / in the expressions of: / which are commonly used in site addresses on the World Wide Web (URLs or Universal Resource Locators).

Non-literal symbols can also have a statistical, customary, and conceptual relationship with literal symbols. For some symbols, it is possible to analyze statistical relationships with each other in relation to verbal warfare. For other symbols, such symbols may never appear in the text, so special software is needed to capture these symbols for user research or statistical analysis. Examples of these may be "backspace", "page up" and other symbols used to edit, inspect, or otherwise manipulate the text.

Given the reference statistics constructed in this way, the next step in assigning symbols to modes is to arrange them in such a way that the statistical, customary and conceptual relations are best met. Another constraint that can be considered is the memory potential in the arrangement. Preferably, all the symbols are arranged in mode in a manner that is "comprehensible", that is, the pattern of symbols is simple, familiar, and preferably visually well constructed. Even well-trained touch-typed hitters can go back to visual scanning to find rarely used symbols on the keyboard. Thus memory potential may be a top concern in the arrangement of modes used for less frequently used symbols. It will be appreciated that memory potential can be quantified using experimental protocols for memory tasks that are well known to psychologists.

Letter symbols [a-z], numbers, and 32 non-letter symbols found on standard keyboards to illustrate this approach. @ # $% ^ & * () _-+ = {[] | \:; <,>. ? Examples of placement of / are provided. This arrangement consists of three mode conversion keys and three modes in which each mode includes sixteen symbol keys, as shown in Figs. 36A-C. This arrangement is designed for PALM PILOT class devices. It does not show that it is optimal by psychological investigation.

The primary mode (FIG. 36A) includes an ambiguous sign for characters, a space / backspace key, a basic punctuation key, and a state transition key in the forward or backward direction.

Secondary mode (FIG. 36B) includes numbers arranged for functioning as a telephone or basic calculator / numeric keypad and keys for specific punctuation.

The tertiary mode (FIG. 36C) can be used as an aid to remember that the shift key has associated meanings (e.g., open and close parentheses) in relation to symbols, or if there is no related meaning, Include additional punctuation marks arranged to relate to related symbologies. This applies the promise that hard symbols, ie angled symbols, are on the left, while soft symbols, ie curved symbols, are on the right. Since all or most of the keys have exactly two symbols, the ergonomic decryption technique can be operated in each mode.

Selectable Keyboard Transparency: Typical implementations of keyboards currently displayed on PALM PILOT-class touchscreen devices consist of keyboards that occupy a portion of the screen, while the rest of the screen receives input from the keyboard-for example, an address book application. Used for program. The display area of these devices is extremely limited, resulting in very small keyboards and applications due to display sharing between the keyboard and applications. When the keyboard used in this apparatus is given in an extremely small size, it is not actually typed in a touch manner, nor can it be. However, through the application of the methods of the present invention, even a limited environment of the touch screen of a personal digital assistant produces a keyboard that is small enough to be typed by touch. If the keyboard can be touched, it is an important observation that it does not need to be displayed to the user. The user's finger "knows" where the keys are without the user seeing the keys. Thus, keyboards can be made transparent while occupying the entire touch screen, while applications can occupy the entire touch screen as well. When displayed in this way, the user can type directly into the application and enter it into the application. In Fig. 35, the keyboard 1003 displayed in this manner is shown with the application 1002- drawing program which displays a picture in this case. Since it is impossible to draw transparent keyboards, in this figure the keyboards are shown in shades of gray while the application is shown in black. In fact, it would be possible to allow a user to select a transparency class of a keyboard, for example, as a function of his or her touch type typing skill class.

Other modes have already described that intimacy with the user may include different classes of symbols. To account for these differences, the transparency of the keyboard can also be adjusted to gradually decrease as the symbols' familiarity of the mode increases, depending on the mode function. In addition to transparency, it should be noted that other visual factors such as image color can be adjusted to achieve the same result as distinguishing a keyboard for an application.

Hybrid coding / ambiguous keyboards: An important aspect of the present invention is that coding patterns where only two keys need to be activated substantially simultaneously, e.g., coding patterns for encoding capital letters in a Quartie keyboard, are easy to learn. It can be adopted by multiple users. However, the prior art has shown that complex coding patterns are generally unacceptable.

We mentioned that there are two main prior art approaches to typing devices with a small number of keys: coding schemes and ambiguous schemes. One aspect of the present invention is to teach how to synergistically combine the two approaches.

Two kinds of coding methods 1) A method of holding a key or several keys for the function of forming chords, an example of which is well known for capitalization by the coding combination of shift and character keys. It is a shift key, and this key will generally be referred to as a shift key, and 2) it is possible to distinguish how a number of character keys are activated substantially simultaneously so that a code is formed. The first of the above methods will be used in this embodiment, and the second will be applied in the next embodiment that follows.

An important insight into this aspect of the invention is that activating a pair of input means substantially simultaneously is easily integrated into a single gesture of the user. Thus, a pair of key presses are hardly or hardly difficult compared to a single key press, but a pair of key presses contain substantially more information than a single key press, and thus a low ambiguity code that can be easily operated and this code It can be used to create a typing device based on the field. Thus, to make the keyboard easier to learn, the code requires a pair of keys to be activated substantially simultaneously. Such an ambiguity using up to two keypresses (symbol encoding) to encode any character (decode) will be referred to as a maximum-2 ambiguity (see FIG. 50). In this embodiment, one of the pair is the key held to form the codes. This coded symbol will be referred to as a combined coded symbol, see FIG. 51, the rest being a key corresponding to at least one decrypted symbol. This key will be referred to as decrypt-sign-assign key or character-assign key. Uncommon symbols will require two or more keys to be activated substantially simultaneously, and very common symbols may be associated with a single input means without going beyond the scope of the present invention.

At least one of the decryption symbols is preferably a symbol of strong correlation. In this way, if the shift key is activated when it does not need to be activated, or is not activated when necessary and the code is improperly formed, the unambiguous software will be able to correct this error.

Thus, in accordance with the teachings of this embodiment, when the coding and the ambiguities can be combined in various ways, the preferred combinations are

In order to encode practically possible symbols, there are no more than two input means required to be activated substantially simultaneously. In other words, a maximum-2 ambiguity code is preferred. See FIG. 50.

Lookup errors and / or query lookup errors are optimized.

Coding is accomplished using mode shift keys, i.e., combined coded symbols.

The (preferably) mode shift key usage probability is minimized, i.e., the combined coded plus decrypt-symbol-assigned-key combination is activated less frequently than with the decryption-symbol-assigned key alone.

When combining coding and ambiguity schemes according to these teachings, surprising synergy results are obtained and will be demonstrated by this embodiment of applying a complex coding / ambiguity scheme to a telephone that implements standard ambiguities. We have already seen that standard ambiguities have rather poor lookup and query lookup errors. Thus, using the hybrid scheme, a keyboard based on the standard ambiguity code can be made for (class C) strong touch typing, and the standard ambiguity with six valid keys can be converted to a code with thirteen valid keys. It is amazing that there is. The purpose of this embodiment is to produce a keyboard using a minimal number of typing gestures that is easy to operate, easy to learn, and fully compatible with standard phones that implement standard ambiguities, which can be typed in a strong touch.

delete

A standard telephone implementing standard ambiguity code is shown in FIG. A number of keys 10000 are used to encode letters and numbers. There are eight keys, two keys 10001 and 10002 encode only numbers, and two keys 10003 and 10004 encode the non-letter symbols * and #, respectively. In this embodiment, one of the keys selected from the group consisting of 10001, 10002, 10003, and 10004 will be used as the mode conversion key for generating the combined coded symbol, and is preferably the key 10001 for encoding the number 1. The key selected in this way may be referred to as a shift key for reasons which will be apparent, although other means may be used for generating the combined coded symbols. The key 10001 can be conveniently activated by the thumb of the left hand when holding the phone with the left hand, and the right hand is used to activate the remaining keys. In an embodiment where the right thumb is used to activate the shift key, hold the phone with the right hand and the key 10004 can be used as the shift key.

