JPH0256058A - Kana to kanji conversion control system - Google Patents

Abstract

PURPOSE:To efficiently execute the reconversion by setting reading of a character- string as an object immediately before punctuation marks or a definite same sound word appearing first after a KANA (Japanese syllabary) character-string from character-string obtained by bringing a reading correction object KANA character- string to reading correction. CONSTITUTION:A homonym which is outputted as a result of KANA to KANJI (Chinese character) conversion is accumulated in a conversion result storage buffer 27 in accordance with reading. An editing work part 24 retrieves a homonym group and delivers it to a screen display processing part 26. When it is detected that a selection/execution key 15 or a conversion/next candidate key 14 is operated, an input control part 21 informs its face to the editing work part 24. The editing work part 24 detects a head position of the succeeding character-string which does not become an object of a reading correction, derives a KANA character-string immediately before that which appears first in punctuation marks or a definite homonym from a character- string after said position and accumulates it in an input buffer. A KANA to KANJI converting part 22 brings the KANA character-string containing a reading correction character delivered from the input control part 21 to KANA to KANJI conversion (reconversion).

Description

[Detailed Description of the Invention] [Objective of the Invention] (Field of Industrial Application) This invention relates to a document creation device using a kana-kanji conversion method that converts an input kana character string into a character string containing kanji. The present invention relates to a kana-kanji conversion control system suitable for correcting and reconverting input characters after converting them into kana-kanji characters.

(Prior Art) Conventionally, in this type of document creation device, it has been common to apply either Method 1 or Method 2 shown below as a correction method when an error in input characters is found after Kana-Kanji conversion. Met.

■ Method 1 involves performing key operations to return all converted character strings (character strings containing kanji) to the character strings before conversion (kana character strings), and then performing key operations to correct incorrect characters (or words). This involves performing key operations and finally converting everything again. With this method, after converting all the characters into the original kana character string, the cursor must be moved by one character position to indicate the character to be corrected. Therefore, if the kana character string has a large number of characters, the number of key presses The disadvantage was that it increased. Furthermore, it is not easy to find incorrect characters in a kana character string.

■ Method 2 is applied to document creation devices that cannot convert character strings after kana-kanji conversion back to kana character strings before conversion. Next, key operations are performed to open up a space in order to secure an input character area for the kana character string necessary for correction (a kana character string that indicates the correct reading of the desired converted character string), or to delete unnecessary character strings. The user performs key operations to manually override the correct kana character string to the character string to be corrected, and finally converts the newly input kana character string. In this method,
The correction procedure is very complicated because it is not possible to correct only the incorrect characters, and a new reading string must be entered for the incorrectly converted character string (i.e., a homophone). Ta.

(Problem to be Solved by the Invention) As mentioned above, conventionally, when attempting to correct characters input by mistake after converting kana-kanji, the device cannot return the converted character string to the pre-converted kana character string. However, since it is not possible to correct only the incorrect characters, and it is necessary to correct the incorrectly converted character strings (i.e. homophones), the correction procedure becomes extremely complicated. . In addition, with a device that can only correct incorrect characters, all character strings that have been converted into kana-kanji must be returned to the original kana character string, making it difficult to find the corrected part and There was a problem that required many cursor movement operations to give instructions. Furthermore,
After correcting the pronunciation, if all subsequent character strings were converted (reconverted) from kana to kanji again, there was also the problem that homophones that had been confirmed would also be converted.

Therefore, the problem to be solved by this invention is that incorrectly inputted kana characters can be easily corrected even after kana-kanji conversion.
Furthermore, the purpose is to efficiently perform reconversion after correcting kana characters (reading).

[Structure of the invention] (Means for solving the problem) This invention instructs to correct the pronunciation of a homophone candidate that is a homophone selection target in a character string containing kanji that has been converted into kana-kanji. and means for converting a corresponding homophone candidate into a kana character string in response to an instruction from the reading correction instructing means, and displaying the same character string as a kana character string to be corrected with a specific display attribute. , a means for instructing to vary the range of the kana character string to be read corrected by one character position, and a means for instructing to vary the range of the kana character string to be read corrected to be displayed on the screen with a specific display attribute according to the instruction of this means. a reading correction means for replacing the kana character string to be corrected that is displayed on the screen with a specific display attribute with another kana character string, and a reading correction means that first appears after the kana character string replaced by the reading correction means. detecting means for detecting a punctuation mark or determined homophone from the character string converted into kana-kanji, and reading the character string from the replaced kana character string to immediately before the detected punctuation mark or homophone. It is characterized by performing kana-kanji conversion on the target again.

