JP3316884B2 - Language translator - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an information processing apparatus using natural language as an input, and more particularly to a language translation apparatus for converting kana-kanji sentences into a programming language such as COBOL.

[0002]

2. Description of the Related Art With the explosive increase in demand for software, there is an increasing demand for lowering software development costs. However, programming languages that have been developed so far take time to learn how to write. Also, it takes time to understand the program, so that debugging and program maintenance are not easy. Therefore,
As a method of solving these problems, there is natural language programming. The conventional techniques include, for example, IPSJ Transactions Vol.25 No.2 (pp.204-215) and IPSJ Transactions Vol.27 No.11 (pp.204-215).
As described in (1112 to 1128), the specification is described in an unambiguous natural language (English), which is analyzed and expanded into a logical expression.
There is a technology to convert to Prolog. The specifications are described hierarchically, and the definitions of the words that appear in the English sentence written at the highest level and the definitions of their meanings are described at lower levels,
As the hierarchy goes down, it becomes a primitive description. A semantic rule is generated from the definition of these words, a logical expression is generated from these semantic rules and semantic rules related to general words (is, every, etc.), and based on the logical expression, is developed into Prolog.

[0003]

In the above-mentioned prior art, it is necessary to define all definitions of words to be used in the specification, and it is necessary to further define words used in the definitions. Also, it does not mention a method for improving the efficiency of specification description in natural language.

It is an object of the present invention to provide (1) a language translator capable of describing a program in natural language.
(2) An object of the present invention is to provide a language translator capable of efficiently generating a program in natural language.

[0005]

SUMMARY OF THE INVENTION The present invention provides a variable dictionary in which a natural language character string or a symbol character string corresponding to data of a variable or the like defined by a user is used as a heading, and a variable dictionary with reference to the data. By providing a language translating apparatus comprising: a generating unit; a reserved word dictionary whose heading is a reserved word in a programming language; and a natural language analyzing unit (morphological analyzing unit) referring to the variable dictionary and the reserved word dictionary.
Solution to Problem (1)

The present invention further comprises means for simultaneously inputting a natural language and analyzing the natural language, notifying an error such as the existence of an unregistered word in real time, and further updating the contents of the variable dictionary. At this time, the above-mentioned problem (2) is solved by providing a language translator capable of reflecting the updated content in the analysis result.

[0007]

[Function] Since the readability is improved by describing the programming in a natural language, the work efficiency of program development, debugging, maintenance, handover, etc. is improved. Variable dictionaries, which are indispensable for analyzing natural words (word division and translation), are automatically generated from user definition information, so users do not have to create dictionaries for natural language programming at all. Help me. Further, by making the format of the variable dictionary the same as the format of the reserved word dictionary provided in advance, the processing efficiency of morphological analysis and the dictionary maintenance work are improved.

In addition, natural language analysis is sequentially performed for each language unit specified by the user, and the presence of unregistered words is immediately displayed as an error. As a natural language input error, and the efficiency of program development is improved.

Furthermore, when definition information such as variables is changed in the course of program editing and the contents of the variable dictionary are updated, the updated contents of the variable dictionary are compared with the results of analysis of the part for which natural language analysis has been completed up to that point. By reflecting the results and automatically correcting the analysis result each time, when the variable dictionary is updated on the way, it is possible to efficiently correct the natural language corresponding thereto.

[0010] Further, since words and phrases used by the user can be defined instead of headings registered in the reserved word dictionary or the variable dictionary, the efficiency of program creation and maintenance are improved.

[0011]

DESCRIPTION OF THE PREFERRED EMBODIMENTS The present invention will be described below based on an embodiment in which the present invention is applied to a program development support system.

The present invention is not intended to be applied to only a specific programming language, but can be applied to various programming languages. In this embodiment, the present invention is applied to a program development support system for a programming language COBOL. A description will be given of the case in which this is done. The input of the system described in this embodiment is a sentence (sentence) in which the contents of a program to be described in COBOL are described in kana-kanji.

FIG. 1 is a block diagram showing the outline of the configuration of the system described in this embodiment. This system is roughly composed of two modules. One is a module (10, 11) for generating a variable dictionary 12, and the other is a module for analyzing a kana-kanji program sentence 2 and performing COBOL.
Modules (3, 4, 6,
8).