For each key of multiple keys 10000, the corresponding characters will be divided into two subsets, each referred to as a shift set and a non-shift set, respectively. That is, the shift set includes a maximum-2 sequence of encoding symbols consisting of one combined coded symbol and one key-assigned coded symbol, and the non-shift set includes a single key-assigned coded symbol. It is noted that the elements of both the shift- and non-shift set of a coded symbol sequence correspond to one or more decoded symbols under ambiguity.
To minimize lookup errors and to minimize query lookups, one of the sets does not lose generality, so that the shift set contains exactly one character per key, i.e., the combined coded and key-assigned coded pairs in the shift set Unambiguously mapped to a single decryption symbol, when the pair of encoded symbols is input, the associated decryption symbol is restricted to be explicitly selected as unambiguous output, see FIGS. 51 and 52, so that the probability of activating the shift key is minimized, Characters will be assigned to sets (shift, non-shift).

delete

delete

delete

delete

delete

Typically, simultaneous optimization of lookup error rate, query query error rate, and other constraints implies a compromise. For example, improved lookup error rates and queries by removing the just established restriction that one of the (shift, nonshift) sets are mapped to a single decryption symbol, thus varying the number of unique characters per key and the number of characters in the shift set from key to key. You can potentially get a query error rate. However, the regularity of division into shift and non-shift sets makes the keyboard easier to learn-this is a high priority constraint. Singleton meetings are probably easier to remember by mnemonic, so choosing a pair can further improve learning. However, this choice can compromise the lookup error rate and query query error rate.

There are 11664 pairs of shift / nonshift sets that follow the constraint that the shift set contains one character per key. This number is small enough that all of these characteristics can be tested. All these sign test results are shown in FIG. 39, where the lookup error rate is plotted against the query query error rate for all 11664 codes as well as the standard ambiguity code SAC. It is noted that all codes are better than standard ambiguities but most are only a few times better. However, this distribution is very large, and the best code CEHLNSTY with a lookup error rate of 431 words / lookup error and query lookup error rate 21 words / query lookup is 15 times more than the standard ambiguity for lookup errors and 10 times more for query lookup errors. Good. This best sign is abC dEf gHi jkL mNo pqrS Tuv wxYz, where elements of the shift set are capitalized. It should be emphasized again that although the code is the best sign with respect to the reference statistic and appears among the best ones for the statistics obtained from many different language wares of English, other statistics can produce other best signs. The same hybrid coding / ambiguity scheme can be applied to any ambiguity, and the ambiguities significant within any ambiguity are not constrained by standard ambiguities, in fact alphabetically or by only eight keys, or possibly even division. It is not constrained to have. By allowing more freedom to choose a code, complex coded / ambiguous codes can be found with very high quality in terms of lookup error rate and query lookup error rate, so other constraints such as minimizing the use of shift keys It can be advantageously combined with the optimization of query query error rate. This optimization is beyond the scope of the present embodiment to find sufficient compatibility with existing telephones, but is within the scope of the present invention.

To further facilitate the learning and operability of the keyboard, as shown in FIG. 38, the characters forming the shift set are preferably represented on the corresponding keys using capital letters, and the characters of the non-shift set are represented by lowercase letters. Alternatively, the two sets can be represented in different sizes, colors, typefaces, and so on.

The use of this device is simple. When the text to be typed contains one character of the shift set, the shift key must be operated substantially simultaneously with the corresponding character key, which character will be represented unambiguously. Conversely, when one character of a non-shift set is needed, the corresponding character key is activated and the character is ambiguously represented.

In view of the teachings of the present invention, the unambiguous technique corresponding to this embodiment may be physically located within the telephone, on the sending side of the communication, and on the receiving side of the communication-for example, on a central computer that the user contacts over the telephone. .

As mentioned above, there are four keys on the standard telephone keypad that can be used to encode non-letter information such as mode switching. Using the method of composite coded symbols described above, the number of non-letter symbols that can be coded using the telephone keypad can be further increased. In particular, additional shift keys may be provided if lower ambiguity classes are required. For example, with each of the four shift keys associated with one character on each character key, an unambiguous text input can be achieved, while the number of key presses per character is increased. In summary, one can think of the association of a subset of decryption symbols with a shift key. Certain classes of assignments will be considered later in the section on internationalization of this teaching, and so far have been described mainly in detail in English.

It can be used to encode at least six non-literal symbols, for example punctuation symbols, state transition symbols, etc., using a shift key that combines the remaining non-magnetic keys, which are the *, #, and 0 keys in the preferred arrangement. Will be understood.

Error correction using standard ambiguity: In particular, for use by inexperienced users, the keyboard of this embodiment operates by pressing the shift key when it should not be pressed when encoding the intended text and without pressing the shift key when it should be pressed. Can be. This operation will often give meaningless decoding if the decoder expects correctly typed coding sequences in the complex coding / ambiguity. In this case, rather than querying, other unambiguization of the equation may be attempted where the shift key activity is ignored and the encoding sequence is interpreted as the encoding sequence of the standard ambiguity code. Frequently, this interpretation will restore the text that the user intended.

In the apparatus shown in FIG. 38, it is noted that a strongly correlated symbol (space) is paired with a weakly correlated or uncorrelated symbol (backspace) in the same key. This pair potentially allows undecrypting software to correct the error, allowing the user to hit the (backspace) key but mean (space) or vice versa.

Touch type typing-oriented query inquiry: In the present embodiment, when query inquiry is permitted as described above, it is preferable to use the selected shift key as a scroll key for query inquiry. It will be appreciated that at any given moment, whether the key functions as a shift key or a scroll key can be determined automatically when given the appropriate software. When the device is in query lookup mode, the key acts as a scroll key, otherwise it acts as a shift key.

Alternative Arrangement of Shift Keys: Referring again to Fig. 38, it is noted that the present embodiment is designed to be able to operate using an existing standard telephone. If one considers manufacturing a telephone using this embodiment, the telephone is preferably equipped with additional keys or additional keys 10005 to act as shift keys. These additional keys are preferably located on the side of the phone, which can be activated by the thumb of the hand holding the phone: This arrangement is shown in FIG. Additional keys can be used that can be actuated with the fingers of the held hand (left or right hand) 10006.

Low frequency word screening in query queries: Very high frequency words are often ambiguous with very low frequency words. For example, in the case of CEHLNSTY for English, the very frequent word "for" is ambiguous with the very low frequency word "fop". By removing these very infrequent words from the dictionary, the effective query query error rate can be improved with little impact on how well the dictionary represents the word in question. For example, the query query error rate for CEHLNSTY can be improved to one query query every 46 words by eliminating words whose total probability is less than one-fifty thousand. Such a selection can be achieved by applying a "gap factor" given by the ratio between two words in the query query, for example, the most frequent and the least frequently. If, for example, the spacing factor is set to 500, the distribution of Fig. 39 is obtained, in which two codes, the standard code (SAC) and the CEHLNSTY code, are specially pointed out.

Internationalization: There are two important problems in the internationalization of this embodiment, both of which can be easily solved by those skilled in the art by applying the teachings of the present invention. These are 1) accent processing and 2) generation of generalized codes that can be applied to many languages at once. These issues will be discussed briefly.