(Operation) According to the above configuration, when a correction instruction is given by the pronunciation correction instruction means, the homophone currently being selected is converted into a kana character string indicating its pronunciation. This converted kana character string is displayed on the screen with a specific display attribute to indicate that it is a kana character string subject to reading correction. The range of kana character strings subject to reading correction that is displayed on the screen with this specific display attribute is determined by the instruction from the reading correction range change instruction means.
Variable character by character. When the pronunciation correction is performed to replace the kana character string determined in this way with another kana character string, the first punctuation mark or Reconversion is performed targeting the reading of the character string immediately before the confirmed homophone.

(Embodiment) FIG. 1 is a functional block configuration diagram showing an embodiment of a document creation device to which the present invention is applied. In the figure, 10 is a keyboard. In addition to various character keys (not shown), the keyboard 10 includes a left cursor key 111 for instructing cursor movement to the left, and a right cursor key 12 for instructing cursor movement to the right. Function keys 13 are provided for realizing various instruction functions in combination with up and down cursor keys and other keys. In this embodiment, the function key 18 is used in combination with another predetermined key, such as the space key 17 to be described later, to instruct correction of the pronunciation of a homophone candidate that is a homophone selection target. The keyboard 10 further includes a conversion/next candidate key 14 for instructing conversion of input kana character strings into kana-kanji characters, switching to the next candidate for the homophone that has been converted to kana-kanji (switching display), A selection/execution key 15 for instructing selection, etc., a cancel key 1B for canceling various instruction states from the keyboard 10, and a space key 17 for inputting a blank space are provided.

21 is an input control unit that performs key manual processing and key discrimination processing from the keyboard IO, and gives instructions to various parts within the device according to the key discrimination results; 22 converts input kana character strings into character strings containing kanji; a kana-kanji conversion unit that performs processing to re-convert (re-convert) a kana character string corrected by reading correction after kana-kanji conversion;
3 is a dictionary consisting of a word dictionary in which independent words and grammatical information are stored, and an attached word dictionary in which dependent words are stored, which are used for kana-kanji conversion by the kana-kanji conversion unit 22; 24 is an input control unit 21;
This is an editing work unit that performs various editing processes in accordance with instructions from the editor. Reference numeral 25 indicates a display device such as a liquid crystal display used for displaying documents, etc.; 2B indicates a screen display processing unit that displays documents on the display unit 25; and 27 indicates homophones (homonyms) converted by the kana-kanji conversion unit 22. This is a conversion result storage buffer in which candidates for different synonyms are stored.

Next, the operation of one embodiment of the present invention will be explained with reference to FIGS. 2 to 6.

When creating a document, the user operates the character keys on the keyboard 10 and inputs a kana character string (reading character string) indicating the pronunciation of a desired sentence using kana manual input, Roman alphabet input, or the like. Here, in order to create the sentence ``It was a wonderfully nice day today.'', we will use the kana character string ``Today was a wonderful day'' to show the pronunciation. It is assumed that you have incorrectly entered "It was a wonderful and good time."

When the character keys on the keyboard 10 are operated and a kana character string with a pronunciation of ``Today was a wonderful day.'' is input, the input control unit 21 displays this kana character string character by character. The data is stored in an input buffer that cannot be shown, and the editing section 24 is notified of this fact.

The editing section 24 uses the screen display processing section 26 to display the characters accumulated in the input buffer by the input control section 21 on the display 2.
The images are sequentially displayed starting from the cursor position currently displayed on the screen in step 5. This device is equipped with a cursor position register (not shown) that stores the coordinates of the X digit and Y line on the screen, and a display memory (not shown) that stores the coordinates of the X digit and Y line on the screen. (which the processing unit 2B has) is managed. Then, the screen display processing section 2B writes the cursor pattern into the display memory according to the contents of the cursor position register and displays the cursor. Further, the editing work unit 24 shown in FIG. 1 and the like are configured so that the current cursor position can be determined by looking at the contents of the cursor position register. In addition, the position of the first line on the screen is also managed separately, so adding the contents of the cursor position register and the X and Y coordinates of the first line on the screen will determine the position of the current cursor in the page. It is now possible to grasp the line column position.