A module for generating the variable dictionary 12 stores and holds a definition editor 9 for a user to define specifications relating to screens, files, forms, database schemas, and the like, and contents defined by the definition editor 9. A user definition information storage table 10 and a variable dictionary generation processing unit 11 that generates a variable dictionary 12 with reference to the information in the table.
Consists of

On the other hand, Kana Kanji sentence 2 is analyzed and COBOL
The modules to be converted into the program 14 include an editing editor 1 for creating the kana-kanji program 2, an input sentence check processing unit 3 which is a pre-processing for morphological analysis of the created kana-kanji program 2, and a kana-kanji sentence 2 Is divided into words and word information is obtained from the dictionaries (12, 13). A word information storage table 5 that stores and holds the word information obtained by the morphological analysis 4. The kana kanji sentence 2 is composed of a kana conversion processing unit 6 for converting half-width kana sentences into an half-width kana sentence, which is an intermediate expression, and a COBOL conversion processing unit 8 for converting half-width kana sentences to a COBOL program 14.

Further, an error information storage table 7 storing message information for supporting a user's work is prepared as output information when an error occurs in each of the above-described processes.

In this embodiment, the morphological analysis processing unit 4
As for, the morphological analysis module used in the machine translation or the interactive system can be used as it is. The reason is that the morphological analysis processing unit 4 used in the present embodiment can perform word division with an accuracy higher than an appropriate level and can extract information on the divided words from the dictionary. Therefore, the morphological analysis module and the dictionary structure adopted in this embodiment are the IPSJ 4th
A publicly-known general-purpose Japanese morphological analysis tool is adopted as described in the 4th National Convention Lecture Paper 3-181. Therefore, in the following description of the morphological analysis processing 4, the internal processing of the general-purpose Japanese morphological analysis tool will not be described, and only the input and output thereof will be described.

Also, the kana sentence is converted into a COBOL program 14
The COBOL conversion processing unit 8 that converts the program into a program is a program development support system EAGLE2 (Effective Appointment) developed by Hitachi, Ltd. for efficient program development.
proach to Achieving HighLevel Software Productivit
y 2, manual # 6180-3-822, 823, 8
24). EAGLE2 is equipped with a precompiler for converting a program described in half-width kana sentences into a COBOL program 14. In the present embodiment, this precompiler is employed for the COBOL conversion processing unit 8. Therefore, from the Kana sentence, COBO
The processing method for converting to the L program 14 will not be described. Hereinafter, a processing method for converting a kana-kanji sentence to a kana sentence and a system configuration for realizing the same will be described.

FIG. 2 is a flowchart showing the outline of the flow of the processing of this embodiment. First, the user selects a mode to create a kana-kanji program or define variables and the like (step 21). When the mode for creating a kana-kanji program is selected, a program editing editor is started, and the user inputs and edits the kana-kanji program (step 1). The program editor is
It can be a full screen editor, a line editor, or any other type of editor. The input means of the Kana-Kanji sentence is a key input from a keyboard. However, as another method, a voice input, a handwritten character input, or the like can be easily substituted.

In this embodiment, in order to improve the conversion processing speed, the natural language analysis processing (3, 4, 6) is executed in parallel with the operation of the user inputting the kana-kanji program. In the present embodiment, whether or not the user inputs a program and presses a line feed key is triggered (step 22), and the kana-kanji sentence of that part is analyzed and converted into an half-width kana sentence as an intermediate expression. These can be easily realized by executing them in a multi-process. The determination method other than determining whether a trigger for analysis is performed based on whether the line feed key is pressed depends on whether the analysis is directly instructed by the user (for example, whether the “analysis execution” switch is turned on). A determination method, a determination method using a punctuation mark as a trigger, or a designation of the number of analysis unit lines can be easily replaced.

When the execution of analysis is instructed by a trigger,
First, it is checked whether the input kana kanji sentence is an appropriate sentence (step 3). In the present embodiment, by performing double-byte character conversion, character number limit check, and the like, it is checked whether the following processing can be executed. For improper sentences (step 23), an error message is notified (step 27) to urge the user to make corrections.