" 또는 "

"로 기록될 수 있다. 이러한 악센트 문자들을 구별하는 것은 중요하다. 악센트 없다면, 예를 들어, 문자 "

l

v e "("학생" 의미) 는 단어 "

l e v

"("올린" 의미) 단어와 혼동될 수 있다. 하나의 자연스런 접근법은 다른 시프트키를 사용하는 것이고, 이것을 악센트 시프트키라 칭할 것이며, 이러한 키는 문자를 부호화하는 키와 결합하여 사용될 때, 그 문자의 악센트 버전을 선택하는 기능을 가진다. 예를 들어 CEHLNSTY를 적용하여 프랑스어를 다룰 때, " 'e " 혹은 " 'e " 를 부호화하기 위하여 "def" 키와 결합하여 악센트 시프트 키를 칠 수 있고, 그 후 이 둘 악센트 중 어느 것이 주어진 단어에 적합한 지를 결정하는 것은 비모호화 기법에 맡긴다. 이러한 접근법을 사용하여, 프랑스어에 대한 단어-빈도 통계의 집합으로, CEHLNSTY에 대한 (룩업에러, 질의조회)율은, 악센트 시프트키를 가지지 않을 경우 (38, 3)이지만 악센트 시프트키를 가진 경우는 (584, 24)이다. 예를 들면 전화 키패드에서, 맨 아래 열에 있는 임의의 키들이 악센트 시프트 키로서 사용될 수 있다. 특별히 제조된 키패드에서, 악센트-시프트 기능을 제공하기 위해 추가 키가 제공될 수 있다. 인간공학적 이유로, 이러한 악센트-키는 일반 시프트키를 작동하는 방식과 비슷한 방식으로 작동되는 것이 바람직하며, 예를 들어 한 방향에서의 엄지 손가락의 움직임은 일반 시프트 작동을 부호화하는데 사용하고, 반대 방향에서의 엄지 손가락 움직임은 악센트 시프트 작동을 부호화하는데 사용될 수 있다. Accent Handling: Many words are written in letters that appear in accented and unaccented form. For example, in French, e means "e", "

" or "

It can be written as ". It is important to distinguish between these accented characters. Without accents, for example, the letter"

l

ve "(" Student "means) is a word"

lev

Can be confused with the word "(" up "meaning) One natural approach is to use another shift key, which will be called an accent shift key, which, when used in combination with a key that encodes a character, For example, when dealing with French by applying CEHLNSTY, you can hit the accent shift key in combination with the "def" key to encode "'e" or "'e", Determining which of these two accents is appropriate for a given word is then left to an unambiguous technique, using this approach, as a set of word-frequency statistics for French, with a (lookup error, query lookup) rate for CEHLNSTY. Is (38, 3) if you do not have an accent shift key, but is (584, 24) if you have an accent shift key, for example, on the telephone keypad, Keys in the right can be used as accent shift keys In a specially manufactured keypad, additional keys may be provided to provide an accent-shift function.For ergonomic reasons, these accent-keys differ from the way in which normal shift keys operate. It is preferable to operate in a similar manner, for example the movement of the thumb in one direction can be used to encode a normal shift operation and the thumb movement in the opposite direction can be used to encode an accent shift operation.

Multilingual Ambiguity: Since the statistics of one language are generally different from the statistics of another language, a sign that is substantially optimal for one language may not be substantially optimal for another. A code that can be typed in a strong touch manner for one language is not a type that can be typed in a strong touch manner for another language.

For example, CEHLNSTY, which is optimized for English, performs poorly than the code specifically selected for optimality for French. In this particular example, CEHLNSTY can be typed in a strong touch on French, but this may not be the case for all languages.

For economies of scale, manufacturers will want to manufacture a single machine that can operate in many local and linguistic environments. Since a typing device, for example a cell phone, may preferably have keys marked with an ambiguous sign, it is useful to have a single sign that applies to many languages, and this marker is useful for all machines manufactured regardless of the target language community. Can be labeled in the same way.

Applying the exact same technique as described above, it is possible to generate ambiguities that simultaneously optimize for several different languages. In a multi-language optimization scheme, the step of weighting constraints on each other may include a step of weighting many languages for each other. In other circumstances different weighting schemes are appropriate. For example, one may choose to optimize for statistics in English and German at the same time, but with respect to English the execution of the sign may be optimized as more important than the implementation for German.

The preferred weighting method is a method of maximizing minimum performance, and this procedure will be referred to as a mini-max procedure.

Consider the optimizations for the set l1, l2, ..., ln of the language and the set e1, e2, ..., em of the ergonomic constraints. Given two ambiguities c1, c2, for each constraint em, c1 is better than c2 if the minimum value of em for language ln for c1 is greater than the minimum value of em for language ln for c2. Grade one. When some constraints are optimized, one sign is better than the other in the minimum-maximal meaning with respect to one constraint, but is bad with respect to other constraints. In this case, as mentioned above, the constraints must be weighted.

As an example of these teachings, consider optimizing for lookup and query lookup errors for a set of languages. In this example, an arrangement other than an alphabetic arrangement is allowed, and using eight ordinary input means, auxiliary input means and accent-shift auxiliary input means, the eight ordinary input means are combined code / ambiguity code as described above. In an embodiment it can be used in combination with one of the auxiliary input means.

First, consider an optimization for a language set consisting of French, Italian, Portuguese, and Spanish, where each language is represented by a set of reference statistics.

Using managed random behaviors (lookup error rate, query error rate) is (3250, 265), (11400, 3800), (4720, 505) and (6280, 400) for French, Italian, Portuguese and Spanish, respectively. Easily find signs like joz m bhx a kn r pw d iy l gq t ev c fu s This code shows relatively poor performance for Dutch, English and German with values of (65, 4.8), (93, 10) and (360, 13), respectively. Although similar computer processing time is used, Dutch, English When optimizing for and German, these languages find the cjk r biy l fv e mo a sz p hx g tu d qw n signs showing (1220, 44), (816, 44) and (480, 47), respectively. Can be. This symbol shows (253, 20), (306, 50), (525, 36) and (4236, 272) for French, Italian, Portuguese and Spanish, respectively. These results are quite good for languages other than sample languages, but the ambiguity about these languages is worse than the results obtained when explicitly optimizing. These results suggest that as languages become more and more different, the performance for one language is less generalized to that for another.

delete

In a real environment, it is more commercial and conceptual to determine which languages are included in a multilingual optimization plan. An important progressive concept by the teachings herein is that the sign can be typed in a strong touch, even with minimal performance for selected languages. In the case investigated above, the optimizations for French, Italian, Portuguese and Spanish find a sign that can be typed in a Class C steel touch on these languages, and the minimum performance for Dutch, English and German is Grade A steel. It can be typed in a touch manner.

Strong touch type handheld device that can be typed using one hand: In the hybrid coding / ambiguity embodiment described above, distinct input means (combined coded symbols) are ambiguously encoded in order to reduce the ambiguity of the entire system. It has been found that an input means (decoded symbol-assigned coded symbols) for encoding coded symbols and how they can be used to form codes. This embodiment shows how the decryption-signal-assigned coded symbols can be used to code coded and ambiguously coded symbols. That is, a given coded symbol functions as a combined coded symbol and a decode-symbol-assigned coded symbol. In this way, the ambiguity code can be represented as a multi-class code: a first sequence of coded symbols is used to select a first subset of decodes, and a second sequence of coded symbols uses a second subset of decoded symbols. Use it to choose, and so on. Preferably, the second subset is a subset of the first subset, the third subset is a subset of the second subset (and therefore a subset of the first subset) and the like. This is a "divide and conquer" approach well known to those skilled in the art. However, a) in the division and acquisition approach, it is common for the final subset to have a unique solution, but the number of consecutive subdivisions of the symbol set may be limited by allowing the smallest subset to contain one or more symbols. And that these smallest subsets can be used to define ambiguities, b) subdivision schemes can be chosen to minimize the ambiguity of the final ambiguities (lookup errors and query lookup errors), c) customary While simultaneously optimizing other constraints such as preservation, ambiguity reduction can be optimized, d) conversion between classes in a hierarchy can be achieved with codes consisting of only a pair of key presses, where a pair of elements Have not understood so far that they may be combined coded symbols and / or decode-symbol-assigned coded symbols, depending on the context.

The specific manifestation of these findings will be described in detail with reference to FIGS. 39 to 47. It will be appreciated that this is just one of a myriad of devices that can be fabricated in accordance with the teachings of the present invention: relying on a division and acquisition approach in sign construction, while minimizing other constraints such as obscurity and / or obsession with conventions. Any typing device is included within the scope of this embodiment.

This embodiment is a device that can be typed in a strong touch with one hand. An additional desirable feature of this is that

1) fully unambiguous text-input mode is possible,

2) a vague text-input mode is possible which is substantially optimal and can be typed in a strong touch,

3) Minimal key press mode for data retrieval is possible,

4) The above three modes can be compatible at best.

Advantageously, the device can be manufactured in a form that is optimized for the new constraint scan time in addition to the aforementioned constraints. Optimization of scan time will be discussed below.

An overview of a method for building a typewriter device based on a multi-class ambiguous code is described with reference to FIG. 39.

