JPS6361702B2 - - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to a character processing device that inputs kana data and creates a sentence containing kanji and kana.

[Prior Art] Conventionally, when this type of device converts input kana data into kanji data, it prepares a dictionary for each field in which readings and homophones are stored in correspondence before starting the task of creating sentences. There was a device that specified a field and selected the dictionary to be accessed when converting kana to kanji according to the specified field.

[Purpose] The present applicant has previously proposed a technique in which an operator specifies the field of an input sentence, but the present invention does not require the operator to specify the field of the input sentence, and the device automatically selects the field of the input sentence. The purpose is to reduce the burden on the operator by making this determination.

An embodiment of the present invention will be described below with reference to the drawings. Before starting the explanation, terms used in this embodiment will be briefly explained.

The first character or first character string refers to a character or character string that is initially input to the device.
The second character or second character string refers to a character or character string resulting from conversion of the first character or first character string by the present device. Further, the second character string is not uniquely determined and may partially include a plurality of converted characters.

The conversion unit refers to a range of characters or character strings specifically designated for conversion into a second character string within the first character string.

The conversion word refers to a character or a character string that corresponds to a conversion unit included in the first character or the first character string, among the second characters or the second character string.

A word is the smallest unit that constitutes a converted word. That is, the conversion word consists of a single word or a plurality of words.

The above terms will be explained using specific examples. Now, if the input character string "Kihon Tekina" is converted to the output character string "Basic", the above terms will be as follows.

First character string “Kihontekina” Second character string “Basic” Conversion unit “Kihontek” Conversion word “Basic” word “Basic” and “Target” Now, here is Figure 1 1 shows a character conversion device which is an embodiment of the present invention. This character conversion device 11 includes a keyboard 12 for inputting the first character string and various commands, Display the second string corresponding to the string
A screen 13 such as a CRT, a recording device (not shown) housed in a character conversion device housing to record the second character string displayed on the screen 13 onto a recording paper 14, and a a second string of characters housed in a separate housing;
It consists of a processing unit that performs conversion processing into a character string.

Before starting a detailed explanation of this character converter,
By explaining the functions of the keys on the keyboard 12, the general functions of the character conversion device will be explained.

In FIG. 1, 15 is a key group KB1 for inputting the first character string, and by selecting input characters using the selection keys described below,
This is to input information corresponding to one of the symbols written on each keytop. That is, 16 is a selection key for selecting whether to input Japanese characters or English characters, and key 1 is for selecting whether to input in Roman characters or kana.
7 to determine the input character as follows.

That is, if the key 16 is set to the Japanese character input mode and the key 17 is set to the Roman character input mode, the keys with alphanumeric characters in the key group 15 can be operated to input Roman characters, and the input Roman characters are converted into kana. It is converted and processed in the processing section, and when displayed, it is displayed in kana or kana mixed with kanji. Also, in this state, key 1
When 7 is set to kana mode, kana characters can be input by operating the keys with kana characters in the key group 15.

When the key 16 is set to the alphabet input mode,
By operating the keys with alphanumeric characters in the key group 15, alphabetic characters are processed in the processing unit as they are (not as Roman characters) regardless of the mode of the key 17, and are displayed as they are when displayed. .

The key indicated by 18 is used when the key 16 is in the Japanese mode, and when the key 17 is in the kana mode, pressing the key 18 displays lowercase letters in hiragana or the right end of the key tops of keys 15-44 and 15-45. The displayed symbols are input, and when the key 17 is in the Roman alphabet mode, the input Roman characters are converted into lowercase hiragana letters and processed by the processing unit.

The key indicated by 19 is a key for selecting whether to input uppercase hiragana letters or lowercase English letters. When the key 16 is in Japanese mode and the key 17 is in kana mode, the key group 15 selects whether to enter uppercase letters in hiragana or lowercase English letters. When the key 17 is in the romaji mode, input the romaji by operating the key with an alphabetic character in the key group 15. It is converted into and input to the processing section.

Further, by setting the key 19 to the lowercase alphabetic character mode while the key 16 is in the alphabetic character input mode, lowercase English letters can be input by operating the keys labeled with alphabetic characters in the key group 15. be.

Note that these keys 16, 18, and 19 are shift keys.
This constitutes SHK1.

The key indicated by 20 is a space key, and pressing this key 20 creates a space between input characters.

Reference numeral 21 indicates keys for inputting katakana lowercase letters and symbols written on the left end of key tops (for example, 15-44, 15-45); key 16 is used for Japanese character input mode, and key 17 is used for kana input mode. If so, by operating the key with a kana in the key group 15, the kana is input as a lowercase katakana, and if the key 17 is set to the Roman character input mode, the input Roman character is converted to the corresponding lowercase character of the kana. and input it.

Use the key indicated by 22 to input katakana capital letters, or
This key is used to select whether to input uppercase English letters. By setting key 16 to Japanese mode and key 17 to Kana mode, it is possible to input Katakana capital letters by operating keys with kana in key group 15. By setting the key 17 to the Roman character mode, the Roman characters input by operating the keys with alphabetic characters in the key group 15 can be converted into katakana capital letters and input. Furthermore, keys 21 and 22 are used as shift keys.
This is what makes up SHK2.

Reference numeral 23 indicates a key that indicates the start or end of inputting characters to be converted to kanji, and when it is detected that the key 23 is pressed an odd number of times, inputting of characters to be converted to kanji begins. It generates a start signal to notify that the key 23 has been pressed for an even number of times, and generates an end signal to notify that the input of characters to be converted into kanji has been completed. be.

The key indicated by 24 is a cursor key, and is used to move the cursor appearing on the screen 13 one character at a time to a desired location.

Reference numeral 25 indicates an initial set key, which is used to set the processing section to an initial state by pressing it before using the character conversion device 11. Reference numeral 26 indicates an edit key, which instructs to search for a location in the second character string displayed on the screen 13 where a plurality of converted words exist.

Reference numeral 27 indicates a print command key, which prints the second character string displayed on the screen 13 onto the recording paper 1.
This key instructs to record on 4.

28 is a key that indicates that when the first character string is converted to the second character string and there are multiple converted words, the one converted word displayed on the screen is the correct converted word. , the key 29 indicates that it is an incorrect conversion word, and this key 29
By pressing , other converted words will be displayed on the screen. Therefore, by pressing the key 29 until the correct converted word is displayed on the screen 13 and pressing the key 28 when the correct converted word is displayed, the correct converted word is selected.

The key 30 has the same function as the key 29, but the display order of converted words is reversed.

The key 31 is a key for instructing to convert the displayed second character string (conversion word) into a first character string (conversion unit).

32 indicates an automatic conversion mode in which the first character string inputted from the key group 15 is converted into a second character string in the processing unit and displayed on the screen 13, or a character corresponding to a single kanji is inputted from the key group 15. This key is used to select a single kanji input mode in which all kanji corresponding to the inputted character are displayed on the screen 13, and a desired kanji is selected from the displayed single kanji.

33 shows the direct input mode in which the kanji code is directly input using the numerical keys 34, and the key group 15.
This key is used to select an indirect input mode in which a second character string is output from the processing section by inputting a first character string.

An explanation will be given of an operation for inputting data according to the document shown in FIG. 2 using keys having the functions described above, and obtaining an output as shown in FIG. 5 on the screen 13.

When inputting in hiragana, first, after initial reset with key 25, key 16 is set to Japanese mode, key 17 is set to kana mode, and key 19 is set to hiragana uppercase mode. However, since the character 37-1 indicated by a circle in FIG. 3 is a lowercase hiragana character, the key 18 is pressed to input this character.

Then, as shown in Fig. 3, press the key 23 to signal the start of inputting characters to be converted into kanji, operate the key group 15 to input the kana ``konkai'' to be converted into kanji, and then press the key 23. Press 23 to notify that the input of characters to be converted into kanji is complete, then input the kana "no", and so on, and then input the kana in sequence as shown in FIG. 3.

As will be described later, in the character conversion device according to this embodiment, by inputting symbols indicating sentence breaks such as dots "," and circles "." from the keyboard 12,
Since the conversion from the first character string to the second character string is started, when the point 36 is input, the second character string is displayed on the screen 13 as shown in FIG. Then, input continues as shown in Figure 3, and when the last circle 37 is input, the fifth
All the second character strings are displayed as shown.

In addition, the symbol indicated by 35 in Fig. 3 is the key 2.
This symbol indicates that 3 has been pressed.

In FIG. 5, the converted words marked with a symbol 38 (*) indicate that there are homophones other than the displayed word. In other words, conversion words to which the symbol 38 is not attached are conversion words that are uniquely determined for input characters.

Once all input has been completed in this way,
Next, when the key 26 is pressed, the cursor 39 is placed at the position of the first symbol 38-1.

The conversion word “basic” indicated by this cursor 39
is a correct conversion word, so when key 28 is pressed, cursor 39 moves to below the next symbol 38-2.

Since the conversion word "after" indicated by the cursor 39 is also a correct conversion word, when the key 28 is pressed, the cursor 39 moves to below the next symbol 38-3.

The conversion word "return" indicated by the cursor 39 is incorrect, so press the key 29 to change the next homophone "return".
is displayed on the screen 13 instead of the above-mentioned "Return".

