JP3004847B2 - Source program generation method and system - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a source program generation method and system for generating a target source program from sentences in which functional specifications of a program are described in an itemized manner.

[0002]

2. Description of the Related Art Conventionally, when a statement for generating a source program is described in a specification, a statement associated one-to-one with one instruction per sentence is registered in a statement information file, and There is a system that generates a source program by describing the registered statement in a book and replacing it with a corresponding instruction word.

FIG. 56 is an explanatory diagram showing a conventional example of creating a specification of a source program using the tree structure editor 100. The tree structure editor 100 describes specifications by combining various symbols S, R, I, and the like representing the structure of a program. In each symbol, a statement registered in the statement information file 110 is described. be able to.

[0004] For example, when the expression "-is and-is" is input as a sentence, the sentence information file 110 contains "-is and-is" in advance. "And its corresponding" ({1 = {@
2) A post-instruction replacement rule composed of the instruction word 112 “AND ($ 3 = $ 4)” is registered as a pair. Therefore,
First, when a symbol 101 for which a sentence is to be input is selected, a list of registered sentences is displayed. From the list, a sentence 111 to be input is selected (“-is and -is”). Then, the selected sentence 111 is input into the symbol 101.

Next, when "-" is selected to input an arbitrary term in "-", a list of data items 120 is displayed.
1, "blank" 122, "name" 123, "blank" 122
Select each. Then, the term selected in "~" is input.

According to the above procedure, the sentence "Date is blank and name is blank" 102 can be input.

[0007] A source program 130 as shown in the figure is generated based on such specifications. The source program 130 includes a source 131 of a tree structure portion generated according to a predetermined source generation rule of the tree structure editor 100 and a source replaced by a command replacement rule 112 of a statement registered in the statement information file 110. 132.

[0008]

However, the above-described method of generating a source program has the following problems.

(1) Since a sentence described in the specification is completed by one sentence, when a complicated conditional expression or the like is described, the expression of the sentence becomes long and its meaning is difficult to understand. Also,
Since it is necessary to register sentences according to the combination of the conditional expressions described, the number thereof becomes enormous.

(2) Since only a simple character string is replaced from the sentence described in the specification, if the specification to be described includes an expression including control of a program, the specification cannot be described.

An object of the present invention is to provide a source program generation method and a system capable of easily writing a functional specification of a program including complicated conditions and generating a target source program even for a person having no programming knowledge.
Is to provide a system.

Further, the specification including a control program easily described, is to provide a source program generation method and system capable of generating a source program of interest.

[0013]

According to the present invention, in order to achieve the above object, basically, a user setting portion is left blank.
Statement and the type of program control structure that corresponds to the statement
Multiple sets of texts with one set of separate data and source generation rules
Sentence information file that stores
Input device for inputting descriptive data of program function specifications
And a location, play the source program corresponding to the stored program specifications data to program specifications file
A source program generation method in a source program generation system for forming, reads pre Kibun information plurality of sets statements information stored in the file, the contents required by the user from the input device to the blank portion of each sentence information A first processing step of inputting, and a second step of storing, in the program specification file, a combination of a plurality of sentences in which contents required by the user are input in blanks as description data of a program function specification in a bulleted text format. And a third processing step of reading out the description data of the program function specification stored in the program specification file by the second processing step, and creating a statement information node representing each statement in the data; and The program control structure is determined from the type data of the program control structure of each statement, and each statement information Fifth creating a fourth processing step of creating a directed graph representing a control relationship between over de, the source program corresponding to the program function specification description data based on the directed graph and each sentence of the source generation rules
And a processing step of

The program specification file
Source program corresponding to the specified program specification data
A source program generation system for generating a system,
Corresponds to the sentence where the user setting part is blank and the sentence
Type control data and source generation rules
Sentence information file that stores a plurality of sets of sentence information with one set of rules
And the sentence information stored in this sentence information file.
Multiple sentences with user-required contents in blanks
A bullet-type sentence-type program machine configured in combination
The description data of the functional specification is stored in the source program generation system.
Received from the input device attached to the
To store the data of
And the profile stored in the program specification file.
Read the data of the program specification description, and
Creates a sentence information node that represents a sentence, and each sentence has
From the type data of the program control structure,
Directed group representing the control relationship between each sentence information node
A directed graph creating means for creating a rough, and the created directed
Based on the graph and the source generation rules for each statement, the program
Generates a source program corresponding to the data in the system specification description
And a source generation unit that performs the operation.

[0015]

According to the above means , the statement information file is referred to,
Functional specification including one or more processing conditions and processing contents
Is an itemized sentence (in this example, Japanese itemized sentence)
Input device attached to the described source program generation system
Compatible with the program specifications by inputting from the device
Is created, and based on this directed graph,
Generates source program corresponding to the program specifications
Is done. For this reason, when creating a new program
Lists the functional specifications of the newly developed program
Newly developed program in this bulleted sentence format
What is necessary is just to input the data of the function specification.

[0016] Enter the functional specifications of the program in a bulleted list.
By that enables the force, including a complex condition program
Easy to use for functional specification of system
And generate the desired source program
Can be.

Further, the sentence information file is stored in the sentence information file.
Each statement is used to generate a source program from that statement.
Has types indicating source generation rules and program control structures
Because of this, the control structure of the program is also
Is described at the same time by the description of
There is no need to describe the control structure by another means. Accordingly
Even people without programming knowledge can control programs
Enter functional specifications without considering the structure
Source programs can be generated.

Further, each of the sentences constituting the itemized sentence has
From the type data of the program control structure
Create a directed graph by determining the system control structure
And the source program has the directed graph and each sentence
Sources for each statement are generated according to the source generation rules.
Generate a program and control it by a directed graph
It is only necessary to combine according to the structure, and it is a complicated control structure.
However, the generation process itself is simplified.

[0019]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Hereinafter, the present invention will be described in detail with reference to the illustrated embodiments.

FIGS. 1 to 3 are explanatory diagrams showing an embodiment of a processing method when a source program is generated by describing a function in an itemized text in a column of a tabular specification 220. FIG.

The tabular specification 220 shown in FIG. 1 includes, for a certain data item, a condition portion 221 for describing a condition for checking the item, and a processing portion 222 for describing a process to be performed when the condition is satisfied. Is done.

With respect to such a specification 220, a function specification 21 "set a code in the account number when the condition that the name is blank and the date is a number is satisfied".
Assuming that 0 has been raised, the condition part 221 contains “all of the following.”, “(1) Name is blank.”
A conditional sentence 223 "(2) Date is a number" is described, and a processing sentence 224 "Set a code in the account number" is described in the processing part 222 in an itemized manner.

FIG. 2 is a flowchart corresponding to the itemized text of the tabular specification 220. When the itemized text is represented by a flowchart, the condition part 231 becomes like the condition flowchart 310, and the processing part 232 becomes the processing flowchart 320. become that way.

Therefore, when a source program is generated by a method described later, a source program 430 as shown in FIG. 3 is generated. This source program 430 is
It comprises a condition source program 410 corresponding to the condition flowchart 310 and a processing source program 420 corresponding to the processing flowchart 320.

That is, the condition source program 410 and the processing source program 420 generated from the itemized sentences described in the individual columns are combined according to the unique rules of the tabular specification 220, and the completed source program 4
30.

In this example, a completed source program 430 is generated by merging the processing source program 420 with the portion described as "$ process" in FIG.

In this case, the combined flowchart is shown as a combining flowchart 330 shown in FIG.

FIGS. 4 to 6 are explanatory diagrams showing an embodiment in which a function is described in an itemized text in a processing column of a tabular specification and a source program is generated.

The table format specification 520 shown in FIG. 4 includes, for a certain data item, a condition portion 521 for describing a condition for checking the item, and a processing portion 522 for describing a process to be performed when the condition is satisfied. Is done.

For such a tabular specification 520,
Functional specification 51 "set code to account number"
If 0 is raised, a conditional statement 523 of "unconditional" is described in the conditional part 521, and a processing statement 52 of "set a code in the account number" is written in the processing part 522.
4 is described.

FIG. 5 is a flowchart corresponding to the itemized text of the tabular specification 520. When the itemized text is represented by a flowchart, the condition part 531 becomes a condition flowchart 610, and the processing part 532 becomes a process flowchart 620. become.

Therefore, when a source program is generated using a method described later, a source program 730 as shown in FIG. 6 is generated. This source program 730 is
It comprises a condition source program 710 corresponding to the condition flowchart 610 and a processing source program 720 corresponding to the processing flowchart 620.

That is, the source programs 710 and 720 generated from the itemized sentences described in the individual columns are combined according to the unique rules of the table specification 520 to generate a completed source program 730.

In this example, a completed source program 730 is generated by merging the source program 720 of the processing part with the part described as “$ processing”.

The flowchart after the combination is represented as a synthesis flowchart 630 as shown in FIG.

FIGS. 7 to 9 are explanatory diagrams showing an embodiment in which a function is described in the column of check conditions and check contents of the tabular specification 820 in the form of itemized sentences and a source program is generated.

The tabular specification 820 shown in FIG. 7 includes a check condition portion 821 for describing a condition for checking an item with respect to a certain data item, and a check content portion 822 for describing a check to be performed when the condition is satisfied. And an error code part 82 set when the condition is not satisfied
And 3.

For such a tabular specification 820,
Functional specification "If the condition that the name is blank and the date is a number is satisfied, check whether the account number is a number, and if the condition is not satisfied , set the error code to 9999." When 810 is raised, the check condition part 521 includes check condition statements 824 such as “all of the following.”, “(1) Name is blank”, and “(2) Date is a number”. Is described in a check content portion 822, and a check content sentence 826 "Account number is a number" is described, and an error code portion 827 "9999" is described in an error code portion 823.

FIG. 8 is a flowchart corresponding to the itemized text in the tabular specification 820.
1 is as shown in a check condition flowchart 910,
The check content portion 832 is as shown in a check content flowchart 920.

Therefore, when a source program is generated using a method described later, a source program 1040 as shown in FIG. 9 is generated. This source program 1
040 is a check condition source program 1010 corresponding to the check condition portion 831 and a check content portion 8
Check source program 1020 corresponding to 32
Consisting of Error code part 823 is tabular specification 8
It is generated as an error code source program 1030 according to 20 unique rules.

That is, the source programs 1010, 102 generated from the itemized sentences described in the individual columns
0, 1030 are combined according to the unique rules of the tabular specification, and are generated as a completed source program 1040.

In this example, the source program 1020 of the check content portion 832
A complete source program 1040 is generated by merging the source program 1030 of the error code with the portion described as “$ error code”.

In this case, the flowchart after the combination is represented as a synthesis flowchart 940 as shown in FIG.

FIGS. 10 to 12 show the tree structure format specification 11.
FIG. 10 is an explanatory diagram showing an embodiment in which a function is described in itemized sentences in 20 symbols and a source program is generated.

The tree structure format specification 1120 shown in FIG. 10 describes the contents to be processed in a conditional symbol 1121 and a processing symbol 1122 representing the structure of the program, and generates a program by combining the respective symbols. .

In response to such a specification 1120, a function specification 111 "set a code to the account number when the condition that the name is blank and the date is a number is satisfied"
When 0 is raised, “all of the following” is included in the conditional symbol 1121 of the tree structure format specification 1120.
An itemized sentence 1123 of “(1) Name is blank” and “(2) Date is a number” is described, and an itemized sentence of “set a code in account number” in the processing symbol 1122 is described. 1124 is described.

FIG. 11 is a flowchart corresponding to the itemized text of the tree structure specification 1120. When the itemized text is represented by a flowchart, a conditional symbol flowchart 1210 and a processing symbol partial flowchart 12 are shown.
It looks like 20.

Therefore, when a source program is generated using a method described later, a source program 1340 corresponding to the itemized sentences 1131 and 1132 of the tree structure format specification 1120 as shown in FIG. 12 is generated. This source program 1340 includes a condition source program 1310 corresponding to the condition symbol 1121 and a processing symbol 11
22 and a tree structure generation rule source program 1 indicating a tree structure generation rule
330.

That is, the source programs 1310 and 1320 generated from the itemized sentences 1131 and 1132 described in the individual symbols are combined by the tree structure generation rule source program 1330 to generate a completed source program 1340. In this case, the combined flowchart is represented as a combined flowchart 1240 as shown in FIG.

In this example, the combination of symbols makes the structure of the program the structure shown in the tree structure partial flowchart 1230 shown in FIG. 11, and the combination of programs uses the tree structure generation rule source program 1330 generated according to this structure. Generated.

FIGS. 13 to 16 are explanatory diagrams showing an embodiment in which a function is described in a conditional section and a processing section of the decision table format specification 1420 in the form of a paragraph and a source program is generated.