In a first step 150, a second set of decryption symbols is selected. These are symbols represented by ambiguities and may include the letters a through z in the English example.

In a next step 151, constraints for over-all multiple class codes are selected. Such constraints can be, for example, strong touch type typing or lookup errors or query lookup errors alone. Many constraints of multiple class codes can be selected simultaneously. In a next step 152, the second degree of decryption symbols are divided into subsets. A coded symbol is assigned to each of the subsets of the second class in such a way that the total code is optimized with respect to the selected constraint. The construction up to this point is no different from optimized ambiguity construction. However, there may be additional constraints on executing the next construction phase, for example, on the number of allowed code symbols. In a next step 153, the second class of encoding symbols are gathered into groups. These groups are treated as decipher symbols for the first class ambiguities. In other words, the coded symbols of the second class code are the decoding symbols for the first class code. Therefore, the coded symbols of the first class are assigned to each class while forming the ambiguity code of the first class. Further constraint optimization can be carried out by assigning second-class symbols to groups. In general, each class of multiple class codes can be optimized for different constraints. These constraints may be the same as or different from those for which the total multiple class codes are optimized. In a final step 154, the multi-class code thus constructed is implemented in the typing device.

It has been described that the construction of the second class code precedes the construction of the first class code. When using a device with a multi-class code implemented, the code is applied in the reverse order in which the codes were constructed. First, the element of the first class code is selected by keyboard activation, and then the element of the second class code is selected by keyboard activation of the second user.
This is the essence of the sharing and acquisition approach. It will be apparent to those skilled in the art that this construction may continue in the same way to include the third and higher class codes.

In practice, the properties of each class of multiple class codes must be optimized simultaneously to obtain the desired properties of the total multiple class codes. This embodiment is presented to specifically illustrate how such simultaneous optimization can be planned and executed.

An overview of the method of building this embodiment is shown in FIG. 40. To present the procedure in a fully generalized manner, three constraints are chosen for the total multiple class codes, two constraints for the first class code, and three constraints for the second class code. The three constraints applied to the total multi-class code in this embodiment are strong touch typing, optimized query lookup error and optimized lookup error. The first class code is optimized for anatomical fitness and alphabetical arrangement, and the second class code is optimized with regard to partition uniformity, anatomical fitness and substantial alphabetical arrangement.

In this embodiment construction, the first step 3100 is to select the (second degree) decryption symbols. These are the letters a-z. The strong touch typing, the query query error optimization, and the lookup error optimization are then selected as constraints for the multiple class codes, respectively, in steps 3101, 3102, and 3103. Then, in step 3104, anatomical fitness is selected as a constraint. Since the device must be able to be typed using the fingers of the hand holding the device, anatomical fitness is optimized when there are four input means and four corresponding first class coded symbols, one for each finger.

Anatomical fitness is selected as a constraint on the second class code in step 3105. Each coded symbol in the first class code will correspond to some coded symbols for the second class code. If each of the four first class symbols corresponds to four second class coded symbols, the anatomical fitness of the second class code is maximized, so the number of symbols of the second class code is 16, the maximum of which is the anatomical fitness. Should be. Sixteen second class coded symbols may be associated with second class coded symbols so that when 26 second class coded symbols are distributed over the second class coded symbols, one or two second class coded symbols may be associated with the 16 second class coded symbols. The equality of division is maximized if associated with each of the coded symbols. This distribution again means that between 4 and 8 second grade deciphers will ultimately be associated with each of the 4 first rank deciphers.

Next, in step 3106, the alphabetical arrangement is selected as a constraint for the first class code. Optimizing for this constraint requires simultaneous optimization of both the first and second class codes. What is needed is that the letters a-z correspond to the input means associated with each finger so that they can be displayed in alphabetical order on the display. Since these outputs are arranged in the order of the fingers, this again means that the letters of the first part of the alphabet must be concatenated with the second degree deciphers, and subsequently, the second degree deciphers are concatenated with the first degree deciphers, and the first degree decodes. The symbol means that the input means is connected to the first finger. In the same way, the second group of characters following the first group of alphabetical order should be assigned to the second class coded symbols, the second class coded symbols are connected to the first order coded symbols, and the first order coded symbols to the input means. This is also the case for the other two first class coded symbols, such as the input means being connected to the next finger. Thus, optimizing for alphabetical order corresponds to selecting an ordered division of 26 characters in the same manner as discussed for other embodiments of the present invention. This time, each of the four elements of the arranged partition must have between four and eight subelements so that all of the listed constraints can be optimized at the same time. As can be seen from the detailed description of the optimal state for this embodiment, when optimized with respect to all of the other constraints considered, one can find signs with even division, even for the first class code.

Finally, in step 3107, the actual alphabetical order is selected as a constraint of the second class ambiguity code. This means that it is possible to arrange the characters in an alphabetical arrangement, given all other constraints on assigning the character to the second class coded symbol. Divergences from a strict alphabetical arrangement can be measured in a number of ways, for example by the number of pair permutations required to make a given arrangement a strict alphabetical arrangement.

Referring to Figs. 41-47, a handheld device which can be typed using one hand, can be typed in a strong touch manner that encodes at least the letter [az] and is implemented according to the method described above. do. In order that the device made according to the present embodiment can be typed in a touch manner, dividing the symbols into subsets, subsets of subsets, and the like should be fixed, i.e., what symbols have been previously entered. It should not change accordingly. This fixation only concerns the requirements of touch-type typing, and the teachings of the present invention can be applied in a wider range of situations. For example, by enabling word completion techniques, the number of key presses can be significantly reduced. However, because word completion techniques are complex and difficult to predict, machines with word completion cannot be typed in a strict sense by touch. Nevertheless, optimizations that provide a strong touch typeable code provide an effective word completion technique because lower ambiguity can produce better word-completion. Thus, increasing the strong touch typing code with the word completion technique does not make the device beyond the scope of the present invention.

For this embodiment, the symbol input typeable part should be able to be held by one hand, and there may be another limitation that the hand holding the device, only the hand, can be typed. In order to limit the requirements for numerical movement, most of the symbols have only five input means: one input means 2104 which can act as the thumb of the hand holding the device, and one finger which can be moved by the finger of the hand holding the device. It can be input through sequences of operations of the four input means 2100-2103. The device shown in FIG. 41 is held by the left hand; It is clear that a symmetrical device designed to be held by the right hand or a two-handed device that can operate with the left or right hand can be constructed.

Preferably, each of the input means 2100-2103 is associated with a visual display 2106-2109 showing the elements of the subset connected with the given input means. Operating the input means selects the corresponding subset. Input means 2104 may be used to sort subset selections and / or to select another subset of symbols. For example, the single symbol "space" can be associated with the input means 2104; This or other symbols connected to the input means 2104 may preferably be displayed on the display means 2110. The letters [a-z] may be distributed to four input means 2100-2103. The character distribution to the input means is preferably selected so as to minimize the ambiguity of the resulting code (lookup error rate and / or query query error rate) while simultaneously preserving the conventions of alphabetical arrangement. This preservation helps inexperienced users find the necessary characters by simply scanning the candidate characters.

42 shows the arrangement of the letters [az], the letters [af] are connected with the first input means 2100, [gl] is connected with the second input means 2102, [mr] is connected with the third input means 2102, Finally [sz] is connected to the fourth input means 2103. These associations form the first class subsets in the first class code. It is generally preferred that 4-8 characters be associated with each of these four input means: whereby a subset of the characters connected to each input means can be further subdivided into four subsets. It will soon be apparent to use this limitation, and it will be apparent to those skilled in the art how the teachings of this embodiment extend to languages with different numbers of symbols and different numbers of input means.