Therefore, the editing section 24 reads the contents of the cursor position register, sequentially increments digits from this position coordinate, and passes the display position coordinate and character code to the screen display processing section 26. In this state, punctuation manually, conversion/next candidate key 14,
Alternatively, when any of the selection/execution keys 15 is operated or the input buffer 28 becomes full,
The input control unit 21 passes the kana character string stored in the input buffer to the kana-kanji conversion unit 22, and activates the kana-kanji conversion. Here, the kana-kanji converter 22
It is assumed that the Kana-Kanji conversion is activated. The subsequent operation is shown in the flowchart of FIG.

The kana-kanji conversion unit 22 performs a kana-kanji conversion process on the kana character string passed from the input control unit 21 in response to a kana-kanji conversion instruction from the input control unit 21 (step SL). That is, the kana-kanji conversion unit 22 converts the input kana character string into the dictionary 23.
Search for independent words and adjunct words, extract clauses by taking into consideration grammatical connections, use of predicates, etc., output homophones of independent words for which the grammar check was successful, and convert them into kana-kanji. Perform the conversion.

Since the method of kana-kanji conversion is well known to those skilled in the art, detailed explanation will be omitted.

The homophones output as a result of the kana-kanji conversion in step Sl are stacked in the conversion result storage buffer 27 in accordance with their readings. FIG. 4 shows the output code when homonyms exist. If the conversion result is a uniquely determined character (plain name, katakana, alphanumeric character, symbol), the character code indicating that character is the document buffer in which the document is stored (not shown/each code storage location in the document buffer is a page). ), but if there is a homophone as a result of the conversion of the reading, the conversion result of this reading will be the fourth one.
It is stored in the document buffer in the form of the code shown in the figure. Homophones corresponding to the readings are stored in the conversion result storage buffer 27, and this code (hereinafter referred to as undefined conversion result code) contains a pointer (homophone) indicating the beginning of the area where the homophones are stored. A synonym buffer pointer) 41 is set. Further, a flag 42 for distinguishing this code from a normal character code, and a flag 43 for indicating that this homophone is a target of selection processing are also provided. That is, the attribute that a homophone exists is indicated by a flag 42, and the corresponding homophone group is indicated by a pointer 4.
The data is stored from the storage position of the conversion result storage buffer 27 indicated by 1.

Here, the conversion result of the above-mentioned input kana character string "Today was a wonderfully good day." is shown in FIG. There are homophones for the pronunciations of ``Kiyou'', ``Suba'', ``Good'', and ``Tenki'', which are ``Ha''.

``Nashiri'' and ``Desu'' have been confirmed as particles, suffixes, and auxiliary verbs in flat names. Therefore, the character string stored in the document buffer is in the form indicated by reference numeral 51. Note that the same unconfirmed conversion result code is embedded for the number of characters of the currently displayed candidate. In addition, a selection flag (43) is set in the undefined conversion result code of the last reading part where the homophone exists. This indicates that this reading is subject to homophone selection. The conversion result storage buffer 27 includes a character code storage section 52, a character count storage section 53, a flag (displaying flag) 54 indicating a word to be displayed as a candidate from a homophone group, and the next homophone for the same-yomi. Pointer storage unit to word (if end code E is stored in the pointer storage unit, it indicates that the word is the last of the homophones) 5
5 storage units are provided. As shown in Figure 5,
There are five homonyms for ``Kyo'': ``Kyoyo/Skillful/Bulletin/Kiyou/Kiyou'', and ``Suba'' and ``Good''.
It has been shown that there are 2, 7, and 3 homonyms for ``tenki'', respectively. In this embodiment, words written in hiragana are also stored as one of the homonyms. A display flag 54 is set for each reading to be displayed as a candidate.