Since this converted word "Kaku" is a correct converted word, when the key 28 is pressed, this converted word "Kaku" is selected.

Once all editing is completed in this way,
By pressing the key 27, the second character string displayed on the screen 13 is transferred directly to the recording paper 14.
This is recorded above.

Next, to explain the device in more detail by taking as an example the case of inputting in the Roman alphabet mode, the order shown in FIG. Enter. However, when entering the character 40 indicated by a circle, the key 18 is pressed.

The contents of the manuscript input as described above are processed within the apparatus and displayed, for example, as a mixture of kanji and kana, as shown in FIG.

In the above process, if there are multiple sets of kanji with the same sound, they are displayed with * as shown in FIG.

(Example: *On the way home) Here, * is used to inform the operator that there are two or more homophone kanji (converted words).

If multiple homophone kanji (conversion words) exist for the reading of the input manuscript, the necessary kanji (conversion words) are selected by the following means.

For example, move the cursor symbol CC to the position of "*" in the "*Basic" display in FIG. Such movement can be performed automatically by operating the EDIT key. The kanji that you want to select now is the kanji that is displayed, so the YES key is operated. Then, the cursor symbol CC automatically moves to the position of "*""after*" to be selected next.
Since the kanji you want to select now is "transition", the NO key is operated to call up the kanji that is not displayed. Then, "later" disappears from the display screen and "majesty" appears in the same place. Further NO
When the key is operated, the above-mentioned display procedure is performed as shown in Table 1, and when the currently required "transition" is displayed, when the YES key is operated, "transition" is selected and the device is activated. Move the cursor symbol CC to the next “*Return”
automatically shifts to the "*" position, and informs the operator of the position of the next selection target.

Table 1 Operation Display EDIT key↓ * Basic YES key↓ 〓Basic...afterwards NO key↓ * Prestige NO key↓ *Intention NO key↓ * Transition YES key ↓ 〓Transition... Return NO key↓ * Plan YES Key ↓ 〓 Plan... Year YES key ↓ 〓 Year... By repeating the same key operations as above, the selection can be completed.

Here, functions for improving the operability of the selection will be explained using Table 2.

When "*Transfer" is displayed as a result of operating the NO key, "*Transfer" is displayed when the NO key is operated further. (Table 2, line 6) All homophone kanji (converted words) corresponding to "Ikou" are now displayed. If the NO key is pressed again here, the first displayed "* and later" will be output again. By constantly displaying it repeatedly like this,
Ease of selection is improved. Next, when you press the BACK key, the display is completely opposite to when you press the NO key. In other words, "*Transfer" is displayed again, and by pressing the BACK key again, "*Transfer" is displayed.
will be displayed. here same as last time
Migration can be selected by operating the YES key. The following is the same as last time.

Table 2 Operation Display EDIT key↓ * Basic YES key↓ 〓Basic... * After that NO key↓ * Prestige NO key↓ *Intent NO key↓ * Transition NO key↓ * Transfer NO key↓ * After BACK key↓ * Transfer BACK key ↓ * Transfer YES key ↓ 〓Transfer... * Return Table 3 Operation Display EDIT key ↓ * Basic idea... INV key ↓ Basic idea... Next, you cannot make corrections using the above methods. For things, it is necessary to be able to easily return them to their original first character (the corresponding conversion unit). Use the INV key for this purpose. First, move the cursor using the method described above (using the EDIT key or YES key) or the CURSOR SHIFT key (CURSOR
The SHIFT key can move the cursor by one character with one operation).
Move the cursor symbol CC to the position of "*" corresponding to the kanji you want to convert to the first character. here
By operating the INV key, the second character is converted to the first character and displayed.

This situation is shown in Table 3.

The method for automatically converting a first character string into a second character string in the process of converting it has been described above.

Next, a method for inputting kanji characters one by one will be described.

There are two methods for inputting kanji units. One is an indirect input method and the other is a direct input method.

First, the automobile kanji selection key AMK is set to the single kanji side, and then the indirect and direct selection key IDK is set to the direction desired by the operator. Thereafter, the cursor signal CC is shifted by the cursor shift key CSK and brought to the corresponding position where the input is desired.

If the indirect direct selection key IDK is set to the indirect side, input kanji using the following procedure.

First, press the kanji shift key KK23 in Figure 1,
Next, enter the reading of the kanji. For the reading of the kanji, enter the reading of the kanji in Roman characters if the Roman character/kana selection key RKK is in Roman characters, and if RKK is in kana, enter the reading in kana. The reading of a kanji can be input as a sound or a kun, or only a part of the reading of the kanji from the beginning can be input. Thereafter, when the kanji shift key KK is pressed again, the corresponding kanji are listed on the screen 13 as shown in FIG. 30 T1. The listed kanji groups correspond to the second row of keys from the bottom in Figure 1B, and are the English character keys A, S,
They correspond to D, F, G, H, J, K, L, respectively in that order.

Next, if a desired kanji is displayed, the user presses the key corresponding to the position of the kanji to input it. If the desired kanji is not displayed, ":"
By pressing the key or space key, a group of kanji following the kanji group on the current screen is displayed, and the desired kanji can be input in the same manner as described above. For example, the 32nd
In the diagram, if you want to change "hon" at position T2 to "ki", move the cursor to position T2 using the cursor shift key CSK and press the single kanji selection key.The entire screen will disappear and the cursor will move to the bottom left corner. Move. After that, as mentioned above, press the Kanji shift key KK,
By pressing the kanji shift key KK for "la", kanji starting with "la" are listed as shown in T1 in FIG. 30.
If you press the key "S" corresponding to "come" here, T2
The character ``Kai'' is inserted at the position of , the data is corrected and displayed on the screen, and the cursor advances to the next character position. A single kanji input control section, which will be described later, performs the above-mentioned control.

In this example, the operator's standard finger usage, for example, the middle finger of the right hand is 9, I, K, Yo, etc.
It is used to input the letters ``d'', ``no'', and ``ne'', but it is structured so that pressing any of the keys will specify the third letter from the left in Figure 30, which corresponds to the middle finger of the right hand. .

When the indirect direct selection key is set to the direct side, the kanji code can be input into the data buffer corresponding to the current cursor position using the kanji direct input means DIN, and the kanji code can be input at the current cursor position. can be displayed. After input, the cursor automatically advances to the next position. This direct input means is capable of inputting not only kanji but also kana alphanumeric characters and other symbols. Since individual kanji characters can be input as described above, words that are not in the word dictionary file can also be easily input, and kanji characters that have been erroneously converted during conversion can also be easily corrected.

In the manner described above, sentences can be created easily and quickly.

Next, when the print key is operated, the displayed text is printed out. At this time, the * symbol indicating the existence of multiple homophones (converted words) and extra spaces (for example, the space after "Year" in FIG. 5) are deleted and recorded in a condensed manner.

Although the most basic operation of the character conversion device according to this embodiment is as described above, this character conversion device will be explained in more detail using the drawings.

FIG. 7 is a block diagram showing the character conversion device according to this embodiment.
Reference numeral 3 indicates keys having the above-mentioned functions, each of which is provided with an encoder for generating a code signal corresponding to the key.

Other functions of each of the keys mentioned above will be explained in more detail below.

The CPU is a processing unit that performs, for example, 16-bit processing.
The processing unit CPU has functions such as data transfer, addition, and comparison, and is composed of, for example, Texas Instruments' TMS9900, National Semiconductor's PACE, Panahuacom's PFL16A, NEC μPD16, etc.

ROM is control memory, keyboard KB1, KB
2. This is a read-only memory that stores control procedures, control data, etc. for determining input instructions from the KB3 and performing processing in accordance with each instruction. This ROM
is divided into blocks as shown in FIG.
A description of each block will be given later. Each control unit performs control regarding its name. for example,
The KB1 input control section performs KB1 input control. The main control section also supervises other control sections.

FIFO is a memory that temporarily stores the output from KB2, is a first-in-first-out memory, and has a capacity of 64W (1W16bit). Information that exceeds the processing capacity of the CPU is stored.

PM1 is a page memory, and as shown in FIG. 5, one document page has a character capacity of, for example, 16 (characters) x 12 (lines), ie, 192 WORD. The contents of the page memory PM are displayed on a CRT (described later) and output to an output device such as a printer, magnetic disk, etc.

In addition, the 14th bit of each WORD of page memory PM,
The following information is stored in the 15th bit. (Please refer to Figure 8) When bit15-0 and bit14-0, the character code is entered in bit13-bit0.

bit15−0, bit14−1 The address of the homophone memory comes from bit9 to bit0.

bit15-1, bit14-1 means love out code. When this code exists, a space is displayed on the display surface of the display means. Also, it is ignored when printing out.

PM2 is a page memory and has the same configuration as PM1. The contents of this memory are the same as PM1.
can be displayed.

WF is a word file, and as shown in FIG. 11, the reading of a word (kanji) is a keyword, and the word character code, grammatical information, etc. for the reading of the word are entered.

In this embodiment, a storage capacity of 32W is prepared for one word, including 10WORD for the reading of the word (kana heading), the reading of the unchangeable kana part, that is, the part that does not change depending on the conjugation of the kanji okurikana. For example, 3WORD for "shi" in "beautiful", 5WORD for word character code, and 5WORD for part-of-speech information.
1WORD, 1WORD for conjugation information, 1WORD for affix information, 4WORD for field information, 1WORD for information on the number of kanji, that is, the number of character codes entered in the character code field, 1WORD for the invariant kana part. For information on the number of reading codes entered
1WORD, 1WORD as frequency information, as vacant
Consists of 4WORD allocation.