FIG. 13 Decision Table Format Specification 1
Reference numeral 420 describes a condition to be checked in the condition part 1421 and contents to be processed in the processing part 1422. If the condition is satisfied in the matrix 1423 of the condition part 1421, "Y" is described, and if not, "N" is described. . In the matrix 1424 of the processing unit 1422, “X” is described in a process performed when the condition described in the matrix 1423 of the condition unit 1421 is satisfied.

[0053] For such a specification 1420, feature called "name is blank, and the date is the condition of there with a number to set the code to the account number Once you have established" specification 14
And 10, and the functional specification 1411 that "name is blank, and the date is set 9999 to the error code in place <br/> if not a number", "name is blank, and a date numbers, And the error code is set to 9999 if the account number is not a number, the decision table format specification 1420 is issued.
In the condition part 1421, “all of the following” is indicated.
A paragraph sentence 1425 of “(1) Name is blank”, “(2) Date is a number” and a paragraph sentence 1426 of “Account number is a number” are described in the processing unit 1422. Describes an itemized sentence 1427 "Set a code in the account number." And an itemized sentence 1428 "Set the error code to 9999." The matrix 142 of the condition part 1421 is described according to the function specification.
If the condition 3 is satisfied, “Y” is described, otherwise, “N” is described.
In “4”, “X” is described in the processing performed when the condition described in the matrix 1423 of the condition part is satisfied.

The itemized sentence described in the specification 1420 is described separately in “condition 1” 1431 and “condition 2” 1432, and “process 1” 1433 and “process 2” 1434.

FIG. 14 shows a decision table format specification 1.
It is a flowchart corresponding to the article 420 of a paragraph,
When the itemized text is represented by a flowchart, “condition 1” 1431 is represented by “condition 1 flowchart” 1510, and “condition 2” 1432 is represented by “condition 2 flowchart”.
1520, “Process 1” 1433 is “Process 1 flowchart” 1530, and “Process 2” 1434 is “Process 2”.
This is represented by a flowchart 1540.

Therefore, when a source program is generated using a method described later, a source program 1660 as shown in FIG. 16 is generated. This source program 16
A “condition 1 source program” 1610 corresponding to “condition 1” 1431, a “condition 2 source program” 1620 corresponding to “condition 2” 1432, and a “processing 1”
“Process 1 source program” 163 corresponding to 1433
0 and a “processing 2 source program” 1640 corresponding to the “processing 2” 1434.

That is, the source program 161 generated from the itemized text described in each condition part and processing part
0, 1620, 1630, and 1640 are generated as a completed source program 1660 by being combined according to the unique rules of the decision table format specification 1420. The completed source program 1660 is represented by a flowchart shown in FIG.

In this example, the condition part matrix 142
3 and the items checked in the matrix 1424 of the processing unit, the structure of the program becomes the structure shown in the decision table flowchart 1550 of FIG. 14E. According to this structure, the decision table source program 1650 of FIG. 16 is generated. . Then, according to the source generation rules of the decision table source program 1650, each source program 1610, 1620, 1
630 and 1640 are combined to form a synthesized source program 1660.

In the present invention, a source program is generated from the itemized text described in the various specifications as described above. Next, a system for generating a source program from the itemized text described in the specification is described. Will be described.

FIG. 17 is a block diagram showing one embodiment of the source program generation system of the present invention.

In FIG. 17, reference numeral 2000 denotes a source program generation system from a paragraph document, 2010 denotes a specification file, 2020 denotes a sentence information file, 2030 denotes a term information file, and 2040 denotes a source program file.

FIG. 18 shows the information stored in these files in more detail.

The statement information file 2020 is a file in which statements used in the specification are registered in advance, and the statement used in the specification, the control attribute of the program of the statement, and the source program from the statement are stored. This file stores a plurality of sets of source program generation rules for generation as one set .

The sentence stored in the sentence information file 2020 is not limited to a sentence that is completed only by the sentence, but a sentence that “is a is” is registered and a user setting contents field is registered.
It is also possible to register a sentence in which an arbitrary term or character can be described in "~".

The term information file 2030 is a file in which terms described in "-" of the sentence registered in the sentence information file 2020 are registered. The term used in the sentence and the data name corresponding to the term are registered. Are stored in pairs.

The specification file 2010 is a file for storing information describing the specification of the program.
Information unique to the specification, for example, information other than the format of the table of the specification or the itemized sentences, and the sentence information file 2020.
And information on the itemized text described using the sentence and the term registered in the term information file 2030.

The itemized sentence holds the information on the order, step, and terminology described in the itemized sentence in units of the specification document. In this example, the itemized sentences described in the specification are stored in units of table columns such as the condition 2101 and the process 2102.

The source program file 2040 is generated based on the sentence information file 2020 and the term information file 2030 from the paragraph text described in the specification file 2010. A source program 2040 in which programs are combined according to a unique procedure of a table specification is stored.

Source program generation system 2000
Is a specification data reading module 2 that reads a paragraph sentence from the specification file 2010 and expands it into a memory.
001, a specification data development area 2004 for developing the paragraph text read by the specification data reading module 2001 in the memory, and a sentence information file 202.
Sentence information reading module 2002 for reading sentence information from 0 and expanding it to memory, sentence information data expansion area 2005 for expanding sentence information read by sentence information data reading module 2002 to memory, and term information from term information file 2030 And a term data reading module 2003 for reading the term information and developing the term information in a memory, and a term information data development area 2006 for developing the term information read by the term data reading module 2003 into the memory.

Further, the specification data development area 2004
It analyzes the contents 2016 of bullets developed in, specifications data analysis module 2007 for instructing the creation of the directed graph representing the control program, sentence information data expand message information 2014 or term information 2015 in accordance with an instruction of each module area 2005 or the term information sentence-term data retrieval module 2008 to be retrieved from the data expansion area 2006, a directed graph generation module 2009 to create a given directed graph in accordance with the instructions of the specification data analysis module 2007, a directed graph generation module 20
Directed graph deployment area 201 for storing information directed graph <br/> off representing a control program generated by the 09
0, information has been created in a directed graph deployment area 2010 2
017 based on, bulleted text source generation module 2012 according to the procedure of execution directed graph instructs the source deployment individual statements, bullets sentences source generation module 2
The source generation rule analysis module 2011 and the source generation rule analysis module 201 that analyze the source generation rule of each sentence designated by the instruction 012 and instruct the source to be expanded.
The source expansion module 2013 expands the contents of a sentence as a source program in accordance with the instruction 1. This
Of the source program generation system 2000, a memory
Including general-purpose computers, displays, keyboards, etc.
Sentence information file.
Item list using a plurality of sets of sentence information stored in 2020
Input data of program function specifications consisting of text
In this case, a plurality of sets of sentence information are displayed on the display,
The necessary sentence information from among them can be
And let the user select it, and set the user for the selected sentence information.
Input necessary information in the column "~" by keyboard operation etc.
It is made to let.

FIG. 18 shows the relationship between the sentence information file 2020, the term information file 2030, the specification file 2010, the specification 2100, and the source program 2040.

[0072] In the above system arrangement will be described below types and bullets has been directed graphs and flow charts obtained from the description of the statements for performing source generated from the text that are bullets.

As shown in FIGS. 19 to 25, the sentence described in the specification is the sentence information 220 as sentence information data.
1, a control type 2202 of the program, and a source generation rule 2203 for generating a source from a sentence.

In the present embodiment, the control type of the program of the statement is a condition,
It is classified into four types: expression, processing, and modification. In addition, elements can be added as needed.

1. FIG. 19 is an example of sentence information of “conditional AND”, and FIG. 20 is an example of sentence information of “conditional OR”. The condition is used for a sentence describing a specification for making a judgment in a sentence. Write it together with expressions and statements with qualification attributes. For example, sentence descriptions are 2210 and 2310, and corresponding flowcharts are 2230 and 2330.

2. Expression FIG. 21 is an example of statement information of “logical AND expression”, and FIG. 22 is an example of statement information of “logical sum expression”. These expressions are used for a sentence describing an expression in a sentence, and are modified. Write together with the statement that has the attribute. For example, sentence descriptions are 2410 and 2510, and the corresponding flowcharts are 2430 and 2510.
It will be 30.

3. Processing FIG. 23 is an example of sentence information of “processing”. “Processing” is used for a sentence describing a specification for performing processing in a sentence. For example, the description of the sentence is 2710, and the corresponding flowchart is as follows. 2730.

4. Modification FIGS. 24 and 25 show examples of sentence information of “modification”, and “modification”
Is used for a sentence describing a specification for making a judgment in a sentence, and when describing a complicated condition, it is used together with a condition and an expression. For example, the description of a sentence is 2810, 2910
Next, the flowchart corresponding to 2830,2
930.

In the present invention, by analyzing such a sentence, a directed graph representing program control is created by dividing the sentence information into the following three types of sentence information nodes, and then a source program is created.

1. Condition node Holds a sentence information node describing one condition. The condition node is a logical product (THEN clause) and a logical sum (ELSE
Section). 2240 directed graph 19 representing a logical product, 2340 directed graph 20 representing the logical OR
Indicated by

Hereinafter, condition nodes are represented by square symbols with rounded corners, logical products are represented by horizontal lines, and logical sums are represented by vertical lines.

[0082] Incidentally, the directed graph for the modified statement 2
4, denoted by reference numerals 2840 and 2940 in FIG.

2. Expression node Holds a statement information node of an expression described in a condition or the like. 2440 directed graph 21 expression node representing the. Logical product having information of the logical product (AND and logical OR (OR),
Directed graph representing a logical sum is represented by 2540 of FIG. 22. Hereinafter, the expression nodes are represented by elliptical symbols, and the logical product and logical sum are represented by lines.

3. Processing node Holds a sentence information node of a description of a single sentence such as processing. The processing node has information when a sentence is listed. 27 of directed graphs 23 representing the cases listed
Indicated at 40. Hereinafter, processing nodes are represented by square symbols.

[0085] Figure 26 shows an actual example of creating flowcharts and directed graph created based on the text was performed itemization in Figure 27.

FIG. 26 is a flowchart 3010 created based on a sentence 3000 that simply describes continuous processing.
It is indicative of the, if you create a directed graph 3020 based on this, "set a calendar date to date" 3001, because the statement of "To set the current to account" 3002 corresponds to the statement with the processing attribute, Each is created as a processing node 3021 of "set calendar date on date" and a processing node 3022 of "set current account", and "set calendar date on date".
The processing node 3022 processing nodes 3021 and in "Loading checking the account" is a directed graph that are connected by vertical lines 3023 to correspond to the cases listed.

FIG. 27 shows a sentence 330 describing a complex condition.
FIG. 33 shows a flowchart 3310 created based on 0. In the flowchart, the condition in which the processing after “(2) all of the following” 3305 is described is step 3 in the flowchart.
311 and 3312 appear in multiple places.

[0088] When you create a directed graph 3320 based on this, "(1) either of which is below" 3302 is waiting for the condition attribute, "is either (1) below" 3302
Does not make sense in the sentence alone, and the sentence having the qualification attribute of “(1) date is a calendar date” 3303 and “(2) date is blank” Based on the description, the condition node 33 of “(1) Date is a calendar date”
21, condition node 33 of "(2) date is blank"
22 is generated.

The line between the condition node 3321 of “(1) Date is a calendar date” and the condition node 3322 of “(2) Date is blank” has a parent sentence of “(1). Any of the following "3302, and is connected as a vertical line 3323 because the sentence attribute corresponds to an attribute having a logical sum.

“(2) Either of the following” 3306
Similarly, the sentence alone does not make sense, and is described in a stepped manner, "(1) Account type is normal."
7 and the expression node 3324 of “(1) Account type is normal.” Based on the description of the statement having the qualifying attribute of “(2) Account type is immediate” 3308, and “(2) Account type is Is generated for the moment. "

The line between the expression node 3324 of “(1) Account type is ordinary” and the expression node 3325 of “(2) Account type is immediate” has a parent sentence of “(2). Either of the following: "3306, which is connected as line 3326 because the attribute of the sentence corresponds to an attribute having a logical sum.

"(2) All of the following" 3305 is described in a stepped manner in "(2) Any of the following."
3306 and a condition node 3327 of “(2) Either of the following.” And a condition node of “(2) Account number is a number” based on the description of “(2) Account number is a number” 3309 3328 is connected.

The line between the conditional node 3327 of “(2) Either of the following.” And the conditional node 3328 of “(2) Account number is a number” has a parent sentence of “(2) 3305, which is connected as a horizontal line 3329 because the attribute of the sentence corresponds to an attribute having a logical product.