As shown in FIG. 5, a display flag 54 is currently set for each of the candidates ``employed,''``fine,''``drunk,'' and ``weather.''

When the conversion is completed, the processing is passed to the editing section 24. The editing section 24 displays the tentative conversion result using the screen display processing section 2B. That is, the editing section 24 reads codes one by one from the storage position of the document buffer corresponding to the cursor position at the time when reading input is started. If the read code is a character code, the editing work unit 24 passes the character code to the screen display processing unit 26, and the screen display processing unit 2B displays the character corresponding to the position specified by the cursor position register and moves the cursor Update the contents of the position register by one digit.

If the code is an undefined conversion result code, pointer (4
Access the conversion result storage buffer 27 from the contents of 1), read the code of the word whose display flag 54 is set,
The character codes corresponding to the number of characters stored in the character number storage section 53 are passed to the screen display processing section 2B with overlined attributes attached. In this state, as shown by reference numeral 31 in Figure 3, the message ``The appointment was wonderful and intoxicating.'' is displayed, and the cursor (the rectangular pattern with diagonal lines in Figure 3) is positioned next to the period point ``.''. will be displayed.

Here, if the conversion result "Kyouse" is not the desired one because the user mistakenly inputs "Kyo" instead of "Kyo", the user corrects the pronunciation as follows.

First, before correcting the pronunciation, the user selects, for example, the first character position "
Perform the operation to move the cursor to "Drunkenness." The contents of this operation are notified to the editing section 24 via the input control section 21, and the editing section 24 reads out the contents of the cursor position register and changes the line column position in the instructed direction, as well as changing the cursor position. is converted into a line column position within the page, the document buffer (51) is accessed, and the code at the position is read. Now, from the position corresponding to "drunkenness" [1:
It is assumed that the undefined conversion result code of [O:C] is read. When the undetermined conversion result code is read, the editing section 24 searches for an undetermined conversion result code that is continuous with this code and has the same content as this code, and sets the selected flag 43 to the undetermined conversion result code.
to "1".

Then, the editing section 24 adds an underlined attribute to the character code pointed to by the pointer (C) of this undefined conversion result code and provides it to the screen display processing section 2B. with this"
The part "sickness" becomes the homophone selection/confirmation target, and the display becomes as shown in the upper part of FIG. 5. At this time, the cursor is at the "drunk" position (not shown).

When the conversion/next candidate key 14 is pressed in the above state and the input control unit 21 notifies the editing unit 24 of this fact, the editing unit 24 displays the unconfirmed conversion result code [1] corresponding to the cursor position. The conversion result storage buffer 27 is accessed from the pointer C of [1: 1: C] and the pointer (pointer storage section) 55 of the word for which the currently displayed flag 54 is set.
to newly search for the next homophone (in this case, "drunkenness"), turn on the display flag 54 of this word (and turn off the display flag 54 of the previous homophone "drunkenness"), and search for this word. Redisplay the following string. At this time, since the number of characters increases or decreases, the editing work unit 24 uses the conversion result storage buffer 27
Based on the number of characters of the homophone read out from It also processes character string padding and tracking. If the newly selected homophone "drunkenness" is not the desired word, the user presses the conversion/next candidate key 14 again. As a result, in the same way as above, the user's desired word "
``Good'' is searched, the display flag 54 for this word is turned on, and the character string after this word is redisplayed. Now, confirmation of the desired word is performed by pressing the function key I3, or when the function key 13 is pressed while the desired word is displayed, the editing section 24 selects the selected word in the document buffer (51). In place of the undefined conversion result code of the part marked as , enter the character code of the homophone pointed to by the pointer (41) of this code. Then, the screen display processing section 2B is instructed to redisplay this part ("good"). At this time, since there is no attribute specification, the display becomes a normal display, indicating that the kana-kanji conversion result of the reading ``good'' is determined to be ``good''. Screen 32 in FIG. 3 shows the display screen state at this time, and by the above confirming operation, the cursor is moved to the position of the next character ``天J'', and the homonym containing this ``天'' is displayed. "Weather" is underlined.