The part of speech information corresponds to each bit of 1WORD as shown in FIG. 10. This includes information such as prefixes and suffixes.

As shown in Fig. 11, the conjugation information corresponds to each bit of 1WORD with a conjugation such as a verb or an adjective.

The affix information corresponds to each bit of 1WORD, as shown in FIG. 12. Affix information includes place name, person name, organization name, ground (connected to place name, e.g., prefecture, city), affix (connected to person's name, e.g., Mr., Kimi), and affix (connected to organization name, e.g., division, etc.). It is categorized as ``Association''.

Frequency information was determined with reference to materials from the National Institute for Japanese Language and Linguistics.

The field information corresponds to each bit of 1WORD, as shown in FIG. 13. The fields in which each word is frequently used are shown.

Note that the words stored in the word file WF can be the converted words.

As shown in Figure 14, the single kanji dictionary KF uses the readings of kanji as keywords (5W), making it possible to search for kanji codes. You can search how to read kanji by both sound and kun.
If there is a different reading for each kanji, a KF is created for that reading.

Note that in this example, we used a structure of one kanji for one kana heading, but if there are multiple kanji with the same sound, it is possible to combine them into a table that stores multiple kanji with the same sound for one heading. teeth,
it is obvious.

PNT1 is a pointer memory and has a storage capacity of 8 bits. This memory serves as an address register for specifying the position of the cursor symbol CC displayed on the display surface of the display means and as an address register for reading and writing the page memory PM1.

PNT2 is an address register for reading and writing the page memory PM2.

Both PNT1 and PNT2 have an address value of 1 to
Numbers up to 192 have meaning. It is invalid when set to any other value.

DCOT is the display control circuit (Fig. 16).
As shown in Figure 5, code converter CCT, selector (SLT), character generator CG, CRT
Consists of controller CCL.

The code converter converts the code of the information output from PM1 according to the following rules.

That is, When bit15=0, bit14=0, Output as is.

When bit15=0, bit14=1, Convert to the character code corresponding to “*”.

When bit15=1, bit14=1, Convert to space code.

When bit15=1, bit14=1, Convert to space code.

The selector selects either the output of PM1 or the output of PM2 via the code converter using the lock key.
Responsible for selection in relation to AMK and IDK.

The CRT controller has the function of displaying the output information from the selector on a display and displays a cursor symbol at the position specified by the pointer PNT1.

Such a CRT controller refers to a character generator CG, generates a character pattern by inputting a character code to the CG, and displays the character code in the page memory PM.

The display control circuit DCOT controls the display to display 12 lines vertically and 16 characters horizontally.

Therefore, if page memory PM1 or PM2 does not contain a carriage return code or line feed code, there is a counter in the display control circuit that counts the number of characters to display 16 characters per line, and 12 lines. counter lines,
A counter is also provided to count the lines so that there are 12 lines.

In addition, the processing unit CPU uses page memory PM and pointer
The time to control PNT is display control control
DCOT is performed during the transition from one line display to the next line display.

A CRT is a display unit that consists of a cathode, a ray, and a tube. A display control circuit is installed on such a display CRT.
Display information under the control of DCOT.

The direct input means DIN is as shown in Figure 16.
Consists of TEN KEY34 (KB4) and encoder ENCD. The encoder is responsible for converting the 4-digit numbers output from KB4 into character codes. As a result, a character defined by a four-digit decimal number is converted into a character code corresponding to that character.

In addition, the direct input means DIN is a kanji input device, for example,
Oki Electric Industry Co., Ltd. KANJI TAB ENTRY−
500, Fujitsu Ltd. FACOM6802B, etc.

OU is an output device that prints the characters displayed on the display unit CRT. For example, it is composed of wire dots, ink jets, thermal printers, etc.

Furthermore, a magnetic disk or the like may be connected as an output device to increase the storage capacity.

RAM is a memory that is provided to control the transfer of data processed by the processing unit CPU. Its configuration, as shown in FIG. 17, consists of various REGISTERs, working areas, conversion parameters 1, conversion sentence buffers, conversion word tables, conversion parameters 2, conversion unit tables, nest tables, word dictionary tables, etc.

Further, the above REGISTER includes a keyword register, a selection number register, an input sentence field register, a conversion word field register, etc.

Next, the details and mutual relationships of each parameter, buffer, table, and each control section will be explained. For individual explanations, the respective drawings will be used, and for explanations of mutual relationships, FIG. 18 will be used.

The converted sentence BUF is shown in Figure 19. The conversion sentence BUF is a buffer for storing sentences to be converted, and is controlled by the KB1 input control section described later.
The reading information of sentences inputted from KB1 is converted and stored according to a predetermined format. The format of the conversion sentence BUF is as shown in FIG. All character information is JIS in kana and symbols.
Expressed by code. Also, a special code is used to indicate the part to be converted into kanji (hereinafter referred to as a conversion unit). The information stored under the control of the KB1 input control section includes the text information and special code. specifies the beginning of the conversion unit, and specifies the end of the conversion unit.

Conversion parameters MPA1 and MPA2 are parameters that specify the range of information stored in the conversion sentence buffer to be converted into the second character string, and KB1
It is set by the input control section and becomes a parameter for subsequent conversion (see FIG. 20).

MPA3 specifies the first address within the range excluding the range where the conversion has been completed. KB
Initialized by the 1-input control section,
Set to MPA3-MPA1.

MPA4 is a register that stores an empty number in a conversion word table, which will be described later. The specific role of MPA4 will be discussed later. MPA4 is initialized by the INT key controller.

MPA5 and MPA6 are the ranges to be converted, i.e.
These indicate the first address and last address to be referenced when converting between MPA1 and MPA2.

In the KB1 input control section, the AMK key is automatic,
In addition, when there is an input from KB1, the operation starts according to a command from the main control section. After the operation is completed, control is returned to the main control section. The KB1 input control unit converts the information input from KB1 into a converted sentence.
Convert to the format specified by BUF,
After that, it writes the information related to the conversion sentence BUF, and at the same time writes parameters necessary for conversion into conversion parameter 1, and plays the role of giving an instruction to the conversion control unit to start conversion. Input from KB1 is RKK
By operating the keys, you can select between Kana input and Romaji input. Furthermore, the odd-numbered kanji shift key KK of KB1 indicates the beginning of the conversion unit, and the even-numbered depression indicates the rear end of the conversion unit. The KB1 input control section converts the above input into a format specified by the above-described converted text buffer, and writes the converted information to the converted text buffer.

The conversion of the information stored in the conversion sentence buffer into a kanji-mixed sentence (second string) starts when one of the following conditions is satisfied:
When any of the following conditions is satisfied, an instruction to start the conversion is given to the conversion control unit.

1 Symbols such as “,,” are included in the input information from KB1.
etc. existed.

2 When a kanji start code is generated by the kanji shift key.

3 When a certain number of characters are entered after the Kanji shift key is pressed twice.

4 When the buffer overflows.

The KB1 input control section sets the conversion start position and end position of the information stored in the conversion sentence BUF in conversion parameter 1 before giving an instruction to start conversion to the conversion control section. Conversion to the second character example is performed within this range. At the same time, a reference range as related information for conversion is also set to conversion parameter 1.

In kana-kanji conversion, the information before and after a word is important, so when converting a word in the middle of a sentence to kanji, the kana-to-kanji conversion is controlled so that the information before and after is always added.

The Romaji-kana conversion is activated by the KB1 input control unit when inputting vowels, ie [A, I, U, E,
Keyboard for [O] or [X] (see Figure 1)
The corresponding kana above, namely [chi, ni, na, i, ra]
Or, it is performed when either [sa], [chi, ni, na, i, la] or [sa] and their respective lowercase letters are input. Conversion is performed with reference to the Romaji-kana correspondence table (FIG. 21). For example, when inputting "A", enter Roman characters on the keyboard shown in Figure 1, and when the "A" key, that is, the "Chi" key is pressed in the hiragana shift state, the Romaji-kana conversion control unit is activated and the 21st "A" is extracted from the table shown in the figure.

As shown in FIG. 22, the word dictionary table is a buffer for extracting and storing only information necessary for conversion from among the information stored in the word dictionary WF, and has a capacity of 32 x 24 W. That is, as mentioned above, since it has a capacity of 32 W per word, information for 24 words can be stored, and the stored information is sequentially numbered from 1 to 24 for each word. Writing to this word dictionary table is performed by conversion control, which will be described later.

The nest table is a table written by the conversion control unit, which will be described later, when finding a conversion word from each conversion unit, and as shown in FIG.
It has a capacity of 8W.

This embodiment has a function of automatically dividing a conversion unit. For example, the conversion unit 〓Kihon-teki〓 is automatically divided into 〓Kihon〓 and 〓Text〓. The number of divisions is not limited to two as in the above example.

In this embodiment, up to eight divisions are allowed. The words obtained by division are either independent words, prefixes, or suffixes. This nest table stores information about this division. For each segmented word, there are 28W of information storage locations.