"(1) Any of the following" 330
2 and “(2) All of the following.” 3305 indicate that the parent sentence is “All of the following.” 3301, and the attribute of the sentence is an attribute having a logical sum, so that a horizontal line 3330,
Directed graph connected by 3331 is generated.

FIG. 28 is a flowchart showing the entire process of generating a source program from an itemized sentence.

First, the specification data reading module 2
001, the specification data composed of itemized text from the specification file 2010 is generated by the source generation system 2000.
And the specification data is developed in the specification data development area 2004 (step 3401). This place
If the specification data stored in the specification file 2010
Is, as described above, the source program generation system 20.
Sentence information file by keyboard operation attached to 00
2020 multiple sets of sentence information stored in the display
Display, and the necessary sentence information
The user is allowed to make a selection by work, etc., and the selected sentence information is
The contents required for the user setting column "~" are used for keyboard operation etc.
Therefore, it is stored by inputting.

Next, the sentence information data reading module 2
002, the sentence information data is input from the sentence information file 2020 into the source generation system 2000, and the sentence information data is developed in the sentence information data development area 2005 (step 3402).

Next, by using the term information data reading module 2003, the term information data is input from the term information file 2030 into the source generation system 2000,
It is developed in the term information data development area 2006 (step 3403).

Next, the sentence / term data retrieval module 2
008, the program control type is acquired from the contents of the specification data development area 2004 and the sentence information data development area 2005, the itemized text is analyzed using the specification data analysis module 2007, and the directed graph is further analyzed. use <br/> full creation module 2009 creates a directed graph representing the control program, directed graph expansion area 20
10 (step 3404). Finally, the contents of a directed graph that has been deployed in directed grayed <br/> rough deployment area 2010 sequentially won by bullets sentence source generation module 2012, on the basis of the acquired content, sentence-term data retrieval module 2008 Is obtained from the sentence information data development area 2005 and the term information data development area 2006 using the necessary source generation rules, and the obtained source generation rules are obtained from the source generation rule analysis module 2
The source program is analyzed using the source generation and development module 2013 based on the analysis result, and is output to the source file 2040 (step 3405).

According to the above sequence, a source program is generated from an itemized sentence.

Next, step 3404 and step 340
5 will be described in detail.

[0103] In fact Yu from bullets sentence 3300 of FIG. 27
Flowcharts for generating a directed graph are shown in FIGS.
As shown in FIG. Also shows a process of creating flow and directed graph 3503 of data when read line by line a statement 3501 deployed in the specification data development area 2004 in FIGS. 32 to 43.

The flowchart for itemized sentence analysis is as follows. The parent sentence information 3501 is input as input data, and the head address 350 of the analyzed child sentence information node is output as output data.
Returns 2. Also, in the data flows of FIGS.
And statements that have been deployed in the specification data development area 2004, Yes
Shows the contents of a directed graph <br/> full deployed in directed graph information expansion area 2010.

First, processing is performed from “all of the following” 3301.

First, "NULL" is set to the head address of the sentence information node of the child to be processed (step 3503). At the same time, “NULL” is set to the previous address of the child sentence information node to perform initialization (step 3504).

Next, one sentence record of a child of the itemized sentence is read from the specification data development area 2004. Then, as shown in FIG. 32A, the first "all of the following" 3301 is read (step 3505).

Next, it is determined whether there is a sentence of the same level. Since the sentence "All of the following" 3301 has already been read, the sentence becomes "there is a sentence of the same level", and the processing of the sentence reading loop is continued (step 3).
506).

Next, it is determined whether or not there is child sentence information.
Then, since “(1) one of the following.” 3302 exists, “Y” is established by “child sentence information = present”.
That (step 3540).

Here, “all of the following” 3301
Is temporarily suspended, and “all of the following” 33
The statement information of No. 01 is stacked, and “all of the following” 3
To call recursively the bullet sentence analysis of the 301 sentence information as input data (step 3523). At this time, the input data 3501 "are all below." 3301, and the contents of a directed graph deployment area 2010 32
As shown in (b), “logical condition (product)” 3820 is stacked as sentence information.

Next, "(1) Any of the following"
Although it is in a state of processing 3302, by recursively calling the itemized sentence analysis processing of step 3523,
The process is newly started from the top of the flowchart. At this time, the input data is "all of the following".
The sentence information 301 is obtained as input data.

First, "NULL" is set to the head address 3502 of the sentence information node of the child to be processed (step 3503). At the same time, “NULL” is set to the previous address of the child sentence information node to perform initialization (step 3504).

Next, one record of a child sentence record of “all of the following” 3301 is read from the specification data development area 2004. Then, the topmost item "(1) Any of the following" 3302 is read (step 3505).

Next, it is determined whether there is a sentence of the same level. Already, “(1) Any of the following” 33
Since the sentence No. 02 has been read, the sentence is “there is a sentence of the same level”, and the processing of the sentence reading loop is continued (step 3506).

Next, it is determined whether or not there is child sentence information.
Then, since “(1) date is a calendar date” 3303 exists, “Y” is established by “child sentence information = present” (step 3540).

Here, the processing of “(1) Any of the following” 3302 is temporarily suspended, and the sentence information of “(1) Any of the following” 3302 is stacked.
Further, using the sentence information of “(1) Any of the following.” 3302 as input data, the itemized sentence analysis processing is performed again.
It is called recursively (step 3523).

[0117] At this time, the input data 3501 "is either (1) below." 3302, and the Yes
The contents of the directed graph development area 2010 are in a state as shown in FIG. 33A, and "logical condition (sum)" is used as sentence information.
3920 is stacked.

Next, "(1) Date is a calendar date" 33
03 is processed, but the process is newly started from the top of the flowchart by recursively calling the itemized sentence analysis process again. At this time, as input data, “(1) Any of the following” 3302
Get sentence information.

First, “NULL” is set to the head address 3502 of the sentence information node of the child to be processed (step 3503). At the same time, “NULL” is set to the previous address of the child sentence information node to perform initialization (step 3504).

Next, one record of a child sentence record of “(1) Any of the following” 3302 is read from the specification data development area 2004. Then, the first “(1) date is a calendar date” 3303 is read (step 3505).

Next, it is determined whether there is a sentence of the same level. Already, “(1) Date is a calendar day” 3303
Sentence is read, so there is a sentence of the same level
And the processing of the statement reading loop is continued (step 3506).

Next, it is determined whether or not there is child sentence information.
Then, because there is no child sentence information, "child sentence information =
"N" is established for "Yes" (step 3540).

If there is no child sentence information, a sentence information node creation process is performed (step 3507).

FIG. 30 shows details of the sentence information node creation processing. First, the control type of the program is obtained from the sentence record of "(1) Date is a calendar date" 3303 (step 3507). ).

Next, the control type of the acquired program is determined (steps 3609 to 3612). Then
Since “(1) Date is a calendar day” 3303 corresponds to the decoration attribute, “Y” of “program control type = modification”
Is satisfied, and the decoration attribute is processed (step 3).
612).

Next, "(1) Date is a calendar date" 33
03 and the parent “(1) Any of the following.”
Program control type of 3302 have the (condition (logical OR))
Passing the directed graph generation module 2009, that instructs the create sentence information node (step 3629).

Next, the sentence information node 3321 of "(1) Date is a calendar date" is set in the sentence information node current pointer.
Is set (step 3630).

[0128] Figure 31 is shows the details of the directed graph creation process, Here, first, sentence information node address before the child is determined whether "NULL". Then, since the child sentence information node preceding address value is not set, the child sentence information node preceding address = NULL, and the condition of “Y” is satisfied (step 3713).

Therefore, the sentence information node 3321 of “(1) Date is a calendar date” is added to the head address of the child sentence information node.
Sets, store text information nodes start address of the child (step 3731).

The sentence information node 3321 of “(1) Date is a calendar date” for which sentence analysis has been completed is set to the child sentence information node previous address (step 3719).

[0131] At this time, a directed graph expansion area 201
The content of 0 is as shown in FIG. 34B, and "(1) date is a calendar date" is created as sentence information node 3321 of the logical condition (sum) as sentence information.

Next, “(2) Date is blank” 33
04.

Next, from the specification data development area 2004, one record is read from the statement record next to “(1) Date is a calendar date” 3303. Then, “(2) Date is blank” 3304 is read (step 35).
20).

Next, it is determined whether there is a sentence of the same level, and if there is a sentence, the processing of the sentence reading loop is continued (step 3521).

Next, it is determined whether or not there is child sentence information (step 3540). Then, it holds the "N" for there is no statement information of the child in the "statement information of the child = yes" (step 3540).

[0136] Therefore, it reads "(2) The date is blank." To gain control type of a statement of the program from the statement record of 3304 (step 3607).

Next, the control type of the acquired program is determined. Then, “(2) date is blank” 330
4 corresponds to the decoration attribute, the condition of “Y” of “program control type = modification” is satisfied, and the processing of the decoration attribute is performed (step 3612).

Next, "(2) Date is blank" 33
04 sentence information and the parent "(1) One of the following."
Program control type of 3302 have the (condition (logical OR))
Passing the directed graph generation module 2009, that instructs the create sentence information node (step 3629).

Next, the sentence information node 3322 of "(2) Date is blank" is set as the sentence information node current pointer.
Is set (step 3630).

Next, the address before the child sentence information node is NU.
When the LL is determined, the child sentence information node previous address value is "(1) Date is a calendar day."
Since 21 is set, the child sentence information node previous address is ≠ NULL, and the condition of “N” is satisfied (step 3713).

Next, when the type of the parent sentence information is determined, since “(1) Any of the following” 3302 corresponds to the condition attribute, the condition of “program control type = condition” is satisfied (step 3717). Further, when the program control type is determined in detail, a condition (logical sum) is established (step 3736).

Next, a line of a condition (logical sum) is placed between the sentence information node 3321 of “(1) Date is a calendar date” and the sentence information node 3322 of “(2) Date is blank”. 332
3 is set (step 3737).

Next, the sentence analysis of the child sentence information node previous address is completed. “(2) Date is blank” 3304
Is set (step 3719).

[0144] At this time, a directed graph expansion area 201
The content of 0 is in the state of FIG. 35B, and the sentence information node 332 of the condition (logical sum) is “(2) Date is blank.
2 and a line of the condition (ethical sum) between the sentence information node 3321 of “(1) Date is a calendar date” and the sentence information node 3322 of “(2) Date is blank”. 33
23 is set.

Next, from the specification data development area 2004, one record is read from the statement record next to “(2) Date is blank” 3304 (step 3520).

Next, when it is determined whether or not there is a sentence, since there is no corresponding sentence, the process exits the sentence reading loop (step 3521).

Next, the sentence information address of "(1) Date is a calendar day" 3303 stored in the head address of the sentence information node of the child is returned (step 3522).

With the above procedure, the procedure returns to the caller of the recursive call, and the processing of “(1) Any of the following” 3302 temporarily held is continued.

The processing is started immediately after recursively calling the itemized sentence analysis (step 3523).

Next, the control type of the program is obtained from the sentence record of the read “(1) Any of the following” 3302 (step 3607).

Next, the control type of the acquired program is determined. "(1) Any of the following" 3302
Corresponds to the condition attribute, so "program control type =
The condition of "Y" of the condition is satisfied, and the processing of the condition attribute is performed (step 3609).

As the current pointer of the sentence information node, the sentence information node address 3321 of "(1) Date is a calendar day" is set as the return value when recursively calling the itemized sentence analysis (step 3626). ).

Next, the address before the child sentence information node is NU.
When the LL is determined, the child statement information node previous address is not set, so the child statement information node previous address =
NULL and the condition of "Y" is satisfied (step 37)
13).

Accordingly, the sentence information node 3321 of “(1) Date is a calendar date” is set as the head address of the child sentence information node.
Is set and stored as the child sentence information node start address (step 3731).

The sentence information node 3321 of “(1) Date is a calendar date” for which sentence analysis has been completed is set to the child sentence information node previous address (step 3719).

[0156] In this case, a directed graph deployment area 2010
36 becomes the state shown in FIG. 36 (b), and by returning to the caller, the stack of the statement information returns by one, and becomes the "logical condition (product)".

Next, "(2) All of the following" 330
5 is processed.

Accordingly, the specification data development area 2004
Therefore, when one record is read from the sentence record following “(1) Any of the following” 3302, “(2) All of the following” 3305 is read (step 352).
0).

Next, it is determined whether there is a sentence of the same level, and if there is a sentence, the processing of the sentence reading loop is continued (step 3506).

[0160] Next, when the determination of the presence or absence of the statement information of the child, "(1) is any of the following." Since 3302 there, it established "Y" in there = statement information of the child (Step 3540).