When the user performs the selection confirmation operation described above, the user places the cursor at the final character position of the homophone that was incorrectly converted due to an error in the reading input, i.e., the word to be corrected (in this case, ``use''). An operation for moving (step S2) is performed. The contents of this operation are notified to the editing section 24 via the input control section 21, and the editing section 24 changes the contents of the cursor position register and also changes the document buffer (51) based on the changed cursor position. access,
Read the code at the relevant position. Assume that an undefined conversion result code of [1:0:A] is now read from the position corresponding to "Yo". When the undefined conversion result code is read out, the editing section 24 searches for an undefined conversion result code that is continuous with this code and has the same content as this code, and changes the selection flag 43 to "1". Editing work department 24
gives an underlined attribute to the character code pointed to by the pointer (A) of this undefined conversion result code to the screen display processing section 26. Now, the "appointment" part becomes the homophone selection/confirmation target, and the display becomes like the screen 33 in FIG. 3. By the above-mentioned movement of the cursor, the homonym "weather", which had been underlined up to that point, is switched to be overlined.

Now, when the function key 13 and the space key 17 are operated for reading correction in the display state of the screen 33 in FIG. The section 24 starts from the pointer of the undefined conversion result code (in this case, "A") in the document buffer (51) specified by the current cursor position, and reads "Kiyo" in the conversion result storage buffer 30. A homophone group is searched and a word consisting of hiragana characters (a character code string corresponding to the reading/here "Kiyou") is read out. Then, the editing unit 24 newly loads the read word (plain name character string) into the input buffer, and also adds an example that indicates that this word is a word (plain name character string) that is subject to reading correction. It is given an attribute of black and white inversion and passed to the screen display processing section 26, and is displayed as shown in the screen 34 of FIG. 3 in place of the previous word ("appointment") (step S5). At the same time, the character string following this word is also redisplayed. At this time, since the number of characters may increase or decrease, the editing work unit 24 creates an unconfirmed conversion result code (here [1: 1: A]) is deleted or inserted, and subsequent character string filling and tracking processing is also performed. Note that the conversion result storage buffer 27 corresponding to the flat name character string (``Kiyou'') read from the conversion result storage buffer 27
There is no change in the status of the homophone group within 7. If another key operation is performed in place of the operation of the function key 13 and space key 17, conventional processing indicated by the contents of the key operation is performed (step S6).

After requesting reading correction by operating the function key 13 and space key 17, if the user wishes to cancel this request, the user operates the cancellation key 16 to cancel the reading correction range (range of kana character strings subject to reading correction). If a change is necessary, operate the left cursor key 11 or right cursor key 12, read the currently highlighted 9 rows of kana characters as they are, and operate the selection/execution key 15 if it is OK to cancel the correction.

If the cancel key 16 is operated and the input control unit 21 detects this (steps S7, S8), the editing unit 24 changes the state of the display screen from the state of the screen 33 to the state of the screen 32 in FIG. That is, step S8. returns to the state before the function key 13 and space key 17 were operated. S
9), the control is returned to the input control unit 21 (the process of step S3). That is, when the cancel key 16 is operated, the editing work unit 24 converts the undefined conversion result code [1: 1:
A] based on the pointer A of the conversion result storage buffer 27
Search for homophones in the list and read out the word (character code) with SORAG 54 on display (in this case, "use")
). Since the selection flag (43) is set on the unconfirmed conversion result code, this character code with an underline attribute is sent to the screen display processing section 26 to be displayed on the screen. At this time, if there is an increase or decrease in the number of characters, the undefined conversion result code (here, [1: 1: A)] is deleted or inserted in the 1-document output (51), and the subsequent character strings are padded, Tracking processing is performed.

Further, when the left cursor key 11 is operated (step 512), the editing work unit 24 shortens the highlighted kana character string to be corrected by one character to the left (step 513).
), deletes the last character of the kana string in the input buffer. Similarly, when the right cursor key 12 is operated (step 514), the editing section 24 extends the highlighted kana character string to be corrected by one character in the right direction (step 515), and extends the text in the document buffer (51). The character code of the corresponding storage position within (that is, the character code of the final character of the new highlighted character string) is loaded into the input buffer.