Next, in a morphological analysis process, the kana-kanji character string is divided into words, word information is extracted from the variable dictionary 12 and the reserved word dictionary 13, and stored in the word information storage table 5.
Hold. If an error occurs during morphological analysis,
An error message is displayed (step 27) to urge the user to make a correction.

As a result of the morphological analysis, the words are converted into kana for the kana kanji sentence that has been word-divided (step 6). First, the result of the word division in the morphological analysis processing is corrected for a part that is not preferable for the present system. Next, referring to the kana information of each word,
Convert to Kana and output the conversion result to a file. At this time, the dictionary (12, 1)
If there is a word (unknown word) not registered in 3), an error message is displayed (step 27), and the user is urged to confirm and correct.

When a variable definition is selected in the mode selection, an editor for the variable definition is provided, and the user is required to input necessary information (step 9). In the present embodiment, when referring to variables, there are names of screens and forms, names of their items, schema information of database, global variables, and the like, and an editor for defining all of them is necessary.

The variable definition information defined by the user is stored in the user definition information storage table 10.
The variable dictionary 12 is generated with reference to the information in this table (step 11). The structure and format of the variable dictionary 12 are the same as those of the reserved word dictionary 13 storing reserved words and system variable information, and have a structure suitable for the above-mentioned general-purpose Japanese morphological analysis tool. The dictionaries (12, 13) described in the following description are assumed to extract necessary minimum information from the information held in these dictionaries and combine them into one table. FIG. 11 shows a configuration of the variable dictionary 12 having the minimum necessary information, and FIG. 13 shows a configuration of the reserved word dictionary 13.

As is clear from FIGS. 1 and 2, the input sentence check processing unit 3, morphological analysis processing unit 4, kana conversion processing unit 6, COBOL conversion processing unit 8, and variable dictionary generation processing unit 11 represent processing. , Kana-Kanji program 2, Word information storage table 5, Error information storage table 7, User definition information storage table 10, Variable dictionary 12, Reserved word dictionary 1
3. The COBOL program 14 is a file (table)
Is shown.

Since each functional block is constituted by program logic, an LSI is provided for each functional block.
It is possible to increase the processing speed as a device that supports the creation of a program using kana-kanji characters.

FIG. 3 is a block diagram of the hardware configuration of the present embodiment. Reference numeral 40 denotes an input / output device, which receives an input from a user via an editing editor and a definition editor,
Further, the device is a device for transmitting a system message to a user by an error message or the like. Reference numeral 45 denotes a processor which performs arithmetic processing based on a program stored in the storage device 50 while exchanging data with the storage device 50 and the input / output device 40. Reference numeral 50 denotes a storage device, which includes the following storage areas. 51 is a kana kanji program storage area, 52 is a word information table storage area, 53 is a user definition information storage area, 54 is an input sentence check processing unit storage area, 55 is a morphological analysis processing unit storage area, 56 is a kana conversion processing unit Storage area, 5
7 is a COBOL conversion processing unit storage area, 58 is a variable dictionary generation processing unit storage area, 59 is a COBOL program storage area, 60 is an error information storage area, 61 is a reserved word dictionary storage area, 62 is a variable dictionary storage area, and 63 is a variable dictionary storage area. Each represents a working area.

FIG. 4 is an explanatory diagram showing an example of input of the kana-kanji program 2. Hereinafter, this program will be described as a specific example. In FIG. 4, the character string "stock screen"
Represents the name of the screen, and is a word to be defined by the user. The data relating to this word is stored in the variable dictionary 12 by the variable dictionary generation processing 11 described below.
“#” On the first and seventh lines is a symbol representing a label, and is accompanied by a numeric string immediately after. In the second line,
The data input to the screen is read, the end condition is described in the third line, the numerical calculation is performed in the fourth line, and the process skips to # 20 in the fifth line. In the eighth line, writing is performed on the screen, and the process ends in the ninth line. As described above, since the program contents can be described in Kana-Kanji and can be used as a program or as a specification as it is, development, maintenance, and handover of the program can be easily performed.