An example of a second class subset that divides the first class subsets shown in FIG. 42 is shown in FIG. 43. 43 is a table with four rows and four columns. Rows are marked with input means that are active in the first stage, and columns represent symbols connected to each input means in the second stage. Thus, for example, if the input means 2100 is activated first, in the second step the symbols ac will be connected to the input means 2100 and be will be connected to the input means 2101. This assignment is chosen to minimize the lookup error rate and query lookup error rate, given the constraints on the subset size. The lookup error rate and query query error rate for this code are (1100,69) using reference statistics. It is very carefully noted that in this example the letters of the first class subsets can be sorted in an alphabetical arrangement, but the letters of the second class subsets can only be partially sorted. In order to generate a sign that can be typed in a strong touch that allows for a good query lookup error rate and lookup error rate, this example has loosely set the alphabetical array constraints in the second class. This shows whether the alphabetical arrangement is optimized or not like other constraints, and shows that weighting the optimization properties may vary from one class to another in the multiple class ambiguities. Again, the advantage of alphabetical arrangement is to reduce scan time, especially for inexperienced users. Since the number of symbols shown in the second class is small, the scan time is small in any case and can also be further reduced by the techniques currently discussed.

In order to type a given desired character, the user first activates one of the input means 2100-2103 corresponding to the first subset containing the desired character. The user then activates one of the input means 2100-2103 corresponding to the set of characters containing the desired character again to select one of the second class subsets. 44 shows an example of operation of a device in which a user types the letter e. Referring to FIG. 42, e is connected to the input means 2100 by the first grade code. The user activates the input means, and the screen output is as shown in FIG. The letter e is now connected with the input means 2101. The letter e is output when the input means is operated. The same sequence of input means operations is also used to select the letter b so that the sign is ambiguous. As in other embodiments, which letter b or e is intended is determined in the context of an unambiguous technique. In this way, the necessary letters are successively selected, the input means 2104 connected with the thumb of two-hand operation is activated, and the words are terminated to enter the words. It is noted that this doublet method can form the basis of one-handed / two-handed embodiments, since a double-tap method can be used to encode each character and a maximum-2 sequence of symbol coding can be used. More specifically, if one hand is used to indicate the first press of each character and the second hand is used to indicate the second press of each character, the first and second press information can be entered at the same time. There are various physical embodiments that can be based on this. For example, the "fingering" technique of [4] can be a physical entity upon which one- and two-handed embodiments can be based. The sign proposed by [4] is based on a motion sensor that can sense several positions per finger to unambiguously sign each character. This requires a relatively sophisticated sensor. However, by applying the two-handed variant of the present embodiment, simpler sensors can be used. These sensors will only need binary (up / down) information for each finger. The complexity of both software and hardware can be reduced in this way. Moreover, machines made in accordance with the teachings of this embodiment can be simpler for the user to learn and operate.

Visual Cache: Scan time is the time it takes to visually locate the desired character from a set of characters. A hunt-and-peck typewriter, typing only on the keyboard's characters, visually scrolls through the keyboard and presses the corresponding key to find the next character. The scan time is determined by several factors including the user's familiarity with the keyboard layout. Typing by typing only on the characters of the keyboard, the typewriter basically knows where the desired key is, and uses visual scanning only for confirmation or accurate positioning. Through familiarity with the alphabetical arrangement of ordinary users, it is to improve the scan time that the alphabetical arrangement is chosen for the first class code of this embodiment. In a variant of the alphabetical arrangement some characters are selected from a group of characters on the key in the distinct and selected area of the visual display associated with the key. These characters are the most selectable characters at any given moment, and placing them in separate locations makes them easier to find. This principle is similar to the cache used in some computer processing devices that store recently used data in registers that can be quickly obtained, assuming that recently used data is likely to be used again. Given language statistics, characters are placed in the cache based on their next availability, not based on their recent use. Thus, the term "visual cache" appears to be appropriate.

One embodiment of the time cache will be described in the present embodiment situation. It is understood that the invention is capable of various modifications without altering its essential quality. For example, variations in cache size and location, how the cache is constructed, how it is marked, and so on.

In the analysis of standard statistics, it can be seen that a is most likely to be a first letter of a word in the letters [a-f] connected to the input means 2100 by the first rank code. Similarly, in [gl] connected to input means 2101, i is most likely to be the first character of the word, o is most likely from [mr] connected to input means 2102, and t is the character [connected to input means 2103; sz] is most likely.

The letters a, i, o, t are placed at distinct positions of the display, for example the upper left hand corner of the display area associated with each input. This makes these characters the first characters found in the standard time scan from left to right and top to bottom of each connected display. Preferably, except this selection of only one character that is not in alphabetical order, the alphabetical arrangement is maintained for the remaining characters in that subset. The distinction between the characters in the cache and the remaining characters may be displayed by selecting, for that font of cache characters, a color, size, shape, etc., different from the remaining characters.

Referring now to FIGS. 45 and 46, it can be seen how this observation is used to reduce scan time. Figure 45 shows how the word "think" is entered without using the visual cache, and Figure 46 shows the same word entered using the visual cache. Thus, in Fig. 45, the letter "t" is first inputted by activating the first input means 2103 corresponding to the letter t. The screen output shown in the second row of the drawing is before the first input means is activated. Once 2103 is pushed, the display changes to the one shown in the third line. Then when the input means 2101 is activated, the letter "t" is output. These displays change similarly when the remaining characters of the word "think" are entered.

In FIG. 46, the difference between characters in the time cache and characters not in the time cache is shown in lowercase letters that are not in the time cache and in uppercase letters in the time cache. According to the reference statistics, 42 percent of the first letters of words are a, i, o, or t. Therefore, when a user starts typing a word, for 42 percent of the time, the desired character will be immediately found in the cache.

When a word is entered, the next most likely letter changes as the context is created by entering that word. Thus, the character selected to be cached must change when a word is entered and will depend on which word is entered.

In the case of the word "think", as shown in the first four columns of FIG. 46-each column corresponding to the input means output-in both before the activation of the first input means and before the activation of the second input means, the letter t is in the visual cache. Is found. Once t is selected, the characters in the time cache are a, h, o, w as shown in the second set of four columns in FIG. After the first input means (input 2101) is selected for entering the letter h, there are only two possible characters that form parts of the words according to the reference statistics, these are the letters h and i, which are shown in the second display. Appears in the visual cache. When continuing this way for the letters i, n, k, we can see that the desired letter is always in the visual cache for this word. In fact, after the first input operation to enter the letter "i", if the user actually enters a word in the database, there is only one character that may have been intended. In this case, the second input means operation is unnecessary, and after the first input means operation, the letter i will be output immediately.

Obvious Unambiguous and Input of Additional Symbols: As already pointed out, it is desirable to provide a method that is not completely ambiguous for inputting symbols in typing devices based on ambiguities. In this embodiment, one simple way of providing an unambiguous input is to provide the additional non-ambiguous input means 2105 shown in FIG. 41 in a position that can be easily activated by the thumb, which is the preferred position. However, other locations may be selected.

In this example, the size of the second class subsets was chosen to be limited to at most two symbols. Thus, the unambiguous technique will always choose the desired symbol correctly, or incorrectly select the remaining symbols paired with it. If the corresponding input means has been selected by the user, any unambiguous software will give a signal indicating which of the two symbols to select. For example, this signal can be used to provide feedback to the user, such as to highlight the character to be selected. If the character to be selected is not the desired character, the user can activate the apparent unambiguous input means 2105 shown in Fig. 41, which allows the user to select the remaining selection, i.e., the inconspicuous symbol.

One example of using this non-ambiguous text input technique is shown in FIG. 47. As in Figures 45 and 46, this figure shows how the word "think" is entered. The fourth line here shows the characters that will be output if the non-ambiguous input means 2105 is activated after activating the first and second input means used to input the word "think" character. For example, if the input means 2103 and then the input means 2101 are activated to input the letter t, the subsequent activation of the input means 2105 will select the letter u. "U" will be the first letter of the word. All possible characters can be entered unambiguously in this way. When the second class subsets contain only one character, this one character will be entered unambiguously even when the input means 2105 is not active, so the input means 2105 is not applicable. The letters d, f, h, l, n and p are always entered unambiguously for a given sign.

Strong touch typing: measurements and thresholds

Strong touch typing is a new and progressive concept that describes a particular machine class. The breadth of this concept has been pointed out in various embodiments near the boundaries of this class, indicating a range.

In order to add more clarity to the initiation of the strong touch typing, the strong touch typing can be measured with any ambiguity and thus determine whether the sign is within the scope of the present invention. Other numerical features of sex will be proposed.