Here, “(2) All of the following” 33
05 is temporarily suspended, the sentence information of “(2) All of the following” 3305 is stacked, and the sentence information of “(2) All of the following” 3305 is input data 3501 and the itemized sentence analysis is performed. Is called recursively (step 352).
3).

At this time, the input data 3501 becomes “(2) All of the following” 3305, the contents of the directed graph development area 2010 become the state shown in FIG. A “logical condition (product)” 4320 is stacked as information.

Next, "(1) Either of the following" 3
306 is processed.

The process is newly started from the head of the flowchart by recursively calling the itemized sentence analysis. At this time, the sentence information of “(2) All of the following” 3305 is obtained as input data.

First, NULL is set to the head address of the sentence information node of the child to be processed (step 3503). At the same time, NULL is set to the previous address of the child sentence information node to perform initialization (step 35).
04).

Next, from the specification data development area 2004, when one statement record of the child record of “(2) All of the following” 3305 is read, the first one is “(1) one of the following. . 3306 is read (step 3505).

Next, it is determined whether or not there is a sentence of the same level.
Since the sentence 06 has been read, it is determined that "there is a sentence of the same level", and the processing of the sentence reading loop is continued (step 3506).

Next, it is determined whether there is a sentence of the same level. Then, “(1) Account type is normal.” 330
In order to 7 is present, the "Y" established in the "statement information of the child = yes" (step 3540).

Here, "(1) Any of the following"
The process of 3306 is temporarily suspended, and the statement information of “(1) One of the following.” 3306 is stacked, and “(1)
One of the following: Using the sentence information of "3306" as input data 3501, the itemized sentence analysis is recursively called again (step 3523).

At this time, the input data 3501 becomes “(1) One of the following.” 3306, and the contents of the directed graph development area 2010 are as shown in FIG. “Logical expression (sum)” 4420 is stacked as sentence information.

Next, "(1) Account type is normal."
A state of processing 3307 is entered.

When the itemized sentence analysis is recursively called, the process is newly started from the top of the flowchart. At this time, the sentence information of “(1) Any of the following” 3306 is obtained as input data.

First, NULL is set to the head address of the sentence information node of the child to be processed (step 3503). At the same time, NULL is set to the previous address of the child sentence information node to perform initialization (step 35).
04).

Next, from the specification data development area 2004, when one statement record of a child of “(1) Any of the following.” 3306 is read, the first “(1)
Account type is normal. 3307 is read (step 3505).

Next, it is determined whether or not there is a sentence of the same level. As already described, “(1) Account type is normal” 3
Since the statement 307 has been read, "statement of the same level exists" and the processing of the statement reading loop is continued (step 3506).

Next, it is determined whether there is a sentence of the same level. Then, since there is no child sentence information, “N” is established for “child sentence information = present” (step 3540).

Next, the control type of the program is obtained from the sentence record of the read “(1) Account type is normal” 3307 (step 3607).

Next, when the control type of the acquired program is determined, “(1) Account type is normal” 3307
Corresponds to the qualification attribute, so "program control type =
The condition of “Y” of “modification” is satisfied, and processing of the decoration attribute is performed (step 3612).

Next, "(1) Account type is normal."
The statement information of 3307 and the parent program "(1) One of the following." 3306, the program control type (expression (logical sum))
The passing the directed graph generation module 2009, and instructs the create sentence information node (step 3629).

Next, the sentence information node 33 of "(1) Account type is ordinary" is set as the sentence information node current pointer.
24 are set (step 3630).

Next, the child statement information node previous address is NU
When it is determined whether the LL is the LL, the previous sentence information node address value is not set, and thus the previous sentence information node address of the child is NU.
LL, and the condition of “Y” is satisfied (step 371).
3).

In the head address of the child sentence information node, "(1)
Account type is normal. Is set and stored as a child sentence information node start address (step 3731).

The sentence information node 3324 of “(1) Account type is ordinary” for which sentence analysis has been completed is set to the child sentence information node preceding address (step 3719).

[0184] At this time, a directed graph expansion area 201
The content of 0 is in the state shown in FIG. 39B, and “(1) Account type is normal” is created as sentence information node 3324 of the formula.

Next, "(2) Account type is immediate."
3308 is processed.

Next, from the specification data development area 2004, when one record is read from the statement record next to “(1) Account type is normal.” 3307, “(2) Account type is immediate.” 3308 is read (step 3
520).

Next, it is determined whether or not there is a sentence of the same level. If there is a sentence, the processing of the sentence reading loop is continued (step 3506).

Next, it is determined whether there is a sentence of the same level. Then, since there is no child sentence information, “N” is established for “child sentence information = present” (step 3540).

Next, the control type of the sentence program is acquired from the sentence record of the read “(2) Account type is immediate” 3308 (step 3607).

Next, when the control type of the acquired program is determined, "(2) Account type is immediate" 3308
Corresponds to the qualification attribute, so "program control type =
Since the condition of “Y” of “modification” is satisfied, processing of the modification attribute is performed (step 3612).

Next, "(2) Account type is immediate."
Program control type (expression (logical sum)) of the statement information 3308 and the parent “(1) One of the following.” 3306
The passing the directed graph generation module 2009, and instructs the create sentence information node (step 3629).

Next, the sentence information node 33 of "(2) Account type is current" is set as the sentence information node current pointer.
24 are set (step 3630). Next, when it is determined whether or not the child sentence information node front address is NULL, the sentence information node 3324 of “(1) Account type is normal” is set in the child sentence information node front address. Therefore, child sentence information node previous address $ NU
LL, and the condition of “N” is satisfied (step 371).
3).

Next, when the type of the parent sentence information node is determined, “(1) Either of the following” 3306 corresponds to the expression attribute, so that the condition of program control type = expression is satisfied (step 3718). ).

Next, "(1) Account type is normal."
An expression line 3326 is provided between the sentence information node 3324 and the sentence information node of the sentence information node 3325 of “(2) Account type is immediate” (step 3734).

[0195] At this time, a directed graph expansion area 201
0 is in the state shown in FIG. 40 (b), and the sentence information “(2) Account type is immediate” is created as the sentence information node 3325 of the logical expression (sum), and “(1) Account An expression line 3326 is provided between the sentence information node 3324 and the sentence information node 3325 of “(2) Account type is immediate.”

Next, the sentence analysis for the child sentence information node previous address is completed. “(2) Account type is immediate.” 33
08 sentence information node is set (step 371).
9).

Next, from the specification data development area 2004, one record is read from the statement record next to “(2) Account type is immediate” 3308 (step 352).
0).

Next, when it is determined whether there is a sentence, since there is no corresponding sentence, the process exits the sentence reading loop (step 3521).

Next, the address 3324 of the "(1) account type is normal." Sentence information node stored at the head address of the child sentence information node is returned (step 352).
2).

With the above procedure, the process returns to the caller of the recursive call, and the processing of “(1) Any of the following” 3306, which has been temporarily suspended, is continued.

Next, the control type of the program is acquired from the sentence record of the read “(1) Any of the following” 3306 (step 3607). When the control type of the acquired program is determined, “(1) any of the following” 3306 corresponds to the expression attribute, so the condition of “Y” of “program control type = expression” is satisfied. , Processing of the expression attribute (step 3610).

Next, "(1) Either of the following" 3
The statement information 306 and the parent “(2) All of the following” 33
05 program control type (Condition (logical product)) passing the directed grayed <br/> rough creation module 2009, and instructs the creation sentence information node (step 3626).

Next, the statement information node 332 of "(1) Any of the following" is set in the statement information node current pointer.
The line of the formula 7 is drawn (step 3627).

Next, the sentence information node 332 of "(1) Any of the following" is added to the sentence information node current address.
7 is set (step 3628).

Next, the address before the child sentence information node is NU.
When the LL is determined, the child statement information node previous address is not set, so the child statement information node previous address =
NULL, and the condition of “Y” is satisfied (step 3
713).

The sentence information node 3327 of “(1) Any of the following” is set to the child sentence information head address,
It is stored as the child sentence information head address (step 37).
31).

The sentence information node 3327 of “(1) Either of the following” is set to the child sentence information node previous address (step 3519).

[0208] In this case, a directed graph deployment area 2010
41B is in the state shown in FIG. 41B, a statement information node 3327 of “(1) Any one of the following.” Is created, and the stack returns to the caller by one. The return is "logical (product)".

Next, "(2) Account number is a number."
3309 is processed.

Next, from the specification data development area 2004, when one record is read from the next sentence record of “(1) Either of the following.” 3306, “(2) Account number is a number.” 3309 is read (step 35)
20).

Next, it is determined whether there is a sentence of the same level, and if there is a sentence, the processing of the sentence reading loop is continued (step 3506).

Next, when it is determined whether or not there is child sentence information, since there is no child sentence information, "N" is established by "child sentence information = present" (step 3540).

Next, the control type of the sentence program is acquired from the sentence record of the read “(2) Account number is a number” 3309 (step 3607).

Next, the control type of the acquired program is determined. Then, “(2) Account number is a number.” 3
Since 309 corresponds to the decoration attribute, the condition of “Y” of “program control type = modification” is satisfied, and the processing of the decoration attribute is performed (step 3612).

Next, "(2) Account number is a number."
The sentence information of 3309 and the parent “(2) All of the following” 3
The program control type (condition (logical product)) 305 is passed to the directed graph creation module 2009 to instruct creation of a sentence information node (step 3629).

Next, the sentence information node 33 of "(2) Account number is a number" is set as the sentence information node current pointer.
The address 28 is set (step 3630).

Next, the address before the child sentence information node is NU.
When the LL is determined, the previous sentence information node address value of the child is “(1) one of the following.”
Since 327 has been set, the child sentence information node previous address = NULL, and the condition of “N” is satisfied (step 3713).

Next, when the type of the parent sentence information node is determined, since “(2) all of the following” corresponds to the condition attribute, the program control type = condition “Y” is satisfied (step 3717). ).

When the program control type is determined in detail, the condition (logical product) is satisfied (step 373).
4).

Next, a condition (logical product) is placed between the sentence information node 3327 of “(1) Either of the following.” And the sentence information node 3328 of “(2) Account number is a number.” (Step 3735).

Next, the sentence information node 3328 of “(2) Account number is a number” whose sentence analysis has been completed is set to the child sentence information node previous address (step 371).
9).

[0222] In this case, a directed graph deployment area 2010
42 (b), "(2) Account number is a number" is created as sentence information node 3328 of the condition (logical product) as sentence information, and "(1) Next A line 3329 of a condition (logical product) is provided between the sentence information node 3327 of "either." And the sentence information node 3328 of "(2) Account number is a number."

Next, from the specification data development area 2004, one record is read from the statement record next to “(2) Account number is a number” 3309 (step 352).
0).

Next, when it is determined whether there is a sentence, since there is no corresponding sentence, the process exits the sentence reading loop (step 3521).

Next, the address of the sentence information node 3327 of "(1) Any of the following" stored in the head address of the child sentence information node is returned (step 352).
2).

With the above procedure, the process returns to the caller of the recursive call, and the process of “(2) All of the following” 3305 temporarily held is continued.

The processing is started immediately after recursively calling the itemized sentence analysis (step 3525).

Next, the control type of the program is acquired from the read statement record of “(2) All of the following” 3305 (step 3607).

Next, the control type of the acquired program is determined. Since “(2) All of the following” 3305 corresponds to the condition attribute, “program control type = condition”
Is satisfied, and the condition attribute is processed (step 3609). The sentence information node 33 of “(1) Any of the following” of the return value when the item text analysis is recursively called as the current pointer of the sentence information node.
27 is set (step 3626).

Next, the address before the child sentence information node is NU.
When the LL is determined, the child sentence information node previous address value is "(1) Date is a calendar day."
Since 21 is set, the child sentence information node previous address = NULL, and the condition of "N" is satisfied (step 3713).

Next, when the type of the sentence information node of the parent sentence is determined, since the sentence information of “all of the following” 3301 corresponds to the condition attribute, the program control type = condition “Y” is satisfied. (Step 3517). Further, when the program control type is determined in detail, the condition (logical product) is satisfied (step 3734).

Next, the sentence information node 3321 of “(1) Date is a calendar date”, “(2) Date is blank”.
Lines 3330 and 3331 for conditions (logical product) are set up between the sentence information node 3322 of “1” and the sentence information node 3327 of “(1) One of the following” (step 3735).

Next, the sentence information node 3327 of “(1) Either of the following” is set to the child sentence information node previous pointer (step 3719).

[0234] In this case, a directed graph deployment area 2010
43 (b), the sentence information node 3321 of "(1) Date is a calendar date." And the sentence information node 3322 of "(2) Date is blank." Sentence information node 332 of “(1) One of the following.”
7, a condition (logical product) is set.