FIG. 5 is an explanatory view showing a program after converting the kana-kanji program of FIG. 4 into kana. Each sentence is divided into words, each word is separated by a half-width space, and each word is described in a half-width kana expression. When such a program in kana notation is sent to the above-mentioned COBOL precompiler, a final COBOL program can be automatically generated.

FIG. 6 is a PAD (Problem Analysis Diagram) showing an outline of the process of the variable dictionary generation process 11. In this process, first, the variable dictionary 12 is initialized (step 111), and then the screen name acquisition process (step 1)
12), data item name acquisition processing (step 113),
By the file name acquisition processing (step 114) and the like,
Variables are collected from various types of definition information, and necessary information is stored in the variable dictionary 12. Here, a description will be given of a process of storing variables relating to a screen name representing a screen name, a data item name which is a name of an item described in a table or the like, and a file name representing a file name in the variable dictionary 12. For other types of variables, the variable information can be stored in the variable dictionary 12 in exactly the same manner as described above, and thus they will not be described in detail here.
As a result of the variable dictionary generation process 11, a variable dictionary 12 as shown in FIG. 11 is generated.

FIGS. 7 to 9 show a variable dictionary generation process 1.
7 is a process for acquiring variable information, FIG. 7 is a screen name acquisition process (step 112), FIG. 8 is a data item name acquisition process (step 113), and FIG. 9 is a file name acquisition process (step 113). It is each PAD figure showing step 114). All the data referred to in these processes are stored in the user definition information storage table 10 shown in FIG. This table stores and holds variable information defined by the user using the variable definition editor. When the user defines variables using the editor,
It is generated by automatically storing such information as shown in FIG. As shown in FIG. 10, the components of the user definition information storage table 10 include a type (5001) for identifying from which definition editor the variable information is acquired, a heading (5002) representing the variable name, and a key corresponding to the heading. Information (5003), classification indicating the type of variable (5004)
Etc. Of course, other information (such as option designation information) that is essential when executing the program is also stored in this table. However, in the following processing in the present embodiment, such information is not necessary, so I will not touch it.

Hereinafter, the screen name acquisition processing of FIG.
Description will be made with reference to the drawings. First, a heading 5002 of a portion representing a screen name is extracted from the user definition information storage table 10 with reference to the type 5001 (step 1120).
1). It is determined with reference to the reserved word dictionary 13 whether or not the heading 5002 overlaps with the heading of the reserved word dictionary 13 (step 11202). , And outputs an error message (step 11203). For example, if the error code here is defined as -5, FIG.
Refer to the error information storage table 7 shown in the table below, and the error message “Registration error:“ the relevant word ”corresponding to the error code -5 This word is a reserved word and cannot be used as a user-defined word. Please correct to another word. 12. Here, in FIG. 12, the portion surrounded by the symbol “＠” is a variable portion, and indicates that the character string is to be replaced with an appropriate character string depending on the situation. , An error code corresponding to the content of the error can be appropriately selected.In step 11202, if there is no overlap, the heading 5002 is stored in the heading column of the variable dictionary 12. (Step 11204) Next, “noun” is stored in the part of speech field of the variable dictionary 12 (Step 11205), which means that all variables are treated as nouns. This can be treated as a noun when viewed as a component of a program sentence.
The kana information 5003 corresponding to the heading is extracted from 0 and stored in the kana information column of the variable dictionary 12 (step 1120).
6). Then, the character string "screen name" is stored in the word classification column of the variable dictionary 12 (step 11207).

The processing procedure is the same for the data item name acquisition processing (step 113) of FIG. 8 and the file name acquisition processing (step 114) of FIG. The character string "data item" is stored (step 11307), and the file name is different only in the processing part that stores the character string "file name" (step 11407).

FIG. 10 shows a user definition information storage table 1.
FIG. 4 is an explanatory diagram showing a configuration of a zero. As described above, this table has a type 5001, a heading 5002, and kana information.
5003, classification 5004, etc.