In FIG. 23A, (A) is an address in the converted sentence buffer where the pronunciation of the word is stored. (B)
is the number of readings of the word. (C) is the number of words when there are multiple words corresponding to the reading of the word. (D) is the leading number of the word dictionary table that stores the word dictionary information corresponding to the word (one of the numbers 1 to 24 of the word dictionary table mentioned above). (E) is a number in the word dictionary table in which word dictionary information corresponding to the word is stored.

Assume that a conversion unit called 〓Kihonteki〓 is stored at the head of a conversion sentence buffer, and a nest table corresponding to the conversion unit has been written by a conversion control unit, which will be described later. At that time, it is divided into "kihon" and "teki", and the character string (word) corresponding to the former is "basic", and the character string (word) corresponding to the latter is "enemy", "drop", "suit", etc. Assume that four types of "targets" are found by the conversion control unit, which will be described later.
An example of the nest table at this time is shown in FIG. 23B.

The conversion parameter 2 is written by a conversion control unit, which will be described later, when converting a conversion unit into a conversion word, and includes various parameters.
As shown in Figure 24, it has a capacity of 9W, and MCW1
They are named from ~MCW9.

MCW1 is a register in which a conversion control unit, which will be described later, writes the start address of a conversion unit found within a specified range of the conversion sentence buffer (range specified by conversion parameter 1).

MCW2 is the length of the conversion unit.

MCW3 is the start address of a word in the conversion unit that is actually searched in the word dictionary WF by a conversion control unit, which will be described later.

MCW4 is the length of the word for which the search is performed.

MCW5 is a register in which the number of words obtained as a result of the search is entered. However, when it is negative, it indicates that the corresponding item did not exist as a result of the search.

MCW6 is an empty space in the word dictionary table that stores word dictionary information for each word obtained as a result of the search.
No. 24 in the word dictionary table is an empty table.

MCW7 is the number of divisions in the conversion unit division.

MCW8 is a register in which the number of converted words corresponding to the conversion unit is written. A positive value indicates the number of converted second character strings, and a negative value indicates that the conversion could not be performed.

MCW9 is the number of words whose "certainty" (described later) exceeds a certain threshold value among the eight MCW conversion words.

The conversion control section starts operating according to an instruction from the KB1 input control section, and returns the control to the KB1 input control section after completion of the operation.

This control unit converts the range defined by conversion parameter 1 of the information (first character string) stored in the conversion sentence buffer into a second character string, and stores the result in the nest table and word dictionary. Write to table, conversion parameter 2. After that, it takes on the role of activating the conversion unit table creation control section and the conversion result processing control section. Here, the concept of the conversion control section will be described. The conversion control unit first searches for a conversion unit within the range from MPA3 to MPA2. Once the conversion unit is found, set MCW1 and MCW2. At the same time, MCW6 is set to 1 to clear the word dictionary table. Next, in order to search the word dictionary WF and find the conversion word corresponding to the conversion unit,
Set the keywords of the word dictionary WF in MCW3 and MCW4. Naturally, first of all,
Set MCW(3) = MCW(1), MCW(4) = MCW(2). Also, give 1 to MCW7 as an initial value.
(The MCW 7 will be described later.) The search is performed by a search control unit located within the conversion control unit. Search control uses dictionary information corresponding to keywords set in MCW3 and MCW4 as a word dictionary.
Search from WF and set the results in the word dictionary table and MCW5. If it is found as a result of the search, the number is set in MCW5, and the dictionary information is stored in the word dictionary table starting from the position indicated by MCW6. If no result is found, MCW5 is set to -1.

After the above operations of the search control section are completed, the conversion control section examines the search results. if,
If MCW5 is negative, 1 is subtracted from MCW4 and the word dictionary is searched again by the search control unit.
This operation is repeated until the corresponding word is found.

If the corresponding word is found, that is, if MCW5 is a positive number, the state at that time is written into the nest table. That is, MCW3 is shown in Figure 23 (A),
MCW4 to (B), MCW5 to (C), MCW6 to (D)
Save to. In addition, (E) indicates the corresponding number in the word dictionary table that stores the dictionary information of the corresponding word.
Fill in only the number of minutes set in MCW5.

If there is a residual part in the conversion unit in the previous search, 1 is added to MCW7, and the word dictionary search for the remaining part in the conversion unit is performed in the same manner as the previous time. At that time, MCW3 and MCW4 are changed to the address and length indicating the remaining portion. However, if the corresponding word is not found in the search for the remaining part, the word found in the search when it was written into the nest table last time is canceled. Therefore, at this time, subtract 1 from MCW7,
Re-write the information already written to the nest table
Return to MCW3~MCW6. And even more
Subtract 1 from MCW4 and execute search control again. If the word dictionary search for the remaining part is successful, save MCW3 to MCW6 in the nest table as before, enter the specified items in section (E) of the nest table, and if there is still a remaining part. repeats the above explanation.

The conversion unit is thus converted as one or more words. Through the above process, the number of divisions of the conversion unit is written in MCW7.

The conversion ends above, but if the conversion cannot be completed in the end, MCW(7)=1, MCW(5)=-
It is clear from the above explanation that MCW(4)=1.

After that, if the conversion is successful, the conversion control section activates the conversion unit table creation control section, and then activates the conversion result processing control section. If the conversion is unsuccessful, MCW8 is set to -1 and the conversion result processing control section is activated.

After the conversion result processing control unit finishes its control,
The conversion control unit again searches the conversion sentence buffer for the next conversion unit to be converted. If it exists, the above control is repeated. After converting the entire range specified by conversion parameter 1, the conversion control section returns control to the KB1 input control section. The KB1 input control section transfers its control to the main control section as described above.

The conversion unit table is a place for storing combinations of words (sometimes only one word) corresponding to conversion units, and is a table in which the certainty of conversion for each combination is entered. As shown in FIG. 25, the conversion unit table has a capacity of 11×24W. It has a capacity of 11W for each combination. Therefore, a total of 24 combinations can be stored. For example, as mentioned above, suppose that for the conversion unit 〓Kihon Teki〓, 1 and 4 words corresponding to ``Kihon'' and ``Teki'' are obtained, respectively (this information is recorded in the nest table). , four combinations of 1×4 can be made as combinations of both. 11W of information is stored for each of these four ways. The first 8Ws define word combinations, and their numbers are the numbers of word information stored in the word dictionary table. The next 1W is the number of words, i.e. the number of numbers written in the first 8W,
The next 1W is a field in which to enter the part of speech of the combination as a whole. For example, the compound word ``basic'', which is a combination of ``basic'' and ``target'', is treated as an adjective verb. The next 1W is a column to enter the "certainty" of the combination. Writing to the conversion unit table is first performed by the conversion unit table creation control unit, and then the “certainty” column is written by the grammar weight control unit, frequency weight control unit,
It is modified by the affix weight control section and the field weight control section.

The main function of the conversion unit table creation control section is to create the conversion unit table. For creation, the nest table and word dictionary table conversion parameter 2 are used. In MCW8,
Enter the number of types of word combinations entered in the conversion unit table, that is, the number of converted words.

Next, write down the part of speech and conjugation of the combination as a whole. The problem here is which one to select from among these multiple conversion words.

Therefore, in order to modify the "certainty" term of the conversion unit table, the grammar weight control section, frequency weight control section, connection weight control section, and field weight control section are activated.

With the above steps, the conversion unit table can be completed.

Next, set MCW9, but as mentioned above, MCW9 can be the number of types of word combinations entered in the conversion unit table, but here we will use another setting method and explain it. .

That is, it is defined in terms of the number of thresholds exceeding a certain threshold, taking into consideration the certainty modified by each of the weight control sections. Also, the threshold may always be constant, or
It may be changed depending on each conversion unit. The threshold is determined by a threshold determination control section included in the conversion unit creation control section. This makes it possible to eliminate the output of converted words that are clearly erroneous.

The grammatical weight control unit performs a grammatical check and weights each second character string converted from the conversion unit. Note that the grammar check is performed within the reference range indicated by MPA5 and MPA6. It is started by the conversion unit table creation control section, and after completion, the control is returned to the conversion unit table creation control section.

The grammar weight control section is roughly divided into two parts. One is that the converted sentence buffer is divided into a front grammar check section that checks the connection with the front part of the conversion unit, and a rear grammar check section that checks the connection with the rear part of the conversion unit.

The front grammar check part checks whether the front part of the conversion unit is a case particle when the second character string is a verb. If it is a case particle, the front grammar check part increases the weight, and in other cases, it increases the weight. Do something to reduce weight.

The rear grammar check section is further divided into two parts.
One is the unchangeable kana part check which checks whether the character following the conversion unit matches the pronunciation of the unchangeable kana part in the word dictionary table, and the other is This is a conjugation check that checks whether it matches the conjugation of a verb, etc.

The conjugations of verbs, adjectives, etc. have regularity in their changes, as is common knowledge, and once the type of conjugation of a verb is determined, the letters that follow it will inevitably be determined. Both the unchanged kana part check and the conjugated form check are
It plays an important role in determining the "certainty" of a string.

Weights are calculated for all second character strings listed in the conversion unit table.