Linguistic Statistics: It has already been mentioned that expressing language with textual warfare is a research concern among linguists. For numerical clarity, we define a representative language battle as a collection of at least ten million words randomly extracted from newspapers of public interest in the target language.

Number of keys: It is necessary to define three kinds of keys: physical number of keys, number of coding keys, and number of valid keys. Physical key count: The number of inputs used to encode the symbols. The minimum qualitative keyboard has 26 keys, shift key and space key labeled, thus this has 28 physical keys. Number of coding keys: The number of distinct combinations of keys that encode the symbols. For a minimum qualitative keyboard, the shift key can be combined with the character keys to form a capital letter, so that the keyboard has a coding key number of 28 + 26-1 = 53. This is because the shift key alone does not encode the symbols. In other words, you can create a keyboard that is exactly the same as the minimum qualitative keyboard with 53 physical keys, each of which encodes only one symbol, either uppercase or lowercase. In fact, some early typewriters had this structure.

Number of valid keys: Given a set of symbols represented by an ambiguity code, a set of linguistic statistics, and the number of physical keys P, there is an optimal ambiguity with the best possible lookup error rate and query query error rate. It is the only constraint on the sign. These ratios are called Pl and Pq, respectively. The ambiguity code for any physical key number will have a valid key number P, if the lookup error rate and query query error rate are equal to Pl and Pq. It is impossible for a keyboard with a physical number less than P to support an ambiguity with a valid number of keys equal to or greater than P. It is sufficiently possible and common for the ambiguity for the physical key number P to have less than the effective key number: P, for example the lookup error rate of the substantially optimal ambiguities, as shown in the experimental results of FIG. It is an experimental observation that the query inquiry error rate is actually related by the power law. The log of the query query error rate that is substantially optimal for English is a log-linear relationship of the substantially optimal lookup error rate.

Through this observation, it is possible to define a single number associated with the lookup error rate and query query error rate of one code: project the point of code (lookup error rate, query query error rate) from the log-log plot to the best-fit line. (projection).

For example, consider standard ambiguities. This code has (29, 2.2) (lookup error rate, query query error rate). Projecting this point over the best-fit line of FIG. 11 to linearly interpolate, we can find the 5.96 value, which is the number of valid keys in the standard ambiguities. Although a standard ambiguity is defined over eight physical keys (and eight coding keys because no coding is included), it is equivalent in ambiguity to the actual best sign on the 5.96 physical (or coding) key. Of course, a physical key with a decimal point is not really possible, but the result shows that you can find a substantial ambiguity in the six keys, which has a better lookup error rate and query query error rate than the standard ambiguity code.

Through these observations, one can define an accurate but arbitrary numerical threshold for the actual optimality of the lookup error rate and query query error rate: One sign is that if the number of valid keys is within 0.01 of the number of coding keys, and Without other ergonomic constraints on the system, one would say that it is practically optimal for these ratios. It is also possible to define an accurate but arbitrary threshold for strong touch typing: The ambiguity for English can define that a strong touch can be typed if the number of valid keys is at least ten. This definition can be extended to other languages by requiring the code to be typed in a strong touch method, and is greater than or equal to the lookup error rate and query lookup error rate of the strong touch type code for English. Should have Since the number of valid keys in the standard ambiguity is less than 10, it cannot be typed according to the strong touch method accordingly.

By measuring the number of valid keys, it is possible to screen for any ambiguous code as to whether or not it possesses the strong touch type magnetic properties. For example, complex coding / ambiguity code: ab c df e gi h jk l mo n pqr s uv t wxz y has 9 physical keys: 1 shift key and 8 standard phone keypad characters key. This is 16 coding keys; This is equivalent to the ambiguity of 16 independent keys without a shift key. Without applying the interval factor, its (lookup, query) error rate is (431, 21), which corresponds to the number of valid keys of 12.8. Applying the spacing factor 500, this is improved to (440,46), which corresponds to 13.75 valid key numbers. With or without a spacing factor, this code can be typed in a strong touch. It is noted that the number of valid keys is less than the number of coding keys, and this sign is substantially optimal given the additional constraints of the alphabetical arrangement. In this way, a complex coding / ambiguous code embodiment for one hand has a sign with (lookup, query) error rates (1100, 69), whose physical key number is 4 and coding key number is 16, The number of valid keys is 15. This is a strong touch typing code, and the difference between the number of coding keys and the number of valid keys is due to additional constraints on the alphabetical arrangement of the first class code. Considering these additional constraints, the sign is practically optimal. On the contrary, Sugimoto US Patent No. The 14 physical key codes pn gt cr zk wj a e hi so ud xf ym vl qb (5,847,697) have an error rate (105, 4) and a valid key number 8.47. This code is practically not optimal and can not be typed in a strong touch, even though the number of physical keys is greater than ten.

While embodiments of the invention have been described with reference to the accompanying drawings, the invention is not limited to these embodiments, and various other variations and modifications can be made by those skilled in the art without departing from the scope of the invention. It must be understood.

Cited Documents

[1] http: /www.fujitsu.co.jp/hypertext/news/1996/jun/ms-txt.html

http: /www.fujisu.co.jp/hypertext/news/1998/May/27-e.html

http: /wwww.fujutsu.co.jp/hypertext/newx/1996/Jul/text.htm

[2] USP 5,818,437 Reduced keyboard disambiguating computer. US5818437, Oct 6, 1998

[3] Reduced keyboard disambiguationg system, PCT / US98 / 01307; WO98 / 33111

[4] "Body Coupled Fingering": Wireless Wearable Keyboard, CHI 97 Electronic Publications: Papers, by FUKUMOTO, Masaaki and TONOMURA, Yoshinobu NTT Human Interface Laboratories 1-1 Hikari-no-oka, Yokosuka-shi, Kanagawa-ken, 239 JAPAN web reference: http://www.acm.org/turing/sigs/sigchi/chi97/preceedings/paper/fkm.htm#U21.

delete

Claims (98)