FIG. 11 shows a variable dictionary 1 generated as a result of the variable dictionary generation processing 11 for the program of FIG.
FIG. 2 is an explanatory diagram showing a configuration of FIG. The variable dictionary 12 includes a heading 5011, a part of speech 5012, kana information 5013, and a word classification 5014. The screen name “stock screen” is obtained by screen name acquisition processing (step 112), and the data items “amount”, “unit price”, “unit price” "Quantity" is a variable acquired by the data item name acquisition process (step 113).

FIG. 12 is an explanatory diagram showing the configuration of the error information storage table. As described above, the error code 5021 and the error message 5022 are stored in a one-to-one correspondence. The portion surrounded by the symbol "@" indicates a variable portion, and indicates that the character string is replaced with an appropriate character string according to the situation.

FIG. 13 is an explanatory diagram showing the configuration of the reserved word dictionary 13. As described above, the reserved word dictionary 13 has the same configuration as the variable dictionary 12, and these configurations are based on the specifications of a general-purpose Japanese morphological analysis tool used in morphological analysis. In FIGS. 11 and 13, only essential information is described. As shown in FIG. 13, the reserved word dictionary 13 includes a heading 5031, a part of speech 5032, kana information 5033, and a word classification 5034. The heading 5031 of the reserved word dictionary 13 can be roughly classified into reserved words and system variables. The reserved words and system variables referred to here are both reserved words provided for converting kana-kanji sentences to kana sentences and reserved words provided for converting kana sentences to COBOL. Including.

FIG. 14 shows the PAD of the input sentence check process 3.
This process will be described below with reference to a PAD diagram.

It is determined whether a half-width alphabet, a special symbol, or a blank character exists in an input sentence to be analyzed (step 301).
The characters are converted to full-width characters (step 302). If not, it is determined whether the number of input characters is 250 characters or more (step 303). If the number of input characters is 250 characters or more, an error message indicating the number of characters is exceeded is output with reference to the error information storage table 7. (Step 30
4).

FIG. 15 is a diagram showing a configuration of the word information storage table 5 generated as a result of the morphological analysis processing 4.
As described above, the present embodiment employs a general-purpose Japanese morphological analysis tool. This tool takes a full-width Japanese character string as input and outputs a word table (not the word information storage table 5) representing the result of word division. The word table stores word information such as headings, parts of speech, inflection information, and codes representing the meaning of attached words. In this case, only word information necessary for subsequent processing is extracted from the word table, The word information storage table 5 is generated. As shown in FIG. 15, the word information storage table 5 includes a heading 5041, a part of speech 5042, kana information 5043, and a word classification 5044.

FIG. 16 is a PAD diagram of the kana conversion process 6. Hereinafter, this process will be described with reference to the PAD diagram.

First, the following processing is performed for each word (step 601). It is determined whether or not the part of speech 5042 is a “numerical string” (step 602). If the part of speech 5042 is a numeric string, the numeric string (full-width) is converted into a half-width numeric string, and “kana information 5043” in the word information storage table is converted. (Step 603). Since the numeric string cannot be registered in the variable dictionary 12 or the reserved word dictionary 13, the kana information (5013,513)
033) is not held. Therefore, this processing is required.

Next, the following processing is performed for each word (step 604). It is determined whether or not the heading 5041 is a label symbol "#" (step 605). If the heading 5041 is a label symbol "#", it is determined whether or not the word immediately after that is a digit string (step 606). If it is a numeric string, the label symbol "#" and the numeric string are combined into one word (step 607). At this time, the kana information 5043 corresponding to the newly created word is a kana character string obtained by combining kana information for each word.

Next, the following processing is performed for each word (step 608). It is determined whether or not the part of speech 5042 is “unknown word”, that is, whether the word is not registered in the reserved word dictionary 13 or the variable dictionary 12 (step 6).
09), if the word is "unknown word", an error message indicating that it cannot be converted to kana is output with reference to the error information storage table 7 (step 610). If the word is not an unknown word, kana information corresponding to the word is written to an output file (step 611).