Further, when the selection/execution key 15 is operated (step S10), the editing section 24 selects the hiragana-kana character string (here, "Kiku") in the input buffer (which matches the highlighted flat name character string). ”) is read out.

Then, the editing unit 24 adds an underlined attribute to the read plate name character string and passes it to the screen display processing unit 26 to display it on the screen (cancels the reverse display), and also changes the contents of the cursor position register to the same value. It is updated to point to the final character position of the character string (step S11).

Now, in this embodiment, the selection/execution key 15 is operated in this state in order to read the kana character string "Kiyou" highlighted on the screen 34 of FIG. 3 and set it as the kana character string to be corrected. . Therefore, the state of the display screen changes from screen 34 to screen 35 in FIG.

After the user operates the selection/execution key 15 to confirm the kana character string to be corrected in reading, if the user wishes to cancel this state, the user operates the cancel key 16 and selects the position of the character to be corrected in the reading indicated by the cursor. If a change is necessary, operate the left cursor key 11 or right cursor key 12, and if a reading correction is necessary, manually input a kana character indicating the correct reading at the character position indicated by the cursor. When the reading correction is completed, the user operates the selection/execution key 15 or the conversion/next candidate key 14.

If the cancel key 1G is operated and the input control unit 21 detects this (step 31B, S17), the editing unit 24 changes the state of the display screen (function key 13 and space key 1
Step 518) returns the state of the screen 32 (immediately before 7 was operated), and returns control to the input control section 21 (the process of step 313). Further, when the left cursor key 11 is operated (step 520), the editing work unit 24 returns the cursor position by one character (toward the left) (step 521),
When the right cursor key 12 is operated (step 522)
, advances the cursor position by one character (to the right) (step 523), and returns control to the input control unit 21 (the process of step S16).

On the other hand, when a kana character is input, the input character is replaced with the character at the storage position in the input buffer corresponding to the current cursor position, and the input character is passed from the input control section 21 to the editing section 24. and the character on which the cursor is displayed is redisplayed as this input character (step 524). At this time, the cursor is advanced by one.

In this embodiment, in order to correct the underlined "yo" in "Kiyo" shown in screen 35 of FIG. 3, the cursor moves to the position of "yo" as shown in screen 3B of FIG. be done. Next, on the screen 3G of FIG. 3, "yo" is overridden at the "yo" position indicated by the cursor, and the one-line reading is corrected as shown on the screen 37 of FIG.

When the user finishes correcting the pronunciation, the user presses the select/execute key 1.
5 or the conversion/next candidate key 14.

The input control unit 21 manually selects the key in step S16.
When it is detected that the execution key 15 or the conversion/next candidate key 14 has been operated (step 519), the editing section 24 is notified of this fact. The editing unit 24 edits the document buffer (
Detect the beginning position of the subsequent character string that was not subject to reading correction in 51), and the character string from this position onwards to the character string immediately before the first occurrence of the punctuation mark or confirmed homophone. (Plate name character string) is obtained and loaded into the input buffer (steps 325 to 527). However, even if the input buffer is full (in this case it is full with N characters), if no punctuation mark or confirmed homophone is found, or if the last character is loaded into the input buffer, the above operation is terminated. Become. In addition, in order to obtain the above-mentioned flat name character string, if the code in the document buffer (51) is an undefined conversion code, it is sufficient to search the conversion result storage buffer 27 using the pointer (41). Needless to say. Here, ``greatness'' is required.

When the editing work unit 24 obtains the above-mentioned flat name character string, it loads this character string in the position subsequent to the flat name character string (“Kiyo”) already stacked in the input buffer, and sends a message to that effect to the input control unit 21. to notify. The input control section 21 includes an editing section 24
Upon receiving the notification from , the plain name character string (``Kiyou wa Subarashidashi'') stored in the input buffer is passed to the kana-kanji converter 22, and the kana-kanji conversion is activated.

The kana-kanji converter 22 converts (re-converts) the kana character string including the reading correction characters passed from the human control unit 21 into kana-kanji (re-converts) in the same manner as in step S1 (step 828). At this time, it goes without saying that the words after the definite word "good" are not subject to reconversion. If we were to establish ``sourei'', it is clear that only ``kyoha'' would be subject to reconversion.