delete delete delete delete delete delete delete delete delete delete delete delete delete delete delete delete delete delete delete delete delete delete delete delete delete delete delete delete delete delete delete delete delete delete delete delete delete delete delete delete delete delete delete delete delete delete Coded symbols; Decryption symbols; A keyboard composed of a plurality of keys responsive to user activation to generate sequences of encoded symbols; An output device for selectively outputting sequences of the decryption symbols in response to user activation of the keyboard; Map sequences of the coded symbols to sequences of the coded symbols, wherein at least one of the sequences of coded symbols is mapped to a sequence of a plurality of the coded symbols, a strong touch type capable of typing Ambiguous code; Formed by assigning the decryption symbols to the keys, the assignment being comprised of a plurality of decryption-symbol-assignment keys characterized by being as uneven as possible in the number of given decryption symbol assignment keys. Touch type typing device based on ambiguity. Coded symbols; Decryption symbols; A keyboard composed of a plurality of keys responsive to user activation to generate sequences of encoded symbols; An output device for selectively outputting sequences of the decryption symbols in response to user activation of the keyboard; A coder for mapping the sequences of encoded symbols to the sequences of decoded symbols, wherein at least one of the sequences of encoded symbols is mapped to sequences of a plurality of the decoded symbols, the ambiguity code being a representative language Given the statistics drawn from the corporation, i) the look-up error rate is at least Class C, ii) the query lookup error rate is at least Class B, and iii) the most effective strong touch scheme, where the number of valid keys is at least one selected from at least 10. A touch type typing device based on an ambiguous code, characterized in that a batter can be typed. Coded symbols; Decryption symbols; A keyboard composed of a plurality of keys responsive to user activation to generate sequences of encoded symbols; An output device for outputting sequences of the decryption symbols in response to user activation of the keyboard; A ambiguity code capable of mapping a sequence of the coded symbols to a sequence of the decoded symbols, wherein a strong touch-type batter code is possible, wherein at least one of the sequences of the coded symbols is mapped to sequences of a plurality of the decoded symbols. Consisting of; The ambiguity code is characterized by an optimum for at least one constraint selected from the group consisting of preservation of conventions, cross platform compatibility and reduced scan time; An ambiguity-based touch type typing device, characterized in that a plurality of decryption symbols are assigned to the keys, the assignment being as uneven as possible. Coded symbols; Decryption symbols; A keyboard composed of a plurality of keys responsive to user activation to generate sequences of encoded symbols; An output device for selectively outputting sequences of the decryption symbols in response to user activation of the keyboard; Mapping sequences of the encoded symbols to sequences of the decrypted symbols, wherein at least one of the sequences of encoded symbols consists of a ambiguity code mapped to a sequence of a plurality of the decryption symbols, the ambiguity being a lookup error rate and Optimal characteristics with respect to at least one of query query error rates; And assigning the decryption symbols to the keys, wherein the assignment is given a number of keys to which the decryption symbols have been assigned, and is as uneven as possible. Coded symbols; Decryption symbols; A keyboard composed of a plurality of keys responsive to user activation to generate sequences of encoded symbols; An output device for selectively outputting sequences of the decryption symbols in response to user activation of the keyboard; Mapping sequences of the coded symbols to sequences of the coded symbols, wherein at least one of the sequences of coded symbols is mapped to sequences of the coded symbols, and the sequences of coded symbols are at least one length 2 maximum An obfuscation code of up to 2 sequences of length 2, consisting of -2 sequences, wherein the sequences of deciphering symbols are the maximum-1 sequences of length 1, whereby the ambiguity code is the encoding symbol of the maximum-2 A mapping between and a subset of the decoded symbols, wherein an input of one of the maximum-2 sequences of coded symbols explicitly selects an ambiguous output of one of the subset of decoded symbols Touch type typing device based on code. 52. The apparatus of claim 51, wherein the apparatus is optimal for at least one constraint selected from a set consisting of a specific partition structure, preservation of conventions, cross-platform compatibility, and scan time reduction. 53. The method of claim 52 wherein preservation of the convention comprises a substantial Qwerty keyboard arrangement and the keys are arranged in at least three rows and one to ten columns, wherein the decryption symbols are a To z, the three columns comprising: a top column; Middle heat; And bottom column; The device assigns the letters to the keys, and as a result the keys include letter assigned keys, and the assignment of q, w, e, to the plurality of letter assigned keys in the top row of the three columns by the assignment; The letters are assigned such that r, t, y, u, i, o and p are assigned, and a, s, d, f, g, h, j with the plurality of letter assigned keys in the middle of the three columns. the letters are assigned such that, k and l are assigned, and the letters are assigned such that z, x, c, v, b and m are assigned to the plurality of letter assigned keys in the lower column of the three columns, The number of letter-assigned keys in rows of is monotonically reduced from the top row to the bottom row, and the number of keys assigned to each of the three columns is from the top row to the middle row. And reduce or maintain a constant from the middle row to the bottom row; The assignment satisfies the given character-assigned keys in each of the three columns as evenly as possible, where the keys have standard numbers with corresponding indicia for digits to optimize cross-platform compatibility. An ambiguity-based touch type typing device, characterized in that it is mounted in a form comprising a subpart of one of a keypad format and a standard telephone keypad format. 54. The apparatus of claim 53, wherein the ambiguous code is optimal for at least one of a lookup error rate and a query query error rate. 52. The apparatus of claim 51, wherein the apparatus is capable of strong touch typing, and the ambiguous code is an ambiguous code capable of strong touch typing. 56. The method of claim 55, wherein the strong touch type batterable ambiguous code is a multi-class ambiguous code, wherein the multi-class ambiguous code consists of a first-class ambiguous code and a second-class ambiguous code, and the coded symbols of the maximum-2 sequences. One of the first coded symbols is extracted from a first class ambiguity code, and in one of the maximum-2 sequences of coded symbols the second coded symbols are extracted from a second class ambiguity code. A touch type typing apparatus based on ambiguities. 52. The apparatus of claim 51, wherein the keys are spatially mounted and the plurality of decryption symbols are associated with the keys in a conventional arrangement in space. 52. The method of claim 51, wherein the plurality of maximum-2 sequences of coded symbols comprise activation of two keys that are nearly simultaneous, such that activation of the second keys occurs during activation of the first keys. And an input responsive to activation of the keys of the touch type typing device. 52. The method of claim 51, wherein the physical motion consists of any movement of body parts selected from, but not limited to, a collection of arms, hands, fingers, thumbs, legs, feet, toes, head and eyes. And an inputting of the sequences of the coded symbols. 52. The apparatus of claim 51, wherein the deciphering symbols are symbols used to represent a natural language consisting of, but not limited to, letters, numbers, and punctuation marks commonly used to write natural language. Touch type typing device based on the ambiguity. 53. The apparatus of claim 51, wherein the plurality of subsets of the decryption symbols comprises the single decryption symbols, whereby a maximum of the encoded symbols mapped to the single decryption symbols by the ambiguous code. The input of one of the two sequences serves to explicitly filter the single decryption symbol for an unambiguous output, the single decryption symbol being a component of a plurality of clearly filtered unambiguous decryption symbols. (ambient), touch type typing apparatus based on the ambiguity code. 62. The apparatus of claim 61, wherein the plurality of maximum-2 sequences of coded symbols consists of one element extracted from the plurality of combined coded symbols, wherein one component is a plurality of key-assigned codings. Extracted from symbols, wherein any single element of the plurality of key-assigned coded symbols is designed with a key-assigned coded symbol, wherein one of the key-assigned coded symbols is Activating one of the assigned keys is inputting the key-assigned encoding symbol, wherein the plurality of decryption symbols are the key-assigned decryption symbols and one of the key-assigned decryption symbols is input. Is generated, a component of a first subset of the key-assigned decryption symbols is output, and one of the combination coded symbols is the key-assigned decryption symbols. When input in combination with one, the component of the second subset of the key-assigned decryption symbols is output, such that any single element of the plurality of key-assigned decryption symbols is A ambiguity-based touch type typing device, characterized in that a key-assigned decryption symbol is designed. 63. The apparatus of claim 62, wherein at least one of the group consisting of the first subset and the second subset is comprised of the explicitly filtered unambiguous decoded symbols. 63. The apparatus of claim 62, wherein the first subset and the second subset consist of characters selected from a to z and their counterparts accented. 65. The method of claim 64, wherein the first subset consists of characters selected from a set consisting of c, e, h, l, n, s, t, x, y. Device. 65. The apparatus of claim 64, wherein the characters are assigned to the keys in alphabetical order. 67. An apparatus according to claim 66, wherein said characters are assigned said keys to a telephone keypad conforming to a standard character assignment. 63. The display device of claim 61, further comprising: a display device coupled to a processor via a suitable interface circuit, wherein the processor receives input sequences of encoded symbols from the keyboard and causes output sequences of the decrypted symbols to be output to the display device; A memory coupled to the processor, wherein the memory includes unambiguation software, the unambiguous software effectively mapping the sequences of the encoded symbols into the sequences of the explicitly filtered unambiguous decoded symbols; And an unambiguous rule database in which the sequences of encoded symbols are associated with the sequences of decrypted symbols. 