FIG. 17 is a diagram showing an example of the kana-kanji program editing editor. The editor mainly comprises an operation menu display section 6001, an editing area 6002, a message display section 6003, and a kana-kanji conversion section 6004.
The operation menu display section 6001 further includes a variable definition icon 6011 to be selected when switching to variable definition, an edit icon 6012 to be selected when switching to kana kanji program editing, and selection when performing conversion when the user likes. A conversion icon 6013, a text processing icon 6014 selected when processing a text by calling another file, a save icon 6015 selected when saving an edited program to a file, and a selection when the user wants to refer to guidance, etc. Help icon 6016,
End icon 6017 to be selected when processing is to be ended
Consists of FIG. 17 shows that the kana kanji program shown in FIG. 4 is being input.

FIG. 18 is an example of a screen indicating that the natural language analysis processing has been completed. As described above, in this embodiment, the natural language analysis is performed in parallel with the input of the natural language, using the line feed symbol as a trigger. In FIG. 18, the part where the natural language analysis has been completed is displayed in Gothic, so that the fact that the analysis has been completed and the analysis has been normally performed is displayed by making a display different from the others. Of course, in addition to the Gothic display, the display may be distinguished by color-coded display or underlined display.

FIG. 19 is a diagram showing error information output when an analysis error is found in natural language analysis. Here, suppose that in the kana kanji sentence of FIG. 4, when the character string “end” existing in the third line is input, the character string “completed” is input due to a conversion error. The character string “completion” is an unknown word that is not registered in the reserved word dictionary 13 or the variable dictionary 12. When the line feed key is pressed, this line is analyzed. As a result of the morphological analysis processing 4, “unknown word” is stored as the part of speech of this word.
When this unknown word is detected in the kana conversion processing 6, an error message is output in real time to the message display section 6003 of FIG. 19 with reference to the error information storage table 7. At the same time, the part of the character string "completion" is displayed in a different manner from the others. As a display method, there are a method of inverting black and white, a method of blinking, a method of displaying in a different color (character color, background color) according to the type of the error message, and the like. In this case, any method can be used. By referring to such message information, the user can
Lexical errors can be easily found.

FIG. 20 is a diagram showing another example of the input of the kana-kanji program 2. The first line in FIG. 20 indicates that the local variable “work 1” is defined in the program. The letter "V" at the beginning of the sentence is a label indicating that the sentence is a variable definition sentence. In this embodiment, an area for defining such a special character is provided at the beginning of the sentence. Further, in the statement, a variable name, its kana notation, and its type and size are defined. In the natural language analysis of this sentence,
First, it is determined whether or not "V" is described at the beginning of the sentence.
011 and the next character string in parentheses is Kana information 501.
Acquire as 3. The information is stored in the variable dictionary 12 and held. Here, the parts of speech 5012 of these variables are all nouns, and the word classification 5013 is a local variable.

On the other hand, when the local variable "work 1" in FIG. 20 is stored in the variable dictionary 12, "work 1" has been analyzed as an unregistered word in the sentences for which natural language analysis has been completed. A search is performed to determine whether or not the word information is present, and the word information now stored in the variable dictionary is added as word information of “work 1” analyzed as an unregistered word. Then, what is displayed on the editor as an unregistered word is switched to display as a word that is not an unregistered word. This processing is performed not only when the local variables are stored in the variable dictionary 12 but also when the user registers new data with the definition editor 9 or updates existing data. Also,
Contrary to the above, even when a word that was not an unregistered word becomes an unregistered word by updating the variable dictionary, a character string search is performed based on the information before update, thereby updating the variable dictionary 12. Character strings that have become unregistered words.

According to the present embodiment, the Kana-Kanji sentence is analyzed with reference to the variable dictionary 12 and the reserved word dictionary 13 and converted to Kana sentence, and further converted from Kana sentence to COBOL, so that the user has excellent readability. It is possible to generate a program by using the kana-kanji sentence. Also, by linking the update of the dictionary contents with the program editing, the update information of the variables can be reflected in the program editing in real time.

As a modified example or extended example of this embodiment, for a reserved word or a defined variable name, a user-familiar word or phrase is registered as a synonym so that it can be described in a program. Can be As shown in FIG. 21, a definition screen is provided for registering a reserved word or a synonym for a defined variable name. In FIG. 21, the character string "read" is registered as a synonym for the reserved word "read". The character string “read” is registered as a heading of the variable dictionary 12, and word information is stored in the variable dictionary 12 by referring to the definition information in FIG. 21 or the word information of “read”. As a result, even if the sentence “Read inventory screen” in FIG. 4 is described as “Read inventory screen”, as a translation result, “Zaiko Gamen O Yom
. ”Can be obtained.