The frequency weight control section has the role of checking the frequency and assigning weights to each of the converted words converted from the conversion unit. This control section is activated by the conversion unit table creation control section, and returns its control to the conversion unit table creation control section after completion. The frequency weight control section performs control such that frequency information of each word listed in the conversion unit table is obtained from the word dictionary table and added to the weight of each second character string.

Connection table 1, as shown in FIG. 26, is a table showing connection relationships between parts of speech, prefixes, or suffixes of independent words. The numbers in the table indicate the degree of connection between each element. However, although this number was determined empirically, it can be approached to a better value by arbitrarily changing it. In the figure, the x mark indicates an extremely low possibility of connection.

The connection table 2 has the 12th connection table as shown in FIG.
This is a table showing connection relationships based on the connection information shown in the figure. The view of the figure is the same as connection table 1.

The connection weight control section is responsible for adding a connection check and weighting each of the single or plural conversion words converted from the conversion unit. This control section is activated by the conversion unit table creation control section, and returns to the conversion unit table creation control section after the operation is completed.

This control section is roughly divided into two parts. One is a part-of-speech connection check section using affix table 1, and the other is a semantic connection check section using connection table 2. In both cases, when the second character string consists of two words, the connection between them is investigated.

The part-of-speech connection check section checks the connection between the two words that make up the second character string based on the part-of-speech. This can be obtained by referring to the word dictionary table, and the weights of these connections can be obtained from the connection table 1. This weight information is added to the "certainty" column (field) of the conversion unit table.

The semantic connection check section checks the connection between two words constituting the second character string based on their semantic relationship. The first classification of words was from this standpoint.
This is the connection information shown in Figure 2. Connection information for each word is
It can be obtained by referring to the word dictionary table from the storage number in the word dictionary table written in the conversion unit table, and the weight of these connections can be obtained from the connection table 2.
Add these weight information to the certainty column (field) of the conversion unit table. The process ends by assigning weights to all second character strings written in the conversion unit table.

Field information as shown in FIG. 13 is set in the converted word field register as field storage means in the REGISTER shown in FIG. 17.
This set has two cases.

One is when a word is uniquely determined by kana-kanji conversion, and the field information of the uniquely determined word in the conversion unit table is set in the converted word field register by the conversion result processing control unit. be done. Of course, if there is only one word with the same sound, it is treated as uniquely determined and treated in the same way as above. One more
In this case, when one word is selected using the YES key or NO key in a state where there are multiple homophones and cannot be uniquely converted, the field information of the selected word is retrieved from the converted word table using the YES key. It is placed in the converted word field register by the control unit. There are 2 words in the conversion word field register, and when a compound word, that is, two or more words constitute one conversion unit, the field information of the two words is entered in the conversion word field register.

The field information entered in the converted word field register is added to the information in the input sentence field register as field storage means for the input document by the field determination control section. The input sentence field register consists of 16 words, with one word allocated to each field. For example, in the past, words with bits in the field of literature were used 28 times, and words in the field of law were used 20 times.
If company is entered 10 times, science is entered 2 times, and home is entered 4 times, then the above input sentence field register is entered.
16 WORDs are respectively 28, 20, 10, 2, 4. That is, if shown corresponding to Figure 13, [0, 28, 20, 10,
2, 0, 4, 0, 0, 0, 0, 0, 0, 0, 0,
0] are accumulated. Here, the field information in which only the literature item is set in the conversion word field register, that is, one of the field information WORD in FIG. ,0,0,
0, 0, 0, 0, 0, 0, 0], the field determination control unit sets each bit of the converted word field register to each bit of the input sentence field register.
Added to WORD. Therefore, in this example, 1 is added to WORD for literature to create W for the field of literature.
RD becomes 119. Therefore, in this example, the values of each WORD in the input sentence field register are [0, 29, 20,
10, 2, 0, 4, 0, 0, 0, 0, 0, 0, 0,
0, 0].

The field weight control section is responsible for checking each of the single or plural second character strings converted from the conversion unit based on field information and assigning a weight to each of them. This control section is activated by the conversion unit table creation control section and returns the control to the activated control section. As shown in the flowchart in Figure 50, weighting is done based on field information.In the first step, the conversion unit table in Figure 25 is checked to see if there are any converted words. Based on the field information for the 22nd word of the word dictionary shown in Figure 9 (the details are shown in Figure 13) and the contents of the conversion unit table in Figure 25, the contents of the input sentence field register mentioned above, singular or After multiplying the field information of each of the plurality of conversion words for each field, in the next step, the results of the above multiplication are summed for each conversion word and used as basic data for the weight information of each conversion word. . In addition, in order to balance with other weight information, such as weight information from a grammar check, etc., in the above-mentioned multiplication, a correction constant, for example 30, is multiplied and added to the AC of each corresponding word in the conversion unit table of FIG. 25, respectively. Weighting based on field information is performed using the method described above.

The conversion word table is a table in which information on the conversion results compiled in the conversion unit table is compiled for display on a CRT. FIG. 28 shows its format. The capacity is 1024W. 1
One block is assigned to one conversion unit. One block consists of a HEAD that provides the definition of the block and multiple records corresponding to multiple conversion words.

The HEAD section consists of five items: iRL, iRN, iDN, iCL, and iCC. iCC is 20W, others are 1W
It is. iRL is the record length, iRN is the number of records, iCL is the number of characters in the kana reading of the conversion unit,
iCC is the character code of the conversion unit in kana reading. iDN is a register indicating the number of the converted word displayed on the CRT. For example, when the converted word entered in the third record of a certain block is displayed, the iDN will be 3. Also, if HEAD
When the kana reading of the part is displayed, it is set to 0.

The record consists of four items: iFL1, iFL2, iAC, and iKC. iFL1, iFL2, iAC are
1W, iKC is generally several W (variable length).

Each record is a record in which the conversion result of the conversion unit, that is, the conversion word, is entered in the form of a character code. iFL
1 is an item for entering the field of the converted word, and is the same as that described in the word dictionary. iFL
2 is for entering the field corresponding to the second word when the converted word consists of a plurality of words. iAC is the "certainty" of the converted word. This is already listed in the conversion unit table.

iKC is the character code of the converted word.

Note that the length iKL of the iKC is determined by the maximum length of the character code string of the conversion word. For example, if "city" and "year" are selected in response to the converted word "toshi", a love out code (all 1s) is added after the year to match the lengths.

The conversion word table is initialized by the conversion result processing control unit. Also, the iDN is initialized to 1 by the conversion result processing control unit,
Changed by the BACK key control section, INV key control section, and NO key control section.

Single Kanji Selection Table FIG. 29 is a comparison table showing the addresses of PM2 with respect to input data from the keyboard, and is used for single Kanji selection control in single Kanji input.

The conversion result processing control unit processes the results obtained by the conversion control unit and the conversion unit table creation control unit.
In addition to writing to page memory PM1 for display on the CRT, press the BACK key by the operator.
It is responsible for creating a conversion word table in preparation for changing the displayed characters according to the operation of the INV key, NO key, etc.

The conversion result processing control unit first moves the contents of the converted sentence buffer in the range indicated by conversion parameter 1 from the beginning into PM1 indicated by PNT1 until a kanji shift code is found. of course
PNT1 is incremented sequentially. if,
If found, determine whether MCW9 is positive or negative.
When MCW9 is negative, that is, a conversion word corresponding to the conversion unit is not found, the conversion text is repeated from the next Kanji shift code to the range indicated by conversion parameter 1, MPA2, or until a Kanji shift code is found again. Move information from buffer to page memory PM1. The transfer of the kanji shift release code will be omitted.

When MCW9 is 1, that is, when one conversion word corresponding to the conversion unit is found, the word dictionary table number corresponding to the conversion word written in the conversion unit table is converted to a character code, and PNT1 is Write to the page memory PM1 shown. At that time, PNT1
are incremented sequentially. Next, the field of the converted word is set in the converted word field register, and the field determination control section is activated. When the control is returned from the field determination control unit, the conversion parameter 1, MPA, is set again from the Kanji shift release code of the conversion unit.
The information is moved from the converted sentence buffer to the page memory PM1 until the range shown by 2 or until the Kanji shift code is found again.

When MCW9 is 2 or more, that is, when multiple conversion words corresponding to the conversion unit are found, bit 15 of conversion parameter MPA4 is set to 0 and bit 14 is set to 1, and written to PM1 indicated by PNT1.
Increment 1. Next, writing of the conversion word table is started. The format of the conversion word table is as described above. The conversion result processing control unit refers to the conversion unit table and extracts nine MCW conversion words from the conversion unit table in descending order of "certainty". By referring to the word dictionary table, the character code and field information of the converted word are found and written into the converted word table. Also, in the HEAD of the block written in the conversion word table,
Read the conversion unit and write other information. At that time, enter 1 in iDN.

Once you have finished writing to the converted word table, update MPA4. After that, return to DM1,
Write the character code corresponding to the first record of the block just written to the conversion word table. At this time, even if a love-out code is written in the character code, it shall be written as is without omitting it.
However, as mentioned above, the love out code is displayed as a space on the CRT. Once you have finished writing the conversion word, write the conversion parameter 1,
Information is moved from the converted sentence buffer to the page memory PM1 until the range indicated by MPA2 or until the Kanji shift code is found again.