Thereafter, the specification data development area 2
From 004, one record is read from the statement record next to “(2) All of the following” 3305 (step 352).
0).

Next, when it is determined whether there is a sentence, since there is no corresponding sentence, the process exits the sentence reading loop (step 3521).

Next, the address of the sentence information node 3321 of “(1) Date is a calendar date” stored in the head address of the child sentence information node is returned (step 3522).

With the above procedure, the process returns to the caller of the recursive call, and the “all of the following” 3
The processing of 301 is continued.

Next, the read "All of the following."
The control type of the program is acquired from the statement record of 3301 (step 3607).

The processing is started immediately after recursively calling the itemized sentence analysis (step 3523).

Next, the control type of the acquired program is determined. Since “All of the following” 3301 corresponds to the condition attribute, the condition of “Y” of “program control type = condition” is satisfied, and the condition attribute is processed (step 3609). The sentence information address 33 of the return value "(1) Date is a calendar date" when the item text analysis is recursively called with the current pointer of the sentence information node.
21 is set (step 3626).

Next, the address before the child sentence information node is NU.
When it is determined whether the LL is the LL, the previous sentence information node address value is not set, and thus the previous sentence information node address of the child is NU.
LL, and the condition of “Y” is satisfied (step 371).
3).

In the head address of the child statement information node, “(1)
The date is a calendar day. The sentence information node of "3303" is set and stored as a child sentence information node head address (step 3731).

The sentence information node of “(1) Date is a calendar date” 3303, for which the sentence analysis has been completed, is set to the child sentence information node previous address (step 3519).

Next, one sentence record next to “all of the following” 3301 is read from the specification data development area 2004 (step 3520).

Next, when it is determined whether there is a sentence, since there is no corresponding sentence, the process exits the sentence reading loop (step 3521). Next, the address of the sentence information node 3321 of "(1) Date is a calendar date" stored in the child sentence information node head address is returned (step 3522).

With the above procedure, a directed graph representing the control of the program is created.

The sentence 330 itemized in FIGS.
Demonstrate the procedure for generating a source a directed graph 3320 created from 0 to the original. 44 to 45 sequentially show the generation process. Among them, FIGS. 44 and 45 schematically show the whole of the generation process, and FIGS. 46 to 51 show each process in detail.

[0249] First, the source generation rules are described in a source to be created corresponding to each sentence can be divided into the following four types of rules.

[0250] 1. Bullet expansion rule This rule is used to expand the source from the bulleted text. If the bulleted text continues, the parent text is repeatedly expanded to the position specified by%, otherwise,
The range specified by #S to #E is expanded.

[0251] 2. Sentence expansion rule This is a rule for expanding a sentence that qualifies the itemized sentence, and expands a source generation rule of the sentence corresponding to the position represented by & NEXT.

[0252] 3. Modification expansion rules These are rules for expanding the qualification statement. [~] Indicates the source generation generated when the relevant qualification statement is used alone, and ｛-｛indicate the corresponding qualification statement is the last qualification statement. In the case of, it represents the range where the source is not generated.

[0253] 4. Term expansion rule This is a rule for expanding the content of the term described in "-" in the sentence. It is expressed by $ 1 to $ 99, and indicates that information of the term specified by the number is expanded.

FIGS. 52 and 53 are flowcharts for generating a paragraph sentence source, FIG. 54 is a flowchart for formula expansion, and FIG. 55 is a flowchart for sentence expansion.

First, the procedure for performing the processing of the sentence information node 3321 of “date is a calendar day” will be described with reference to a flowchart.

In the flowchart for generating the itemized sentence source shown in FIGS. 52 and 53, the sentence information node 3 of the first “date is a calendar day” in the directed graph is used as the input data 5701.
The processing is started from 321.

First, the control type of the program is acquired based on the information of the sentence information node 3321 of “date is a calendar day” and has a condition (logical sum) attribute. Using the sentence / term data search module 2008. If there is a condition (logical sum) line at this time,
The latter part 2360 of the source generation rule is replaced with the former part 2250 of the source generation rule of “all of the following.” And read as the source generation rule (step 570).
2).

Next, the read source generation rules are obtained one by one, and the first “& NEXT” 4510 is obtained (step 5703).

Next, it is determined whether or not there is a source generation rule. If there is a source generation rule, a process in a horizontal source generation rule acquisition loop is performed (step 570).
4).

Subsequently, when the type of the acquired source generation rule is determined, since “& NEXT” 4510 corresponds to the sentence expansion rule, “Y” is satisfied in the determination of source generation rule = sentence expansion rule, and Processing of an expansion rule is performed (step 5706).

Next, when the attribute of “date is a calendar day” 3303 is determined, since it corresponds to the decoration attribute, “Y” is established in the determination of sentence attribute = modification expansion, and the processing of the decoration expansion is performed (step). 5729).

Next, in order to expand the source generation rule 2950 of “date is calendar date” 3303, the processing of the statement information node 3321 of “date is calendar date” is temporarily suspended, and “ The source generation rule name of “Date is a calendar day” 3303 is passed and the qualified expansion is called (step 5743).

In the sentence development flowchart of FIG. 55, “date is a calendar day” 3303 as input data 6001
Receives the source production rule name of. "The date is a calendar day"
Is obtained using the sentence / term data search module 2008 (step 600).
2).

Next, the read source generation rules are acquired one by one, and the first “CALL'REKI_CHK'U
SING # 1 R_CODE "4520 is obtained (step 6003).

Next, it is determined whether or not there is a source generation rule. If there is a source generation rule, a process in a statement source generation rule acquisition loop is performed (step 6004).

Subsequently, when the type of the acquired source generation rule is determined, since “$ 1” corresponds to the sentence expansion rule, “Y” is satisfied in the determination of source generation rule = term expansion rule, and Process expansion rules (step 600)
7).

Next, the term expansion rule corresponds to a rule that replaces the “date” 4532 part of “is a calendar day” 4531, so “CALL'REKI_CHK'U
The “$ 1” part of SING $ 1 R_CODE is replaced with “DATE” 4599 corresponding to “date” 4532 (step 6018).

Next, “END-CALL” 4533,
The source generation rule of “IF (R_CODE = 0)” 4534 has no corresponding rule, and is thus output directly to the source (step 6008).

Since there is no next source generation rule, the process exits the qualified source generation rule acquisition loop and ends the sentence expansion flowchart (step 6010). The source generated so far is as follows.

CALL'REKI_CHK'USING
DATE R_CODE END-CALL IF (R_CODE = 0) This is indicated by reference numeral 5270 in FIG.

[0271] When the sentence development of "Date is a calendar day" 3303 is completed, the processing is started from the generation of the paragraph document source for which the processing is temporarily suspended. The processing is started from the position where the sentence expansion is called (step 574).
3).

The source generation rule of the next “THEN” 4535 has no corresponding rule, and is output to the source as it is (step 5707). The source generated so far is as follows.

In FIG. 46, this is denoted by reference numeral 5271.
Indicated by.

Since the next “%%” 4536 corresponds to the item development rule, “Y” is satisfied in the determination of the source generation rule = item development rule, and the item development rule is processed (step 5705).

Next, when it is determined whether or not there is a node in the horizontal direction, the statement information node 3327 of “any of the following” exists, so that the condition of “Y” is satisfied with “horizontal node = present”. Then, the item is expanded (step 571).
8).

Next, “any of the following” 3306
Temporarily expands the processing of the sentence information node 3321 of “date is a calendar day” to expand the source generation rule of
A sentence information node 3327 of “any of the following” is passed as the input data 5701, and the generation of the itemized text source is recursively called (step 5722).

The process is newly started from the beginning of the flowchart by recursively calling the generation of the paragraph text source. At this time, a sentence information node 3327 of “any of the following” is obtained as input data 5701.

First, the control type of the program is acquired based on the information of the statement information node 3327 of “any one of the following.” Since it has a condition (logical product) attribute, “one of the following: Source generation rules 2350, 2360
Is acquired using the sentence / term data retrieval module 2008 and is read as a source generation rule (step 5).
702).

Next, a source generation rule is acquired for each of the read source generation rules one by one, and the first “& NEXT” 52
13 is acquired (step 5703).

Next, it is determined whether or not there is a source generation rule. If there is a source generation rule, a process in a horizontal source generation rule acquisition loop is performed (step 570).
4).

Subsequently, when the type of the acquired source generation rule is determined, since “& NEXT” 5213 corresponds to the sentence expansion rule, “Y” is satisfied in the determination of source generation rule = sentence expansion rule, and Processing of an expansion rule is performed (step 5706).

Next, when the attribute of “any of the following” 3306 is determined, since it corresponds to the expression attribute, “Y” is established in the determination of statement attribute = expression expansion, and the expression expansion process is performed ( Step 5727).

Next, “any of the following” 3306
In order to expand the source generation rule, the process of the sentence information node 3327 of “any of the following.” Is temporarily suspended, and as the input data 5901, the sentence information node of “the account type is normal”. 3324 is passed, and expression expansion is called (step 5726).

In the expression expansion flowchart of FIG. 54, a sentence information node 3324 of “account type is ordinary” is received as input data 5901.

Next, the source / generating rule 2450 of “any of the following.”
008 and obtain (step 5902).

Next, the read source generation rules are acquired one by one, and the first "[IF (] (& NEXT) {OR}
%% [)] "4537 (step 590)
3).

Next, it is determined whether or not there is a source generation rule. If there is a source generation rule, processing in an expression source generation rule acquisition loop is performed (step 590).
4).

Subsequently, when the type of the acquired source generation rule is determined, since "[IF (]""corresponds to the modification expansion rule," Y "is satisfied in the determination of source generation rule = modification expansion rule. Then, the processing of the modification expansion rule is performed (step 5907).

Next, when the type of the source generation rule is obtained, “Y” is established in the determination of the source generation rule type = [〜], and the processing of the expansion rule of [〜] is performed (step 592).
1).

Next, when it is determined whether or not the expression “any of the following” is nested and described, it is determined that the expression “nested” exists because the expression is not nested and described. If the determination is "N", the sentence expansion rule is processed (step 5926).

Since expansion processing in the range of [-] is performed,
F () is output to the source (step 5927).

Next, since the source generation rule of “(” has no corresponding rule, it is directly output to the source (step 5909).

Next, since “& NEXT” corresponds to the sentence expansion rule, the source generation rule = “Y” in the judgment of the sentence expansion rule.
Holds, and the processing of the sentence expansion rule is performed (step 590).
6). Next, when the attribute of “account type is ordinary.” 3324 is determined, since the attribute is a qualified attribute, “Y” is established in the determination of sentence attribute = qualified development, and a qualified development process is performed (step 5914). .

Next, "the account type is normal" 330
7, the processing of the sentence information node 3327 of “any of the following.” Is temporarily suspended, and the source generation rule name of “account type is normal.” Handover and call for a qualified expansion (step 591
5).

In the sentence expansion flowchart of FIG. 55, the input data 6001 is “account type is ordinary”.
The source generation rule name 307 is received. The source generation rule 2850 of “account type is normal” is obtained by using the sentence / term data retrieval module 2008 (step 6002).

Next, the read source generation rules are acquired one by one, and the first “[IF] $ 1- $ 2 [)]” 453 is obtained.
8 is acquired (step 6003). Next, it is determined whether there is a source generation rule. If there is a source generation rule, a process in a statement source generation rule acquisition loop is performed (step 6004). Subsequently, when the type of the acquired source generation rule is determined, [IF (] corresponds to the qualification expansion rule, so that “Y” is satisfied in the determination of source generation rule = qualification expansion rule, and Processing is performed (step 6006).

Next, when the type of the source generation rule is obtained, “Y” is established in the determination of the source generation rule type = [〜], and processing of the expansion rule of [〜] is performed (step 601).
2).

Next, "the account type is normal" 330
When it is determined whether expression 7 is described in a nested manner, "any of the following" is nested in the parent, so that "Y" is satisfied in the determination of "nested". ,
Skip the output of "IF (" (step 601
3).

Next, when the type of the acquired source generation rule is determined, since “$ 1” corresponds to the term expansion rule, “Y” is established in the determination of source generation rule = term expansion rule, The term expansion rule is processed (step 600).
7).

Next, in the term expansion rule, since the part of “account type” 4539 of “is is” is replaced by “[IF] ＄ 1- ＄ 2 [)]”
Is replaced with “KOUZA_CODE” 4540 corresponding to “account type” 4539 (step 6018).

Next, since there is no corresponding rule for the source generation rule of "=", it is output to the source as it is (step 6008).

Next, since “$ 2” corresponds to the sentence expansion rule, “Y” is established in the determination of the source generation rule = term expansion rule, and the term expansion rule is processed (step 600).
7).