[0053]

According to the present invention, the readability of the program is improved, and the work efficiency of coding, debugging, testing, maintenance management, takeover, etc. of the program is improved. Also,
Since a variable dictionary is automatically generated from data defined by the user, dictionary creation costs can be reduced. In addition, since the format of the variable dictionary and the format of the dictionary of reserved words specific to the programming language have the same format structure, maintenance of the dictionary is facilitated and the processing efficiency of morphological analysis is improved. In addition, since the user analyzes the input natural language while editing the program, an error message such as the presence of an unregistered word can be reported as an error at the time of inputting the natural language. Furthermore, when the variable dictionary is updated, the program can be efficiently corrected by reflecting the updated content in the program edited so far.

[Brief description of the drawings]

FIG. 1 is a block diagram of one embodiment of the present invention.

FIG. 2 is a flowchart of a process according to an embodiment of the present invention.

FIG. 3 is a block diagram of hardware according to the embodiment of the present invention.

FIG. 4 is an explanatory diagram showing input of a kana-kanji program.

FIG. 5 is an explanatory diagram showing a program after conversion into kana;

FIG. 6 is a PAD showing an outline of a variable dictionary generation process;
FIG.

FIG. 7 is a PAD diagram of a screen name acquisition process.

FIG. 8 is a PAD diagram of a data item name acquisition process.

FIG. 9 is a PAD diagram of a file name acquisition process.

FIG. 10 is an explanatory diagram showing a configuration of a user definition information storage table.

FIG. 11 is an explanatory diagram showing a configuration of a variable dictionary.

FIG. 12 is an explanatory diagram showing a configuration of an error information storage table.

FIG. 13 is an explanatory diagram showing a configuration of a reserved word dictionary.

FIG. 14 is a PAD diagram of an input sentence check process.

FIG. 15 is an explanatory diagram showing a configuration of a word information storage table.

FIG. 16 is a PAD diagram of kana conversion processing.

FIG. 17 is an explanatory diagram showing a kana-kanji program editing editor.

FIG. 18 is an explanatory diagram showing the end of natural language analysis.

FIG. 19 is an explanatory diagram showing output of error information.

FIG. 20 is an explanatory view showing another example of the input of the kana-kanji program.

FIG. 21 is an explanatory view showing a synonym registration screen.

[Explanation of symbols]

DESCRIPTION OF SYMBOLS 1: Kana-kanji program editing editor, 2: Kana-kanji program, 3: Input sentence check processing unit, 4: Morphological analysis processing unit, 5: Word information storage table, 6: Kana conversion processing unit, 7: Error information storage table, 8 COBOL conversion processing unit, 9 variable definition editor, 10 user definition information storage table, 11 variable dictionary generation processing unit, 12 variable dictionary, 13 reserved word dictionary, 14 COBOL program.

Continuation of the front page (72) Inventor Takao Kawamura 5030 Totsukacho, Totsuka-ku, Yokohama-shi, Kanagawa Pref. Software Development Division, Hitachi, Ltd. (56) References JP-A-62-267829 (JP, A) JP-A-3 JP-A-271828 (JP, A) JP-A-62-282329 (JP, A) JP-A-59-125447 (JP, A) JP-A-2-129756 (JP, A) (58) Fields investigated (Int. . ^7, DB name) G06F 9/06 - 9/45

Claims (1)

(57) [Claims] And 1. A means for inputting a natural language program, the editor for editing the natural language program the input, the definition editor to define the data such as variables or constants used in the program, the definition editor Means for generating a variable dictionary to which a character string or a symbol character string of the natural language corresponding to the data defined in the above or the data defined in the program as a heading and to which semantic information is added, a reserved word in a programming language of the program A natural word analyzing means for analyzing the edited natural language based on a reserved word dictionary to which semantic information is added as a heading, and the variable dictionary; and programming the natural language based on the analysis result of the natural language A language translation device comprising: translation means for translating into a language; and output means for outputting the translation result.