69. The apparatus of claim 68, wherein the keys cause a response to user activation transmitted by one of physical sensations of sight, hearing, touch, taste and smell, and the display device is visual and auditory. And an ambiguity-based touch type typing device, characterized in that it is selected from the group consisting of tactile, gustatory and olfactory display devices. 69. The method of claim 68, wherein the unambiguous software is word-based, sequence-based, prefix-based, and any combination based. A touch type typing apparatus based on an ambiguity code, characterized in that selected from a set consisting of. 71. The computer readable medium of claim 70, wherein the unambiguous software performs an operation to apply filtering the plurality of unambiguous rules within the database of the unambiguous rule, wherein the plurality of filtering components of the unambiguous rules And / or wherein the sequences of decryption symbols in filtering the unambiguous rules of include at least one of the explicitly filtered unambiguous decryption symbols. 72. The method of claim 71, wherein the plurality of unambiguous rules filtering components are rules such that the explicitly filtered unambiguous decoded symbols may occur at the beginning, inside, or end of sequences of the unambiguous decodes. Touch type typing device based on the ambiguity. 62. The apparatus of claim 61, wherein the decryption symbols are explicitly filtered to optimize at least one lookup error rate and query query error rate. Coded symbols; Decryption symbols; A keyboard composed of a plurality of keys responsive to user activation to generate sequences of encoded symbols; An output device for selectively outputting sequences of the decryption symbols in response to user activation of the keyboard; Mapping sequences of the coded symbols to sequences of the coded symbols, wherein at least one of the sequences of coded symbols is mapped to sequences of the plurality of coded symbols, and a strong touch type typing is possible Ambiguous code; Assigning the decryption symbols to the keys; And wherein said ambiguous code with said assignment is optimal for at least one constraint selected from the group consisting of preservation of conventions, cross-platform compatibility, and scan time. 75. The apparatus of claim 74, further comprising: a thumb operable input device for efficiently inputting the plurality of unambiguous decoded symbols formed by the unambiguous association of the plurality of encoded symbols with the plurality of decoded symbols; A palm grip that allows the hand to slide along the expression by the pressure from the palm while the typing is in a comfortable position; And a display device for effectively displaying the plurality of decryption symbols upon input. 75. The apparatus of claim 74, wherein the ambiguity code is optimal for at least one of a lookup error rate and a query query error rate. 75. The apparatus of claim 74, wherein the plurality of keys are decryption-signal-assigned keys, and the number of such decryption-signal-assigned keys is approximately half of the number of decryption symbols, and at least of the number of decryption symbols. It is about half, The touch type typing apparatus based on the ambiguity code. 78. The apparatus of claim 77, wherein the plurality of decryption symbols consist of letters a to z. 75. The method of claim 74, wherein the preservation of the custom comprises a Qwerty keyboard arrangement and the keys are assigned within at least three rows and one to ten columns, and the decryption symbols are a through. consisting of a letter z, the three columns comprising: a top column; Middle heat; And bottom column; The device assigns the letters to the keys, such that the keys include letter assigned keys, wherein the plurality of letter assigned keys in the top column of the three columns include q, w, e, r, t, The letters are assigned such that y, u, i, o and p are assigned, and a, s, d, f, g, h, j, k and l with the plurality of letter assigned keys in the middle of the three columns. The letters are assigned such that the letters are assigned, and the letters are assigned such that z, x, c, v, b and m are assigned with the plurality of letter assigned keys in the lower column of the three columns, and in the respective columns The number of letter-assigned keys decreases monotonically from the top row to the bottom row, and the number of key assigned keys in each of the three rows is from the top row to the middle row and the middle row. To the bottom row or remain constant, the assignment being Satisfies the given character-assigned keys in each of the three columns as equally as possible, wherein the keys have a standard numeric keypad format with corresponding indicia for digits to be optimized for cross-platform compatibility. And a subpart of one of the standard telephone keypad formats. 80. The apparatus of claim 79, wherein the ambiguity code is optimal for at least one of a lookup error rate and a query query error rate. 75. The apparatus of claim 74, wherein the preservation of the custom is at least one of an alphabetic keyboard sequence and a quarttie keyboard sequence in assigning the decryption symbol to the keys. . 84. The apparatus of claim 81, wherein the preservation of the custom is made in the order of the qualitative keyboard layout and the keys are assigned in at least three columns and one to ten rows. 84. The apparatus of claim 82, wherein the decryption symbols consist of letters a to z, the three columns comprising: a top column; Middle heat; And bottom column; The keys are composed of the letter-assigned keys, such that q, w, e, r, t, y, u, i, o and p are assigned to the plurality of letter-assigned keys in a top column of the three columns. The letters are assigned, the letters are assigned such that a, s, d, f, g, h, j, k and l are assigned to the plurality of letter assigned keys in an intermediate column of the three columns, and An ambiguous sign based touch type typing device, characterized in that the letters are assigned such that z, x, c, v, b and m are assigned to the plurality of letter assigned keys in a lower row of the columns. 84. The apparatus of claim 83, wherein the number of letter-assigned keys in each of the columns monotonically decreases from the top row to the bottom row, and the number of keys assigned to each of the three columns. The touch type typing device according to claim 1, characterized in that from the upper row to the middle row and the middle row to the lower row to reduce or maintain a constant. 75. The apparatus of claim 74, wherein the keys are mounted in one of a standard numeric keypad format and a standard telephone keypad format, whereby preservation of conventions is optimal. . 75. The apparatus of claim 74, wherein the assigning of the decryption symbols to the keys comprises dividing the keys across according to a specified partitioning structure. 87. An apparatus as recited in claim 86, wherein said embodied split structure is as uniform as possible. 87. The method of claim 86, wherein the embodied splitting structure is an arc-based touch, characterized in that the keys assigned as the decryption symbols are assigned the decryption symbols as the keys are as uniform as possible given the number. Type typing device. Coded symbols; Decryption symbols; A keyboard composed of a plurality of keys responsive to user activation to generate sequences of encoded symbols; An output device for selectively outputting sequences of the decryption symbols in response to user activation of the keyboard; Maps sequences of the coded symbols to sequences of the coded symbols, wherein at least one of the sequences of coded symbols is mapped to sequences of the coded symbols, a ambiguity characterized by strong touch typing sign; Assigning the decryption symbols to the keys; And an ambiguous code based touch type typing device, characterized in that the conventional keyboard layout is preserved. 90. The apparatus of claim 89, wherein the plurality of keys are decryption-signal-assigned keys, and the number of such decryption-signal-assigned keys is an integer between one greater than one and one less than half the number of decryption symbols. And at least half of the number of decryption symbols. 93. The apparatus of claim 90, wherein the plurality of decryption symbols are comprised of letters a through z. 90. The apparatus of claim 89, wherein the conventional keyboard layout is a Qwerty array and the decryption symbols consist of letters a to z. A keyboard comprising keys; The plurality of keys are arranged in a top, middle, and bottom row such that each of the top, middle, and bottom rows comprises at least one of the keys; Characters a through z are assigned to at least one of the top, middle, and bottom columns of the keys such that the top column has the number of top columns of the letter-assigned key, and the break column is the break of the letter-assigned key. Has a number of columns, and the lower row has a lower number of columns of letter-assigned keys, the assignment of the letters a through z has an upper column letter-assigned number of 10 or less, and an interrupted column letter-assigned number of 9 or less Is characterized by satisfying at least one constraint in a set of constraints consisting of seven or less lower column character-assigned numbers; A first sub-assignment of the key assignment of the letters is such that q, w, e, r, t, y, u, i, o, p letters are assigned to the at least one key of the top row; A second sub-assignment of the key assignment of the letters is such that a, s, d, f, g, h, j, k, l letters are assigned to the at least one key in the break sequence; A third sub-allocation of the key assignment of the letters is that z, x, c, v, b, n, m letters are assigned to the at least one key of the lower row; At least one of the keys is assigned to one or more of the letters, And further comprising unambiguous software that effectively selects any of the characters to be input by the user of the keyboard when the user presses at least one of the keys to which one or more of the characters assigned by the character assignment are assigned. A qualitative-like device for inputting text. 94. The apparatus of claim 93, wherein the first partial assignment is given the number of possible character-assignment keys in the top row. 94. The apparatus of claim 93, wherein the second partial assignment is given the number of possible character-assignment keys of the break sequence. 94. The apparatus of claim 93, wherein the third partial assignment is given the number of possible character-assignment keys in the lower row. 94. The method of claim 93, wherein the assignment is a qualitative-like order, wherein the qualitative-like order is that the keys may be marked such that each of the letter assigned keys is a mark corresponding to a letter assigned to each of the letter assigned keys. Labeled, q, w, e, r, t, y, u, i, o, and p from left to right along the top column, and from left to right along the break column; and s, d, f, g, h, j, k, l and z, x, c, v, b, n, m from left to right along the bottom row. Quartie-like device for text entry. 95. The method of claim 93, wherein the number of letter-assigned keys in the top row is greater than or equal to the number of break column text-assigned keys, and the number of letter-assigned keys in the break columns is the bottom column text-assigned key. A quaternary-like device for text input, characterized in that it is greater than or equal to the number of.

Images