Next, in the term expansion rule, since the part of “normal” 4541 of “is is” 2850 falls under the rule to be replaced, “{IF} ＄ 1- {2 [)]” The part “2” is replaced by “′” corresponding to “normal” 4541.
0 ′ ”4542 (step 6018).

Next, [)] corresponds to the modification development rule, so that “Y” is established in the determination of the source generation rule = modification development rule, and the modification development rule is processed (step 600).
6).

Next, when the type of the source generation rule is obtained, “Y” is established in the determination of the source generation rule type = [〜], and the processing of the expansion rule of [〜] is performed (step 601).
2). Next, the formula of “account type is normal.” 3307
Is nested, it is determined that “any of the following” is nested in the parent, so “Y” is satisfied in the determination of “nested” and “ ) "
Is skipped (step 6013).

Since there is no next source generation rule, the process exits the statement source generation rule acquisition loop and ends the sentence expansion flowchart (step 6010).

The source generated so far is KOUZA
_CODE = '0'.

[0308] By completing the sentence expansion of "Account type is normal." 3307, the processing of the expression expansion temporarily suspended is started. The processing is started from the position where the sentence expansion is called (step 5915).

Next, the source generation rule of ")" is directly output to the source because there is no corresponding rule (step 5909). Since the next “{OR}” corresponds to the qualification expansion rule, “Y” is established in the determination of the source generation rule = the qualification expansion rule, and the qualification expansion rule is processed (step 5907).

Next, when the type of the source generation rule is obtained, “Y” is established in the determination of the source generation rule type = ｛to｝, and processing of the expansion rules of ｛to｝ is performed (step 592).
2).

Next, "Account type is normal" 330
When it is determined whether there is a sentence next to No. 7 because there is "Account type is current." Is performed (step 5918). "OR" is output to the source by performing the expansion processing in the range of (1) to (5). When "%%" is further read, the source generation rule = "Y" in the judgment of the item development rule because it corresponds to the item development rule
Holds, and the processing of the item expansion rule is performed (step 5).
905).

Next, when it is determined that there is a node, "Y" is satisfied in the determination that there is a node, and the processing of the expression expansion rule is performed because "Account type is current" 3308 exists (step 5912). .

Next, "the account type is immediate" 330
In order to expand the source generation rule of No. 8 , the processing of the sentence information node 3324 of "account type is normal." Is temporarily suspended, and the sentence information node 33 of "account type is immediate."
25 and recursively invoke expression expansion (step 59)
13).

The processing is newly started from the head of the flowchart by recursively performing the expression expansion.

At this time, a sentence information node 3325 “Account type is current” is obtained as input data.

Next, the source / generating rule 2450 of “any of the following.”
008 and obtain (step 5902).

Next, the read source generation rules are acquired one by one, and the first “[IF (] (& NEXT) {OR}
%% [)] "4543 (step 590).
3).

Next, it is determined whether or not there is a source generation rule. If there is a source generation rule, processing in an expression source generation rule acquisition loop is performed (step 590).
4). Subsequently, when the type of the acquired source generation rule is determined, since “[IF (]” ”corresponds to the modification expansion rule,“ Y ”is established in the determination of source generation rule = modification expansion rule, and the modification is performed. The expansion rule is processed (step 590).
7).

Next, when the type of the source generation rule is obtained, “Y” is established in the determination of the source generation rule type = [〜], and the processing of the expansion rule of [〜] is performed (step 592).
1).

Next, "the account type is immediate" 330
When it is determined whether the expression 8 is nested and described, "any of the following" is nested and described in the parent, so "Y" is satisfied in the determination of "nested". Then, the output of “IF (” is skipped (step 592).
6).

Next, since the source generation rule of “(” does not have a corresponding rule, it is directly output to the source (step 5909).

Next, “& NEXT” corresponds to the sentence expansion rule 510.
Since “5” is satisfied, “Y” is established in the determination of the source generation rule = sentence expansion rule, and the processing of the sentence expansion rule is performed (step 5906).

Next, “Account type is immediate.” 3308
Is determined to be a decoration attribute, "Y" is established in the determination of sentence attribute = modification expansion, and a decoration expansion process is performed (step 5914).

Next, "the account type is immediate" 330
In order to expand the source generation rule of No. 8 , the processing of the sentence information node 3325 of “Account type is immediate” is temporarily suspended, and the source generation rule name of “Account type is immediate” 3308 is passed. , Call the expansion (step 591)
5).

In the sentence expansion flowchart of FIG. 55, the input data 6001 is “Account type is immediate.” 3
A source generation rule name 308 is received.

The source generation rule 2850 of “the account type is current” is set to the sentence / term data search module 200
8 (step 6002). Next, the read source generation rules are obtained one by one, and the first “[IF] ＄ 1 = $ 2 [)]” 4544 is obtained (step 6003).

Next, it is determined whether or not there is a source generation rule. If there is a source generation rule, processing in a statement source generation rule acquisition loop is performed (step 600).
4). Subsequently, when the type of the acquired source generation rule is determined, [IF (] is a qualified expansion rule, so that “Y” is satisfied in the determination of the source generation rule = modified expansion rule, and the qualified expansion rule is satisfied. (Step 600)
6).

Next, when the type of the source generation rule is acquired, “Y” is established in the determination of the source generation rule type = [〜], and the processing of the expansion rule of [〜] is performed (step 601).
2).

Next, “the account type is immediate” 330
When it is determined whether the expression 7 is described in a nested manner, "Y" is established in the determination of "nested" because "any of the following" is nested and described in the parent. Then, the output of “IF (” is skipped (step 601).
3).

Next, when the type of the acquired source generation rule is determined, since “$ 1” corresponds to the term expansion rule, “Y” is established in the determination of source generation rule = term expansion rule, The term expansion rule is processed (step 600).
7).

Next, in the term expansion rule, the rule [[IF)] {1 =} corresponds to the rule that replaces the “account type” part of the source generation rule 2850 of “is is”.
2 [)] ”is replaced with“ KOUZA_CODE ”4546 corresponding to“ account type ”4545 (step 6018).

Next, since the source generation rule of “=” has no corresponding rule, it is output to the source as it is (step 6008). Next, since “$ 2” corresponds to the term expansion rule, “Y” is established in the determination of the source generation rule = term expansion rule, and the term expansion rule is processed (step 6007).

Next, in the term expansion rule, “[IF]] {1 =} since the“ interim ”portion of the source generation rule 2850 of“ is is ”corresponds to the replacement rule.
2 [)] ”is replaced with“ 1 ”4548 corresponding to“ quick ”4547 (step 601).
8).

Next, since []] corresponds to the modification development rule, “Y” is satisfied in the determination of the source generation rule = modification development rule, and the modification development rule is processed (step 600).
6). Next, when the type of the source generation rule is acquired, “Y” is established in the determination of the source generation rule type = [〜], and [〜]
(Step 6012). Next, when it is determined whether the expression of “Account type is immediate” 3308 is described in a nested manner, it is determined that “any of the following” is nested in the parent. "Y" is satisfied in the determination of "nesting", and the output of ")" is skipped (step 6013).

Since there is no next source generation rule, the process exits the source generation rule acquisition loop and ends the flowchart of sentence expansion (step 6010).

The source generated so far is KOUZA
_CODE = '1'.

Next, "the account type is immediate" 330
By terminating the sentence expansion of No. 8, the processing of the expression expansion whose processing has been temporarily suspended is started. The processing is started from the position where the sentence expansion is called (step 5915).

Next, the source generation rule of “)” is output to the source as it is because there is no corresponding rule (step 5909). Since the next “{OR}” corresponds to the qualification expansion rule, “Y” is established in the determination of the source generation rule = the qualification expansion rule, and the qualification expansion rule is processed (step 5907).

Next, "the account type is immediate" 330
When it is determined whether or not there is a next sentence of No. 8, since there is no next sentence, “N” is established in the determination of “next sentence” and “OR”
Is skipped (step 5918). Further, since “%%” corresponds to the item development rule, “Y” is established in the determination of the source generation rule = item development rule, and the item development rule is processed (step 5905).

Next, "the account type is immediate" 330
When it is determined whether there is a next sentence of No. 8, since there is no next sentence, "N" is determined in the determination of the presence of the next sentence, and the output of the item development rule is skipped (step 5912).

Since “[]]” corresponds to the modification development rule, “Y” is established in the determination of the source generation rule = modification development rule, and the modification development rule is processed (step 590).
7). Next, when it is determined whether the expression of “Account type is immediate” 3308 is described in a nested manner, “any of the following” is nested in the parent and described. Therefore, "Y" is established in the determination of "nested" and ")"
Is skipped (step 5926).

Since there is no next source generation rule, the flow chart of the expression expansion through the source generation rule acquisition loop is ended (step 5911).

The source generated so far is (KOUZ
A_CODE = '1').

[0344] After terminating the expression of "account type is immediate." 3308, the process of the expression development whose processing has been temporarily suspended is started. The processing is started from the position where the sentence expansion is called (step 5915).

Next, the source generation rule of ")" is directly output to the source because there is no corresponding rule (step 5909).

Since “[]]” corresponds to the modification development rule, “Y” is satisfied in the determination of the source generation rule = modification development rule, and the modification development rule is processed (step 590).
7).

Next, when the type of the source generation rule is obtained, “Y” is established in the determination of the source generation rule type = [〜], and the processing of the expansion rule of [〜] is performed (step 592).
1).

Next, if it is determined whether the expression “any of the following” is described in a nested manner, since the expression is not nested and described, “N” is determined in the determination of “nested”.
Holds, and the processing of the modification expansion rule is performed (step 59).
26).

Since expansion processing of [-] is performed, ")" is output to the source (step 5927).

Since there is no next source generation rule, the process exits the source generation rule acquisition loop and ends the expression expansion flowchart (step 5911).

IF ((KOUZA_CODE = '0')
OR (KOUZA_CODE = '1'))
In FIG.

[0352] By completing the sentence development of "either of the following" 3305, the generation of the itemized sentence source for which the processing is temporarily suspended is started. The processing is started from the position where the expression expansion is called (step 5726).

The source generation rule of the next “THEN” 4535 is directly output to the source because there is no corresponding rule (step 5707).

This is indicated by reference numeral 5273 in FIG.

Since the next “%%” 4536 corresponds to the item development rule, “Y” is satisfied in the determination of the source generation rule = item development rule, and the item development rule is processed (step 5705).

Next, when it is determined whether or not there is a node in the horizontal direction, the sentence information node 332 of “account number is a number”.
8 exists, the condition of "Y" is satisfied when the node in the horizontal direction is present, and the processing of item development is performed (step 57).
22).

Next, "the account number is a number" 330
In order to expand the source generation rule of No. 9, the processing of the sentence information node 3327 of “any of the following.” Is temporarily suspended, and the sentence information node of “the account number is a number” as the input data 5701. 3328 is passed, and the itemized text source generation processing is recursively called.

[0358] By recursively calling the itemized text source generation processing, the processing is newly started from the top of the flowchart. At this time, a sentence information node 3328 of “account number is a number” is obtained as input data.

First, the control type of the program is acquired based on the information of the sentence information node 3328 of “the account number is a number.” Since it has the condition (logical product) attribute, “all of the following. Are acquired using the sentence / term data retrieval module 2008 (step 5702).

Next, the read source generation rules are obtained one by one, and the first “& NEXT” 4549 is obtained.
Next, it is determined whether there is a source generation rule. If there is a source generation rule, a process in a horizontal source generation rule acquisition loop is performed (step 5704). Subsequently, when the type of the acquired source generation rule is determined, “& N
Since “EXT” 4549 corresponds to the sentence expansion rule, “Y” is established in the determination of the source generation rule = the sentence expansion rule, and the sentence expansion rule is processed (step 5706).

Next, “Account number is a number” 3309
Is determined, the expression attribute is satisfied, so that "Y" is established in the determination of sentence attribute = modification expansion, and a modification expansion process is performed (step 5729).

Next, "the account number is a number" 330
In order to develop the source generation rule of No. 9, the processing of the sentence information node 3328 of “Account number is a number” is temporarily suspended, and the source of “Account number is a number” 3309 as input data 6001. The generation rule name is passed, and sentence expansion is called (step 5743).

In the sentence development flowchart of FIG. 54, “account number is a number.”
309 is received.

Next, the source generation rule 4450 of “account number is a number.”
008 (step 6002).

Next, the read source generation rules are acquired one by one, and the first “[IF () ＄ 1 = NUMERI”
C [)] ”4550 (step 6003).
Next, it is determined whether there is a source generation rule. If there is a source generation rule, a process in a statement source generation rule acquisition loop is performed (step 6004).