Which range is to be re-converted may be determined by determining whether or not the conversion result is affected by the character string after reading correction. Now, in this example, step 32
As a result of the re-conversion of step 8, the conversion result storage buffer 27 contains the corrected pronunciation of ``kyo'' as shown in FIG.
Homophones of "Suba" are stored. Further, the contents of the cursor position register are changed to the line column position corresponding to the storage position of the first code of the undefined conversion result code string of the same content with the selected flag (43) set in the document buffer (51), From this storage location, as shown in Figure 6, instead of the undefined conversion result code [1: 1: A] corresponding to the homophone of "Kyo", the first word of the homophone group of "Kyo" is An undefined conversion result code [1: 1: D] including a pointer (D) to "today" is written for the number of characters to be displayed (here, two characters). The writing of character codes or undefined conversion result codes to subsequent storage locations is also performed as shown in FIG. Then, according to the contents of the document buffer (51) and the conversion result storage buffer 27, the screen 38 shown in FIG. 3, that is, the kana-kanji conversion result for the correct reading is displayed.

In the above embodiment, the reading correction instruction is given by a combination of the function key 13 and the space key 17, but it is also possible to provide a dedicated reading correction instruction key.

[Effects of the Invention] As detailed above, according to the present invention, when an incorrect kana-kanji conversion result is obtained due to input of an incorrect pronunciation, the incorrect word (homonym) after conversion is selected. By doing this, it is possible to convert only this word into a kana character string that indicates its pronunciation, so by correcting the reading based on this converted kana character string, it is extremely easy to correct only the incorrect characters. can be done. Furthermore, according to the present invention, the kana-kanji conversion is performed again only for the readings of the character string from the kana character string whose reading has been corrected to just before the next punctuation mark or the determined homophone. It is possible to prevent the inconvenience of converting existing homonyms again.

[Brief explanation of the drawing]

Fig. 1 is a functional block configuration diagram showing an embodiment of a document creation device to which the present invention is applied, Fig. 2 is a flowchart showing a reading correction processing procedure, and Fig. 3 is a state transition of a document input screen during reading correction processing. Figure 4 is a diagram showing the format of an undefined conversion result code, and Figures 5 and 6 are diagrams showing the state of the conversion result storage buffer shown in Figure 1 during reading correction processing. FIG. 6 is a diagram showing the state in the document buffer. 10...Keyboard, 11...Left cursor key 12
...Right cursor key, 13...Function key 14...
・Conversion/Next candidate key 15...Cancel key, 17...
- Space key 21... Input control section, 22... Kana-kanji conversion section, 24... Editing section, 25... Display device, 27... Conversion result storage buffer.

Claims (1)

[Scope of Claims] A kana-kanji conversion means that divides the input reading data string into phrases and finds corresponding homophone candidates for each phrase,
In a document creation device that extracts one candidate for each clause as a homophone selection target candidate from among the homophone candidates obtained by this kana-kanji conversion means and displays it on the screen as a character string after kana-kanji conversion for the above-mentioned reading data string. , a pronunciation correction instruction means for instructing to correct the pronunciation of a homophone candidate that is a homophone selection target among the character strings after the above-mentioned kana-kanji conversion; , means for specifying a character string to be read and corrected for converting the homophone candidate to be selected as a homophone into a kana character string, and displaying the same character string as a kana character string to be read and corrected on the screen with a specific display attribute indicating that the character string is read as a kana character string to be corrected; and a pronunciation correction target range change instruction means for instructing that the range of the kana character string to be pronunciation corrected is varied in units of characters, and the display is displayed on the screen with the specific display attribute according to the instruction of the instruction means. A means for changing the range of a kana character string to be corrected; a reading correction means for replacing the kana character string to be corrected, which is displayed on the screen with the above-mentioned specific display attribute, with another kana character string; and replacement by the reading correction means. detecting means for detecting a punctuation mark or a determined homophone that first appears after the kana character string, from the kana-kanji converted character string, and detecting from the kana character string replaced by the reading correction means. A kana-kanji conversion control method characterized in that the kana-kanji conversion is performed again on the pronunciation of the character string immediately before the punctuation mark or homophone detected by the means.