This concludes the explanation of the cases where MCW9 is negative, the case where it is 1, and the case where it is 2 or more. Next, the conversion result processing control unit uses a parameter that specifies the conversion end range +1.
Set up MPA3.

The set value of MPA3 is set to the last address at which information was transferred from the converted text buffer to PM1 +1.

Next, the kanji result processing control unit checks whether the number of conversion words corresponding to the conversion unit is 1 or not, and if it is 1, writes the field of the conversion word into the conversion word field register. Start the field determination control section. When control is returned from the field determination control section, the conversion result processing control section returns the control to the conversion control section.

The field determination control section adds the data entered into the converted word field register as described above to each field of the input sentence field register. Therefore, if the contents of the learning word field register are added one after another, the input sentence field register will overflow, but the field determination control section takes care to prevent such overflow by adding all the contents when the addition result is 31 or more. Control is performed to subtract 1 from the input sentence field register. The field determination control unit is a conversion result processing control unit or YES
It is activated by the key control section, and after the control is completed, the control is returned to the activated control section.

The print key control unit has the role of outputting information stored in the page memory to the outside. It is started by the main control section, and after the operation is completed, control is returned to the main control section again. The print key control section is
The information from the beginning to the end of page memory PM1 is output to the output device OU in order.
Anything other than bit15=bit14=0 is a control code and should not be output.

The cursor shift key control unit has the role of moving the cursor displayed on the CRT one by one. It is activated by the main control unit and returns the control to the main control unit after the operation of the cursor shift key is completed. The cursor shift key control unit has a function of incrementing PNT1 in order to shift the cursor.

The EDIT key control unit has the role of searching the second character string on the display screen for locations where a plurality of converted words exist. The EDIT key control section is activated by the main control section, and returns its control to the main control section again after its operation is completed.

The EDIT key control unit scans the page memory PM1 and searches for a position where bit15=0 and bit14=1. When scanning page memory PM1, PNT1 is sequentially incremented, so
By ending the operation of this control section at the position where the information is found, the PNT1 can be automatically set to the position of the information. The above information means that there are multiple conversion words that follow it, so
With the above operations, it is possible to search for locations where multiple conversion words exist in the second character string,
On the CRT, the cursor symbol CC will be displayed at the position of the * mark displayed in front of the conversion word.

The YES key control unit converts the second character string displayed on the CRT out of the plurality of second character strings when the first character string is converted to a second character string and there are a plurality of converted words. In order to determine the field of the input sentence by using the selection function and the selected second string,
It has the role of activating the field determination control section and the function of searching for the position of the second character string to be selected next on the CRT when control is returned from the field determination control section.

The YES key control section is activated by the main control section, and returns its control to the main control section after completing its operation.

The YES key control unit first checks whether the information in PM1 pointed to by PNT1 is bit15=0 and bit14=1. If bit15=0 and bit14=1
If not, the control is immediately returned to the main control section. In other examples, bit15=0 and bit14=1
Otherwise, the control may be transferred to the control from line 14 on the next page. If bit15=0 and bit14=1, perform the following operation.

YES key control unit is instructing PNT1
The conversion word table is referred to by the address pointed to by the lower bits 0 to 9 of the information in PM1. The HEAD part of the block in the conversion word table currently contains
The record number that stores the converted word displayed on the CRT is stored in the iDN. The field of the converted word is stored in the record.
The YES key control section writes the field information into the converted word field register and activates the field determination control section. When control is returned from the field decision control section,
The YES key control unit sets bit 15 of the contents of the page memory PM1 corresponding to the * mark on the CRT currently pointed to by the cursor symbol CC to 1. This results in
The DCOT (display control circuit) will change the * mark that was previously displayed on the CRT to a space.

Next, search for the conversion word on the CRT to be selected. This function is basically the same as the EDIT key control section. However, while the EDIT key control section scans from the beginning of PM1, this control section currently scans PNT1.
The difference is that it scans the area after the point pointed to.

This completes the operation of the YES key control section.

When the first character string is converted to the second character string and there are multiple second character strings, the NO key control section is used to convert It plays the role of displaying the second character string.

The NO key control section is activated by the main control section and returns its control to the main control section after its operation is completed.

The NO key control unit first checks whether the information in PM1 pointed to by PNT1 is bit15=0 and bit14=1. If bit15=0 and bit14=1
If not, the control is immediately returned to the main control section. If bit15=0 and bit14=1, perform the following operation.

The NO key control unit first refers to the conversion word table using the address pointed to by the lower bits 0 to 9 of the information in the PM1. The record number storing the converted word currently displayed on the CRT is stored in the iDN in the HEAD portion of the block in the converted word table.

The NO key control increments the current iDN. At this time, if the iDN exceeds the total number of records iRN, change the iDN to 1. Therefore, the NO key control will be activated every time it is activated.
The iDN will be rotated in a fixed direction within the range from 1 to iRN. After that, the NO key control section
Pick up the character code stored in the record corresponding to the iDN just changed, and change the corresponding information in page memory PM1. After the above operations are completed, control is returned to the main control section.

The BACK key control section is almost the same as the NO key control section. The difference is that the NO key control section rotates iDN in the conversion word table in the increasing direction within the range of 1 to iRN, whereas the BACK key control section rotates it in the decreasing direction.

The INV key control part is the second character string (conversion word)
It plays the role of converting into the first character string (conversion unit).

The INV key control section is activated by the main control section, and returns control to the main control section after its operation is completed.

The INT key control unit first refers to the conversion word table using the address pointed to by the lower bits 0 to 9 of the information in the PM1. The reading corresponding to the conversion word currently displayed on the CRT, that is, the first character string, is written in the HEAD portion of the block in the conversion word table. The INV key control unit changes the second character string on the page memory PM1 to the first character string. Generally, the first string has more characters than the second string, so the page memory
By shifting all the information in PM1 by that amount, no information is missing.

After the above operations are completed, control is returned to the main control section.

The INT key control unit initializes all control units, various memories, registers, etc., and is activated by the main control unit when inputting text starts. After the operation is completed, control is returned to the main control section.

The DIN input control section reads data from DIN, the means for directly specifying and inputting kanji as described above, and performs control to input it to PM1. i.e. AMK
is a single kanji input, and when IDK is in direct input mode, it reads data from DIN and outputs it to PM1.

The single kanji input control unit sends the data from the FIFO, that is, the reading of the kanji, to the keyword register when the AMK is set to input a single kanji. The KF search control section is
KF with the reading of the kanji entered in the keyword register,
That is, the kanji file is referred to and the corresponding kanji group is sent to address PNT2 of page memory 2 (PM2).
The kanji group that has entered PM2 is displayed on the CRT through the DCOT under the control of the single kanji input control section. As described above, only the bottom row is displayed, and the other rows are hidden, as shown in FIG. That is, in the single kanji input state, the DCOT selects PM2 data and displays it on the CRT under the control of the single kanji input control section. A desired kanji is selected from the displayed kanji group by pressing the key corresponding to the display, and the kanji code is sent to PM1.

As a modified example, in the single kanji input described above, the sentence and the enumeration of the single kanji were not displayed at the same time.
By performing the procedure as shown in FIG. 31, the operability can be improved.

If the operation of the character processing device is explained using the flowcharts shown in FIGS. 33 to 60, FIG. 33 shows the main control flowchart, and in the figure, Y stands for YES and N stands for NO. .

The main control section has mode keys (DIN, IDK,
AMK), FIFO, print key, cursor shift key, EDIT key, YES key, BACK key, INV
Reads information from DEVICE such as key, NO key, INT key, determines the corresponding job (JOB),
Activate the corresponding control unit. Here, the mode key has a function that can be read at any time.
In addition, other devices have a function that indicates whether information has been received by the device, that is, whether there is information that has not been read yet, and
It shall have a function to save the data until it is read.

When there is input from FIFO, the control section that is activated differs depending on the state of the mode key. In other words, when the AMK key is set to automatic, the KB1 input control section,
When the AMK key is a single kanji and the IDK key is indirect, the single kanji input control
When the IDK key is direct, it activates the DIN input control section.

If there is an input from the print key, the print control part, if there is an input from the cursor shift key, the cursor shift key control part, if there is an input from the EDIT key, the EDIT key control part If there is an input from the YES key, the YES key control part, BACK key If there is an input from the , the BACK key control section, if there is an input from the INV key, the INV key control section, if there is an input from the NO key, the NO key control section, if there is an input from the INT key, the INT key control section. shall be started.

Figure 34 is a flowchart of KB1 input control, and the part surrounded by dotted lines 101 is a control routine that leaves the last six characters of the kana data that have already been converted into the converted sentence buffer as necessary. This shows that. In this flowchart, N(I) indicates data at address 1 of buffer or memory N.

Also, IHBUF is the conversion statement buffer, ILA is the address information of IHBUF used to transfer the first data part of the reference range for conversion of the next input data after the data in IHBUF is converted, and INPUT is the initial state or conversion control. It is 0 when input is completed, and 1 when input is in progress. IKS indicates a shift state of a kanji, α indicates a shift to a kanji, and β indicates a shift to a non-kanji. When the initial value of "INPUT" is 0, that is, when input is started or immediately after kana-kanji conversion is completed, according to the flow in 101, in the former case, the conversion range is In the latter case, the last 6 characters of the part that has already been converted are packed from the beginning of IHBUF, i.e., IHBUF1, and then the beginning of the conversion range, i.e., MPA1, and the beginning of the conversion reference range, i.e., MPA5. decide.