Subsequently, when the type of the acquired source generation rule is determined, since “[IF (]” ”corresponds to the modification expansion rule,“ Y ”is satisfied in the determination of source generation rule = modification expansion rule. Then, the processing of the modification expansion rule is performed (step 6006).

Next, when the type of the source generation rule is acquired, “Y” is established in the determination of the source generation rule type = [〜], and the processing of the expansion rule of [〜] is performed (step 601).
2).

Next, "the account number is a number" 330
When it is determined whether expression 9 is described in a nested manner, "nesting exists" because it is not described in a nested manner
"N" is established, and the processing of the sentence expansion rule is performed (step 6013). Since the expansion processing in the range of [-] is performed, "IF (" is output to the source (step 601).
4).

Next, when the type of the acquired source generation rule is determined, since “$ 1” corresponds to the term expansion rule, “Y” is established in the determination of source generation rule = term expansion rule, The term expansion rule is processed (step 600).
7).

Next, in the term expansion rule, since the "account number" 4551 part of "is a number" 5161 falls under the rule to be replaced, "[IF () @ 1 = N UME
RIC [)] ”is replaced with“ account number ”455
1 is replaced with “KOUZA_NO” 4552 (step 6018).

Next, the source generation rule of “= N UMERIC” is output to the source as it is because there is no corresponding rule (step 6008). Since “)]” corresponds to the sentence expansion rule, “Y” is established in the determination of source generation rule = qualification expansion rule, and processing of the expansion rule is performed (step 6006).

Next, when the type of the source generation rule is acquired, “Y” is established in the determination of the source generation rule type = [〜], and processing of the expansion rule of [〜] is performed (step 601).
2). Next, when it is determined whether or not the expression of “account number is a number” is described in a nested manner, “N” is satisfied in the determination of “nested” because the expression is not nested and described. Perform the processing of the sentence expansion rule (step 601)
3).

By performing the expansion processing in the range [-], ")" is used.
Is output to the source (step 6014).

Since there is no next source generation rule, the process exits the source generation rule acquisition loop and ends the sentence expansion flowchart (step 6009).

IF (KOUZA_NO = N UMERI)
C) This is indicated by reference numeral 5274 in FIG.

[0376] By completing the sentence expansion of "Account number is a number." 3309, the generation of the paragraph sentence source whose processing has been temporarily suspended is started. The processing is started from the position where the sentence expansion is called (step 5743).

The source generation rule of the next “THEN” 4553 is directly output to the source because there is no corresponding rule (step 5707).

This is indicated by reference numeral 5275 in FIG.

Since the next “%%” 4554 corresponds to the item development rule, “Y” is satisfied in the judgment of the source generation rule = item development rule, and the item development rule is processed (step 5705).

Next, when it is determined whether or not there is a node in the horizontal direction, since there is no corresponding sentence information node, the condition of “N” is satisfied with the node in the horizontal direction = Yes, and the expansion of #S to #E Processing is performed (step 5718).

Since the expansion processing in the range from #S to #E is performed, "$ processing" is output to the source (step 5719).

Since the next source generation rule “#>” 4555 is a symbol indicating the end of the horizontal source generation rule,
Exit the horizontal source generation rule acquisition loop (step 5)
709).

{Processing} This is indicated by reference numeral 5276 in FIG.

Further, the read source generation rules are acquired one by one (step 5810).

Next, it is determined whether or not there is a source generation rule. If there is a source generation rule, a process in a vertical source generation rule acquisition loop is performed (step 581).
1).

Next, “ELSE”, “CONTINU”
Since “E” and “END-IF” 4556 do not all have a corresponding source generation rule, they are directly output to the source (step 5815).

Since there is no next source generation rule, the process exits the vertical source generation rule acquisition loop and terminates the paragraph document source generation (step 5817).

ELSE CONTINUE END-IF This is indicated by reference numeral 5278 in FIG.

[0389] After the generation of the itemized sentence source of the item "Account number is a number.", The sentence information node 332 of "any of the following" has been temporarily suspended.
7. Start generating the source of the paragraph text from step 7.

The processing is started immediately after recursively calling the generation of the itemized text source (step 5722).

Since the next source generation rule “#>” is a symbol indicating the end of the horizontal source generation rule, the process exits the horizontal source generation rule acquisition loop (step 570).
9).

Further, the read source generation rules are acquired one by one (step 5810).

Next, it is determined whether or not there is a source generation rule. If there is a source generation rule, processing in a vertical source generation rule acquisition loop is performed (step 581).
1).

Next, "ELSE", "CONTINU"
E "and" END-IF "4598 are not output to the source without any corresponding source generation rule (step 5815). Since there is no next source generation rule, the process exits the vertical source generation rule acquisition loop, and ends the generation of the paragraph document source (step 5817).

ELSE CONTINUE END-IF This is indicated by reference numeral 5279 in FIG.

[0396] The generation of the itemized sentence source of the item "one of the following" is completed, and the generation of the itemized item source from the sentence information node 3321 of "the date is a calendar date." To start.

The processing is started immediately after recursively calling the generation of the itemized text source (step 5722).

Since the next source generation rule “#>” 4557 is a symbol indicating the end of the horizontal source generation rule,
Exit the horizontal source generation rule acquisition loop (step 5)
709). Further, the read source generation rules are acquired one by one (step 5810).

Next, it is determined whether or not there is a source generation rule. If there is a source generation rule, a process in a vertical source generation rule acquisition loop is performed (step 581).
1).

Next, “ELSE” 4558 is output to the source as it is because there is no corresponding source generation rule (step 5815).

ELSE This is indicated by reference numeral 5280 in FIG. next"%
% ”4559 corresponds to the item development rule, so that“ Y ”is established in the determination of the source generation rule = the item development rule,
The item development rule is processed (step 5812).

Next, when it is determined whether there is a node in the vertical direction, since the sentence information node 3322 of “date is blank” exists, the condition of “Y” is satisfied when the node in the vertical direction is present, and Perform expansion processing.

Next, to expand the source generation rule of “date is blank.” 3304, “date is a calendar date”
Temporarily suspend the processing of the sentence information node 3321, passes the statement information node 3322 of "date is blank." As the input data 5701, recursively call the bullet sentence source generation.

[0404] By recursively calling the generation of the itemized text source, the process is newly started from the top of the flowchart. At this time, a sentence information node 3322 of “date is blank” is obtained as input data.

First, the control type of the program is acquired based on the information of the sentence information node 3322 of “date is blank.” Since it has a condition (logical product) attribute, “all of the following” The source generation rules 2250 and 2260 are obtained using the sentence / term data retrieval module 2008 (step 5702).

Next, the read source generation rules are obtained one by one, and the first “& NEXT” 4560 is obtained.
Next, it is determined whether there is a source generation rule. If there is a source generation rule, a process in a horizontal source generation rule acquisition loop is performed (step 5704).

Subsequently, when the type of the acquired source generation rule is determined, since “& NEXT” 4560 corresponds to the sentence expansion rule, “Y” is satisfied in the determination of source generation rule = sentence expansion rule, and Processing of an expansion rule is performed (step 5706). Next, when the attribute of “date is blank.” 3304 is determined, it corresponds to the decoration attribute, so “Y” is established in the determination of sentence attribute = modification expansion, and the processing of decoration expansion is performed (step 5729).

Next, in order to expand the source generation rule of “date is blank” 3304, the processing of statement information node 3322 of “date is blank” is temporarily suspended, and as input data 6001 The source generation rule name of "date is blank." 3304 is passed, and the sentence expansion is called (step 5714).

In the sentence expansion flowchart of FIG. 55, “date is blank” 330 as input data 6001.
4 is received.

The source generation rule 4561 “Date is blank” is obtained by using the sentence / term data retrieval module 2008 (step 6002). Next, the obtained source generation rules are acquired one by one, and the first "[IF
(] $ 1 = SPACE [)] ”4561 is acquired (step 6003).

Next, it is determined whether or not there is a source generation rule. If there is a source generation rule, a process in a statement source generation rule acquisition loop is performed (step 6004).
Subsequently, when the type of the acquired source generation rule is determined, since [[IF (]) corresponds to the modification expansion rule,
“Y” is established in the determination of the source generation rule = the modification development rule, and the modification development rule is processed (step 600)
6).

Next, when the type of the source generation rule is obtained, “Y” is established in the determination of the source generation rule type = [〜], and the processing of the expansion rule of [〜] is performed (step 601).
2). Next, when it is determined whether the expression “date is blank” is described in a nested manner, since the expression is not nested and described, “N” is established in the determination of “nested”, The processing of the expansion rule [-] is performed (step 601).
3).

[0413] By performing expansion processing in the range of [~], "IF
() Is output to the source (step 6014).
When the type of the acquired source generation rule is determined, $ 1
Corresponds to the term expansion rule, so that “Y” is established in the determination of the source generation rule = term expansion rule, and the term expansion rule is processed (step 6007).

[0414] Next, in the term expansion rule, since the "date" 4563 part of "is blank" 4562 falls under the rule to be replaced, "[IF () $ 1 = SPAC
E [)] ”is replaced with“ DATE ”4564 corresponding to“ Date ”4563 (step 60).
18).

Next, the source generation rule of “= SPACE” is output to the source as it is because there is no corresponding source generation rule (step 6009). Since “[]]” corresponds to the qualification expansion rule, “Y” is satisfied in the determination of the source generation rule = the qualification expansion rule, and the qualification expansion rule is processed (step 6006).

Next, when the type of the source generation rule is obtained, “Y” is established in the determination of the source generation rule type = [〜], and the processing of the expansion rule of [〜] is performed (step 601).
2). Next, when it is determined whether the expression “date is blank” is described in a nested manner, since the expression is not nested and described, “N” is established in the determination of “nested”, The processing of the sentence expansion rule is performed (step 6013).

Performing the expansion process in the range of [~] results in ")".
Is output to the source (step 6014).

Since there is no next source generation rule, the process exits the source generation rule acquisition loop and ends the sentence expansion flowchart (step 6010).

IF (DATE = SPACE) This is indicated by reference numeral 5281 in FIG.

[0420] Upon completion of the sentence expansion of "Account number is a number" 3309, the generation of the item source temporarily suspended is started. The processing is started from the position where the sentence expansion is called (step 5915).

The source generation rule of the next “THEN” 4565 is directly output to the source because there is no corresponding rule (step 5909).

This is indicated by reference numeral 5282 in FIG.

Since the next “%%” 4566 corresponds to the item development rule, “Y” is satisfied in the determination of the source generation rule = item development rule, and the item development rule is processed (step 5705).

Next, when it is determined whether or not there is a node in the horizontal direction, there is a sentence information node 3327 of “any of the following”. The item is expanded (step 572).
2).

Next, “any of the following” 3306
Temporarily expands the processing of the sentence information node 3322 of “date is blank” to expand the source generation rule of
The sentence information node 3327 of “any of the following” is passed as the input data 5701, and the generation of the paragraph document source is recursively called in the same procedure as in 4536 (step 5722).

When the generation of the item text source of “any of the following” is completed, the generation of the item text source of the item “any of the following” sentence information node 3322 whose processing is temporarily suspended is started. I do.

The following source generation rule “#>” 4567 is
Since the symbol indicates the end of the horizontal source generation rule, the process exits the horizontal source generation rule acquisition loop (step 5708).

IF ((KOUZA_CODE = '0')
OR (KOUZA_CODE = '1')) THEN IF (KOUZA_NO = N UMERIC) THEN {Process ELSE CONTINUE END-IF ELSE CONTINUE END-IF This is indicated by reference numeral 5283 in FIG.

Further, the read source generation rules are acquired one by one (step 5810). Next, it is determined whether or not there is a source generation rule. If there is a source generation rule, a process in a vertical source generation rule acquisition loop is performed (step 5811).

Next, “ELSE”, “CONTINU”
"E" and "END-IF" 4568 are directly output to the source because there is no corresponding source generation rule (step 5815).

Since there is no next source generation rule, the process exits the vertical source generation rule acquisition loop and terminates the generation of the paragraph document source (step 5817).

ELSE CONTINUE END-IF This is indicated by reference numeral 5284 in FIG.

According to the above procedure, the source is generated from the itemized text.

[0434] will be described with reference to the flowchart of FIGS. 29 to 31 is an example of creating a directed graph 3020 from bullets sentences Figure 26.

First, “set a calendar date for date” 30
01 is processed.

First, NULL is set to the head address of the sentence information node of the child to be processed (step 3503). At the same time, NULL is set to the previous address of the child sentence information node to perform initialization (step 35).
04).

[0437] Next, when one sentence record of a child of the itemized sentence is read from the specification data development area 2004, the first item "set calendar date to date" 30 is read.
01 is read (step 3505).

Next, it is determined whether there is a sentence of the same level. Already, “Set calendar date to date.” 3001
Has been read, the sentence becomes "there is a sentence of the same level", and the processing of the statement reading loop is continued (step 3506).

Next, when it is determined whether or not there is child sentence information, since there is no child sentence information, "N" is established when the child sentence information is present (step 3540).

Next, the control type of the program is obtained from the sentence record of the read “set calendar date in date” 3001 (step 3607). Next, the control type of the acquired program is determined. Then, “set calendar day in date” 3001 corresponds to the processing attribute,
The condition of “Y” of “program control type = processing” is satisfied, and processing of the processing attribute is performed (step 3611).

[0441] passing the directed graph generation module 2009, to instruct the creation of a sentence information node (step 362
9).

Next, the address of the sentence information node 3021 of “set calendar date in date” is set as the sentence information current address (step 3630). Next, when it is determined whether or not the child sentence information node preceding address is NULL, the child sentence information node preceding address is set to NULL because the child sentence information node preceding address value is not set, and the condition of “Y” is satisfied. (Step 3713).

The sentence information start address of the child is set to the sentence information node 3021 of “set calendar date in date”, and the child sentence information start address is stored (step 363).
1).

The sentence information node 3021 of “set calendar date for date” for which the sentence analysis has been completed is set to the child sentence information node previous address (step 3519).

[0445] At this time, "to set the calendar to date." Directed graph is created as a statement information node 3021 of processing.

Next, “Set an immediate account” 30
02.

Next, from the specification data development area 2004, when one sentence record following “Set calendar date in date” 3001 is read, “Set interim account” 3001 is read ( Step 352
0).

Next, it is determined whether there is a sentence of the same level. If there is a sentence, the processing of the sentence reading loop is continued (step 3506). Next, when it is determined whether or not there is child sentence information, since there is no child sentence information, "N" is established when the child sentence information is present (step 354).
0).

Next, the control type of the sentence program is obtained from the sentence record of the read “set an immediate account” 3002 (step 3607). Next, the control type of the acquired program is determined. Since “Set an interim account” 3002 corresponds to the processing attribute, the condition of “Y” of “program control type = processing” is satisfied, and processing of the processing attribute is performed (step 3611). Passing the directed graph generation module 2009, and instructs the creation of the sentence information node (step 3629).

Next, the sentence information node 3022 of “set an immediate account” is set in the sentence information node current pointer.
Is set (step 3630).

Next, the child statement information node previous address is NU
When the LL is determined, the sentence information node 30 of the child sentence information node previous address value is "set calendar date to date."
Since 21 is set, the child sentence information node previous address = NULL, and the condition of "N" is satisfied (step 3713).

Next, when the type of the parent sentence information is determined,
Since “set calendar day in date” 3002 corresponds to the processing attribute, the condition of program control type = processing is satisfied (step 3015). Next, a processing line 3023 is set up between the sentence information node 3021 of "Set the calendar date in the date." And the sentence information node 3022 of "Set the current account." (Step 3733).

[0453] At this time, "to set the current to account." Directed graph is created as a statement information node 3022 of the process, "to set the calendar to date." Sentence information node 3
A processing line 3023 is provided between the statement information node 3022 of the statement information node 3022 and the statement information node 3022 of “Set an immediate account”.

Next, one record is read from the specification data development area 2004 as a sentence record next to “Set an immediate account” 3003 (step 3520).

Next, when it is determined whether there is a sentence, since there is no corresponding sentence, the process exits the sentence reading loop (step 3521). Next, the address 3021 of the sentence information node stored in the child sentence information node head address, "set the calendar date in the date." Is returned (step 3522).

[0456] directed graph 30 in FIG. 26 by the above procedure
20 are created.

[0457]

As described above, in the present invention, sentence information
Referring to a file, one or more processing conditions and processing
The function specification including the content is a bulleted sentence (Japanese
Source program generation system described in
By inputting from the input device attached to the stem,
A directed graph corresponding to the program specification is created.
Source corresponding to the program specifications based on the directional graph
Program is generated. Because of this, new programs
When creating a program, the function specifications of the newly developed program
Is written in a bulleted sentence format, and this bulleted sentence format
Enter the data of the functional specifications of the newly developed program
No. Program function specifications can be entered in bulleted text
By doing so, the function of the program including complicated conditions
Specifications can be easily written by people without programming knowledge
And generate the desired source program.
You.

[0458] Furthermore, it is stored in the sentence information file.
Each statement is used to generate a source program from that statement.
Has types indicating source generation rules and program control structures
Because of this, the control structure of the program is also
Is described at the same time by the description of
There is no need to describe the control structure by another means. Accordingly
Even people without programming knowledge can control programs
Enter functional specifications without considering the structure
Source programs can be generated. Further
In addition, the program included in each sentence that constitutes the bulleted statement
Discriminates program control structure from type data of program control structure
To create a directed graph,
The program follows the directed graph and the source generation rules of each sentence.
To generate a source program for each sentence
Then, according to the control structure shown by the directed graph,
It is only necessary to combine them.
This has the effect of simplifying the body.

[Brief description of the drawings]

FIG. 1 is an explanatory diagram showing an embodiment of a function specification description when a source program is generated by describing a program function specification in a tabular specification.

FIG. 2 is a flowchart for the function specification description of FIG. 1;

FIG. 3 is an explanatory diagram showing an example of a source program generated for the functional specification description of FIG. 1;

FIG. 4 is a second functional specification description when a source program is generated by describing a program functional specification in a tabular specification.
It is explanatory drawing which shows Example of this.

FIG. 5 is a flowchart for the function specification description of FIG. 4;

FIG. 6 is an explanatory diagram showing an example of a source program generated for the functional specification description of FIG. 4;

FIG. 7 is a third functional specification description when a source program is generated by describing a program functional specification in a tabular specification.
It is explanatory drawing which shows Example of this.

FIG. 8 is a flowchart for the function specification description of FIG. 7;

FIG. 9 is an explanatory diagram showing an example of a source program generated for the functional specification description of FIG. 7;

FIG. 10 is an explanatory diagram showing a fourth embodiment of the functional specification description when a source program is generated by describing a program functional specification in a tree structure specification.

11 is a flowchart for the function specification description of FIG.

FIG. 12 is an explanatory diagram showing an example of a source program generated for the functional specification description of FIG. 10;

FIG. 13 is an explanatory diagram showing a fifth embodiment of the function specification description when a source program is generated by describing a program function specification in a decision table specification.

FIG. 14 is a flowchart for the function specification description of FIG.

FIG. 15 is an overall flowchart for the function specification description of FIG. 13;

FIG. 16 is an explanatory diagram showing an example of a source program generated for the functional specification description of FIG. 13;

FIG. 17 is a configuration diagram showing an embodiment of a source generation system from a paragraph sentence.

FIG. 18 is a file-related diagram of a system for generating a source from an itemized sentence.

FIG. 19 is an explanatory diagram illustrating an example of a conditional conjunction statement.

FIG. 20 is an explanatory diagram showing an example of a conditional ethics Japanese sentence.

FIG. 21 is an explanatory diagram showing an example of a logical product expression statement.

FIG. 22 is an explanatory diagram showing an example of a logical sum expression sentence.

FIG. 23 is an explanatory diagram showing an example of a processing sentence.

FIG. 24 is an explanatory diagram showing an example of a modification sentence.

FIG. 25 is an explanatory diagram showing another example of a modification sentence.

FIG. 26 is an explanatory diagram showing a flow from an itemized sentence and an example of generation of an analysis tree.

FIG. 27 is an explanatory diagram showing a flow from an itemized sentence and a second generation example of the analytic tree.

FIG. 28 is a flowchart showing the entire generation of a source from an itemized sentence.

FIG. 29 is a flowchart of generating a parse tree from an itemized sentence.

FIG. 30 is a flowchart of generating a parse tree from an itemized sentence.

FIG. 31 is a flowchart of generating a parse tree from an itemized sentence.

FIG. 32 is a data flow diagram for generating a parse tree from an itemized sentence.

FIG. 33 is a data flow diagram for generating a parse tree from an itemized sentence.

FIG. 34 is a data flow diagram for generating a parse tree from an itemized sentence.

FIG. 35 is a data flow diagram for generating a parse tree from an itemized sentence.

FIG. 36 is a data flow diagram for generating a parse tree from an itemized sentence.

FIG. 37 is a data flow diagram for generating a parse tree from an itemized sentence.

FIG. 38 is a data flow diagram for generating a parse tree from an itemized sentence.

FIG. 39 is a data flow diagram for generating a parse tree from an itemized sentence.

FIG. 40 is a data flow diagram for generating a parse tree from an itemized sentence.

FIG. 41 is a data flow diagram for generating a parse tree from an itemized sentence.

FIG. 42 is a data flow diagram for generating a parse tree from an itemized sentence.

FIG. 43 is a data flow diagram for generating a parse tree from an itemized sentence.

FIG. 44 is a data flow diagram showing a process of generating a source from a parse tree.

FIG. 45 is a data flow diagram showing a process of generating a source from a parse tree together with FIG. 44.

FIG. 46 is an explanatory diagram showing an example of a source program generated from a parse tree.

FIG. 47 is a data flow diagram showing in detail a process of generating a source from a parse tree.

FIG. 48 is a data flow diagram showing the first part of the process of generating a source from a parse tree in more detail.

FIG. 49 is a data flow diagram showing in further detail the second part of the process of generating a source from a parse tree.

FIG. 50 is a dataflow diagram showing in further detail the third part of the process of generating a source from a parse tree.

FIG. 51 is a dataflow diagram showing in further detail the fourth part of the process of generating a source from a parse tree.

FIG. 52 is a flowchart showing details of processing for generating a source program from an itemized sentence.

FIG. 53 is a flowchart showing details of a process of generating a source program from an itemized sentence.

FIG. 54 is a flowchart showing details of processing for expanding an expression.

FIG. 55 is a flowchart showing details of processing for expanding a sentence.

FIG. 56 is an explanatory diagram showing an example of a conventional source program generation method.

[Explanation of symbols]

210: functional specification, 220: tabular specification, 221: conditional part, 222: processing part, 223: conditional statement, 224 ...
Processing statement, 430: source program, 1120: tree structure format specification, 2000: source creation system, 2010
... Specification file, 2020 ... Sentence information file, 203
0: Term information file, 2040: Source file, 2
202: control type, 2203: source generation rule, 302
0 ... directed graph.

────────────────────────────────────────────────── ─── Continuation of the front page (56) References JP-A-3-286226 (JP, A) JP-A-4-140834 (JP, A) JP-A 5-265765 (JP, A) Transactions, Vol. J71-D, No. 12 (1988-12) P. 2555-2562 Minoru Harada, "All about CASE" Ohmsha (1991) 441-443 (58) Field surveyed (Int. Cl. ⁷ , DB name) G06F 9/06

Claims (2)

(57) [Claims] 1. A sentence in which a user setting portion is blank.
And the type data of the program control structure corresponding to the statement
And multiple sets of sentence information with one source generation rule set
Sentence information file and program in bulleted sentence format
An input device for inputting the description data of the functional specification.
For example, a source program generation method in source <br/> over scan program generation system for generating a source program corresponding to the stored program specifications data to program specifications file previously stored Kibun information file a plurality of sets A first processing step of reading the sentence information of each sentence and inputting the content required by the user in the blank portion of each sentence information from the input device; and a plurality of sentences in which the content required by the user is input in the blank portion. A second processing step in which the combination is stored in the program specification file as description data of the program function specification in a bulleted sentence format, and a description of the program function specification stored in the program specification file by the second processing step Third process of reading data and creating a sentence information node representing each sentence in the data A fourth processing step of determining a program control structure from steps and type data of a program control structure of each sentence and creating a directed graph representing a control relationship between each sentence information node; and a source generation rule of the directed graph and each sentence. And a fifth processing step of creating a source program corresponding to the program function specification description data on the basis of the above. 2. A program stored in a program specification file.
Source program corresponding to the program specification data
The source program generation system to generate, and the sentence where the user setting part is blank corresponds to the sentence
Type control data and source generation rules
Sentence information file that stores a plurality of sets of sentence information with one set of rules
And the blank using the sentence information stored in this sentence information file.
Combining multiple sentences that set the contents required by the user
Function specifications in the form of bulleted sentences composed
Descriptive data of the source program generation system
From the input device of the
Means for store the data in the program specification file, said program specification file to a program stored in
Reads the specification description data and displays each sentence in the data.
Create a sentence information node and
The program control structure is determined from the type data of the ram control structure.
And a directed graph representing the control relationship between each sentence information node.
Based on the directed graph creation means to be created, based on the created directed graph and the source generation rules for each sentence
Source code corresponding to the data of the program specification
Source generation means for generating a program.
Source program generation system.