When “INPUT” is not 0 or 101 above
If the key input after passing through the flow within is a Kanji shift code, check whether the current state is a shift state to a non-Kanji character, that is, IKS=.
When IKS=, change the Kanji shift code to 100 or
Move to 130. The flow starting with 130 is for Romaji input, and conversion from Romaji to Kana is performed as necessary.

In FIG. 34C, it is determined whether or not kana-kanji conversion should be started, and if so, the conversion range and reference range are determined. Further, in FIG. 34B, 102 is a step for determining whether it is a hiragana or katakana, and 103 is a step for determining whether it is a hiragana or katakana, and 103 is a step for determining whether it is a hiragana or katakana.
This is a step to determine whether the character is a capital letter or a letter other than lowercase.

In FIG. 34C, in step 104, an overflow occurs when the current address is greater than 20, and in step 105, the conversion start code is a circle ".", a dot ",", ",", EOP, EOF, NL, etc. It is.

FIG. 35 is a flowchart of Romaji-kana conversion.

FIG. 36 is a flowchart for inputting kanji, etc. using DIN, that is, direct input means.

FIG. 37 is a flowchart of the control unit for inputting a single kanji, that is, inputting the pronunciation of the kanji by sound or kun in kana or romaji, displaying the enumeration, and selecting and inputting the desired kanji from among them. 177) = { is “{” in address 177 of PM2
PM2(182)=} indicates that "}" is inserted in the address 182 of PM2.

Further, step 107 indicates that data is to be input to address PNT2 of PM2.

Also, in Figure 37C, steps 108 and 109 are hiragana chi, ni, na, i, ra, sa, katakana chi, ni, na,
Step 110 is a step to determine whether the name is other than a, a, or sa, and step 110 is a step to determine whether it is a character other than hiragana or katakana.

Fig. 38 is a flowchart of KF search control, Fig. 39 is a flowchart of single kanji selection, Fig. 40 is a flowchart of conversion control,
Each step will be explained in this flowchart.

1 If MPA(3)>MPA(2), the string to be converted does not exist in the conversion statement buffer, so return.

2 Conversion statement BUF within the range of MPA3 and MPA2
Search for conversion units within.

3 If the conversion unit does not exist, return. When this routine is started for the first time and reaches step 2, there is generally a conversion unit, so the process proceeds to step 4.

4 Set the conversion unit range found in step 2 to MCW.
1. Set to MCW2.

5 Initialize the word dictionary table NO.

6. Set initial values to MCW3 and MCW4 as the first character examples to be searched from the word dictionary WF.

7 Initialize NEST DEPTH to 1, which corresponds to the number of words in the compound word. That is, initially it is assumed that the conversion unit consists of a single word.

8 Start the search control unit and search for words using the character strings specified in MCW3 and MCW4 as keywords.

9 If the corresponding word is found as a result of the search, MCW(5)>0
Then, proceed to step 10. If the corresponding word is not found, MCW(5)=-1 and proceed to step 25.

10~13 Reserve MCW3~MCW6 in the NEST table.

14 Write the start address of the word dictionary table that stores the dictionary information of the words found as a result of the search into the E field of the nest table for the number of words found. In this embodiment, up to 24 homophones can be written.

15 Update the word dictionary table free number.

16 When MCW3, MCW4 and MCW1, MCW2 do not match, that is, when the conversion unit is created from two or more words, the value of MCW(3) + MCW(4)
By comparing with the value of MCW(1)+MCW(2), it can be determined whether all searches for conversion units have been completed. If the result of the determination is that there is a remaining part, the process proceeds to step 19, and if there is no remaining part, the process proceeds to step 17.

17 Activate the conversion unit table creation control unit and prepare to write conversion words into the conversion word table.

18 Activate the conversion result processing control unit and write to PM1, conversion word table, etc.
Proceed to step 1 to check whether there is any remaining first character string that has not been converted into the conversion statement BUF.

19 Update NEST DEPTH.

20 Check whether NEST PEPTH overflows or not. If it overflows, it is determined that the search is impossible and the process proceeds to step 24.

21 Check whether the word dictionary table overflows or not. If it overflows, proceed to step 24.

22, 23 Set MCW3 and MCW4 to values indicating the range of the remainder, and proceed to step 8 to search the word dictionary WF again.

24 Parameter when conversion word is not found
Set MCW8 to -1. Proceed to step 18.

25, 26 When the search result word is not found
Subtract 1 from MCW4, if MCW(4)0≠0
If so, proceed to step 8 to perform the search again.
If MCW=0, proceed to step 27.

27 If MCW(7)=1, there is no point in searching any further, so proceed to step 33. If MCW(7)≠1, the process proceeds to step 28 to redo the search at the stage of MCW(7)-1.

28 MCW(7)=MCW(7)-1 29 If the number of readings at the stage of MCW7 is 1, proceed to step 27 again. If it's not 1
It is possible to search again at the MCW7 stage and proceed to step 30.

30, 31, 32 Reserved MCW3,
Restore MCW4 and MCW6, and subtract 1 for MCW4. Proceed to step 8 for re-search.

FIG. 41 is a search control flowchart.

Figure 42 shows a conversion unit table creation control flowchart, where I is the conversion unit table.
This is the ROW number, and the operation of each step will be explained below.

1 The search results by the conversion control unit are entered in the nest table. From this nest table, create a compound word (i.e. conversion word) consisting of a combination of words, and then
Write the number of the word dictionary table in the second column.
(Of course, it may not be a compound word.) 2. Write the number of combinations of compound words in MCW8.

3 Enter the number of words for the combination in the word count column of the conversion unit table.

4 Steps 4 to 24 are for determining the respective parts of speech for the converted words created in step 1.

In step 4, create the Homma conversion unit table.
Initialize ROW No. to 1.

5 If MCW≠, that is, a compound word, proceed to step 6.

If MCW(7)=1, that is, the converted word consists of one word, the process advances to step 24.

6 If the word is a compound word consisting of two words, proceed to step 7. If it is a compound word consisting of three or more words, go to step 23.
Proceed to.

7 Check the part of speech of the first word.

8 If it is a prefix, proceed to step 20. If it is not a prefix, proceed to step 9.

9 If it is a noun, go to step 10; if it is not a noun, go to step 23.

10 Check the part of speech of the second word.

11 If the part of speech of the second word is a noun, proceed to step 21. If it is not a noun, proceed to step 13.

12 If the part of speech of the second word is a noun, proceed to step 22.

If it is not a sa-modular noun, proceed to step 13.

13 If the part of speech of the second word is a suffix, proceed to step 14. If it is not a suffix, proceed to step 23.

14, 15 If the character code of the second word is "target", proceed to step 16; otherwise, proceed to step 23.

16 Enter that it is an adjective verb in the part of speech column of the conversion unit table.

17 Write in the usage column that it is da usage.

18, 19 Steps 5 to 24 are applied to the conversion word of Subete written in the conversion unit table.

20 Enter the part of speech and conjugated form of the second word in the conversion unit table.

21 Enter the noun in the part of speech column of the conversion unit table.

22 Write in the part of speech column of the conversion unit table that it is a sa-variant noun.

23 Enter 0 in the part of speech and conjugation columns of the conversion table.

24 Enter the part of speech and conjugation of the word in the part of speech and conjugation columns of the conversion table.

25 Initialize the certainty column of the conversion unit table to 0.

26-29 Activate the grammar weight control section, frequency weight control section, connection weight control section, and field weight control section to set the certainty column of the conversion unit table.

30 Activate the threshold determination control unit to determine the threshold to be adopted as the conversion word.

31 Set MCW9.

FIG. 43 is a threshold determination control flowchart, in which the threshold is determined using the larger value of the maximum certainty value entered in the conversion unit table minus 1000 and 0.

FIG. 44 shows a grammar weight control flowchart, where I indicates the ROW number of the conversion unit table.

In this flowchart, if a character string to be referenced exists at the front of the conversion unit, a front grammar check routine is activated.

If there is a character string to be referenced at the end of the conversion unit, an unchanged kana part check routine and a conjugated form check routine are activated.

The grammar check described above is applied to all conversion words written in the conversion unit table.

FIG. 45 shows a front part grammar check flowchart, which increases the certainty value when the converted word is a verb and the front part of the converted unit is a case particle.

Figure 46 is an unchangeable kana part check flowchart. When the number of words in a converted word is 1 and it has an unchangeable kana part, it is checked whether the unchangeable kana part matches the character example following the conversion unit. Gives the value of certainty.

FIG. 47 is a conjugation form check flowchart. When the converted word is a verb, adjective, or subverb, once the conjugation form is determined, the character string that follows it is also limited to a few characters. Taking advantage of this,
Give a probability value for each converted word.

FIG. 48 is a frequency weight control flowchart, where I is the ROW number of the conversion unit table. Here, when the converted word consists of one word, the frequency of the word is given as the certainty value, and when it consists of two or more words, the sum of those frequencies is given as the certainty value.

Figure 49 is a connection weight control flowchart.
Regarding the converted word when the number of words is 2, the strength of the connection is determined using connection tables 1 and 2, and certainty is given.

The above operation is performed for all conversion words entered in the conversion unit table.

FIG. 50 is a field weight control flowchart, where IWD is a word dictionary table and IBR is an input sentence field register (16W). This flowchart shows the flow for determining the weight of each conversion word based on the contents of the input sentence field register IBR and the field information of each conversion word.

FIG. 51 is a conversion result processing flowchart, and each step will be explained below.

1 Initialize current address J to MPA3.

2 Repeat the following operation until the current address J becomes MPA(2)+1.

3 Read the contents of the current address J from the conversion statement BUF.

4 If the read content is a kanji shift code, proceed to step 5 instead of proceeding to step 9.

5 The read contents are displayed as PM1 indicated by PNT1.
Move it inside.

6 Increment PNT1.

7 Increment the current address J.

8 At step 8, the processed address +1 is stored in the current address J, and that value is set in MPA3.

9 When MCW(9)=-1, that is, the conversion word does not exist, proceed to step 10. When MCW(9)≠-1, that is, the conversion word exists, proceed to step 17.

10 to 16 Read the contents of the conversion statement BUF and write them to the page memory PM1 in order. At this time, the current address J and PNT1 are sequentially incremented.
If the current address J reaches MPA(2)+1 on the way, proceed to step 8. If the read content is a kanji shift code, the process advances to step 8. Furthermore, if the read content is a Kanji shift release code, writing the code to PM1 is omitted.

17 If there is one converted word, proceed to step 23; if there are two or more converted words, proceed to step 18.

18 Write bit15=0 and bit14=0 of MPA4 to PM1 indicated by PNT1.

19 Increment PMT1.

20 From the conversion unit table, in descending order of “certainty”
Extract only the number shown in MCW9.

21, 22 Refer to the word dictionary table, find the character code and field information of the converted word, and write it into the converted word table.

At that time, set iDN=1.

23 Update MPA4.

24-26 Extract the conversion word with the highest "certainty" from among the conversion words entered in the conversion unit table, and convert the conversion word into a character code by referring to the word dictionary table. After that, the character code is written into PM1.

27 Update current address J.

28 If there is one converted word, proceed to step 29; if there are two or more converted words, proceed to step 10.

29 Find the field information of the word constituting the converted word from the word dictionary table, enter it in the converted word field register, and start the field determination control section.

<Note> In reality, it is necessary to limit the increment of PNT1 so that it does not take a value greater than 193.

FIG. 52 is a field determination control flowchart for automatically determining the field of an input sentence, where IHBR is a conversion word field register and IBR is an input sentence field register (16W).

FIG. 53 is a cursor shift key control flowchart, in which PNT1 is cyclically incremented from 1 to 192.

Figure 54 is an EDIT key control flowchart, starting from the beginning of page memory PM1 with bit15=0,
Find the location where bit14=1 and set that value to PNT1
Set to .

Figure 55 is a NO key control flowchart, and only when the cursor is at the position marked *
Enable ROuTiNE.

From the contents of PM1 indicated by PNT1, it is possible to know the address where the converted word information stored in the converted word table exists. The iDN written in the conversion word table is cyclically incremented within the range of 1 to iRN, and the conversion word indicated by the iDN is written to PM1.

FIG. 56 is a BACK key control flowchart, which differs from the NO key control section only in the following points. That is, the iDN is cyclically decremented between iRN and 1.

FIG. 57 is an INV key control flowchart. When the cursor is at the * mark position, by checking the contents of PM1 indicated by PNT1, reading information stored in the conversion word table (first
character string) exists. PM the first string found in this way
Move within 1. At that time, since the length of the second character string is generally shorter than the length of the first character string, in order to secure a place to store the first character string in PM1,
The character string that exists after the PM1 cursor must be moved.

FIG. 58 is a YES key control flowchart, and this routine operates only when the cursor is placed on the * mark. The corresponding position in the conversion word table is manipulated based on the contents of PM1 pointed to by PNT1.
After storing the field information stored in the record corresponding to the iDN in the field register, the field determination control section is activated.

Next, bit 15 of PM1 information indicated by PNT1
By setting 0 to 0, the * mark will be removed from the CRT display.

Next, search for the * mark after the cursor and move the cursor to that position. At that time, if there is no * mark after the cursor, PNT(1) = 1 and the cursor is brought to the top of the CRT display.

FIG. 59 is a print control flowchart, in which code data is output to the printer from the beginning of PM1. At this time, only when bit 15 and bit 14 of the code data are both 0, it is character data and is output.

FIG. 60 is an INT key control flowchart, which initializes various parameters and BUF.

According to the present invention, with the above-described configuration, the operator does not need to specify the field of the input sentence, and the device automatically determines the field of the input sentence, thereby reducing the burden on the operator. can.

[Brief explanation of drawings]

1A is a perspective view of a character processing device according to the present invention, FIG. 1B is a front view of a partial keyboard of the character processing device shown in FIG. 1A, FIG. 2 is a front view showing a document, and FIG. Figure 4 is an explanatory diagram showing the input procedure,
FIG. 5 is a front view showing the screen, FIG. 6 is a block diagram showing the character processing device, FIG. 7 is an explanatory diagram showing the ROM, FIG. 8 is an explanatory diagram of the page memory PM1, and FIG. 9 is a word dictionary. FIG. 10 is an explanatory diagram showing parts of speech. FIG. 11 is an explanatory diagram showing conjugated forms. FIG. 12 is an explanatory diagram showing connection information. FIG. 13 is an explanatory diagram showing field information. The figure is an explanatory diagram showing the word dictionary structure, Fig. 15 is a block diagram showing DCOT, Figs. 16A and 16B are block diagrams showing DIN, Fig. 17 is an explanatory diagram showing RAM, and Fig. 18 The figure is a block diagram showing a character processing device, FIG. 19 is an explanatory diagram showing a converted sentence buffer, FIG. 20 is an explanatory diagram showing conversion parameters, and FIG.
Figure 22 is an explanatory diagram showing a romaji-kana correspondence table, Figure 22 is an explanatory diagram showing a word dictionary table, and Figure 23 is an explanatory diagram showing a word dictionary table.
Figures A and 23B are explanatory diagrams showing nest tables, Figure 24 is an explanatory diagram showing conversion parameters,
Figure 25 is an explanatory diagram showing the conversion unit table,
6 is an explanatory diagram showing a connection table, FIG. 27 is an explanatory diagram showing a connection table, FIG. 28 is an explanatory diagram showing a conversion word table, FIG. 29 is an explanatory diagram showing a single kanji selection table, FIG. 31 and 32 are front views showing the screen, FIG. 33 is a main control flowchart, FIGS. 34A, 34B, and 34C are KB1 input control flowcharts,
Figure 5 is the Romaji-kana conversion flowchart, Part 3.
Figure 6 is the DIN input control flowchart, Figure 37A, Figure 37B, and Figure 37C are the single kanji input control flowchart, Figure 38 is the KF search control flowchart, and Figure 39 is the single kanji selection control flow. Chart, Figure 40 is a conversion control flowchart, Figure 41
The figure is a search control flowchart, Fig. 42A, Fig. 4
2B and 42C are conversion unit table creation control flowcharts, FIG. 43 is a threshold determination control flowchart, FIG. 44 is a grammar weight control flowchart, FIG. 45 is a front grammar check flowchart, and FIG. Figure 47A, Figure 47B, and Figure 47C are conjugated form check flowcharts.
Figure 8 is a frequency weight control flowchart, No. 49.
Figure 50 is a flowchart for connection weight control, Figure 50 is a flowchart for field weight control, Figure 51 is a flowchart for conversion result processing, Figure 52 is a flowchart for field determination control, Figure 53 is a flowchart for cursor shift key control, and Figure 54 is a flowchart for field determination control. The figure is an EDIT key control flowchart, Figure 55 is a NO key control flowchart, Figure 56 is a BACK key control flowchart, Figure 57 is an INV key control flowchart, Figure 58 is a YES key control flowchart, and Figure 59 is a flowchart for INV key control. The figure is a print control flowchart,
FIG. 60 is an INT key control flowchart. Here, KB1, KB2, KB3, KB4 are keyboards, 13 is a screen, 15 to 33 are keys,
ROM is read-only memory, RAM is random access memory, KF is Kanji file, and CPU is central processing unit.

Claims (1)

[Scope of Claims] 1. A character processing device that inputs kana data from an input means and converts the kana data corresponding to the reading of kanji out of the kana data into kanji data to create a document, comprising: a storage means that stores kanji data to be used and field information to which the kanji data belongs; a conversion means for reading data; a field storage means for storing field information of the kanji data converted by the conversion means; a document field storage means for storing field information of the document currently being created; If there is only one converted kanji data, the field information indicating the field to which the kanji data belongs is determined to be information for modifying the field information to which the document currently being created is stored in the field storage means. If the content of the document field storage means is modified according to the content, and the converted kanji data consists of a plurality of kanji data, the current document is created according to the field information indicating the field to which each of the plurality of kanji data belongs. field determination control means for determining field information to which a document belongs to be modified information, and modifying the contents of the document field storage means with a plurality of contents stored in the field storage means; field weight control means for creating weight information based on the field information created by the field weight control means; and control means for adjusting the priority of the kanji data based on the field weight information created by the field weight control means. Character processing device.