JPH09204301A - Program generation system - Google Patents

Abstract

PROBLEM TO BE SOLVED: To easily produce a software product of high quality by process management more efficient than conventional one. SOLUTION: This program generation system for automatically generating a desired source program by analyzing program specifications described based on unitarily managed many kinds of definition information and components and developing the definition information and the components in a pattern provided with a standard logic structure is provided with a new program generation part 16. The program generation part 16 automatically generates the desired source program 17a by developing rule definition components supplemented by description in the program specifications 14a in the pattern based on relation information 23a for making many kinds of the rule definition components 21 for defining rules relating to plural data items shared between programs and the respective kinds of the definition information respectively correspond to each other beforehand.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a program generation system, and more particularly to a program generation system for easily producing a high-quality software product with more efficient process management than before.

[0002]

2. Description of the Related Art Conventionally, in order to automate and improve the work process of software development, given parts (subroutine group), patterns (framework of program), definition specifications (files, records, databases, screens / forms) "Program synthesis" is performed to automatically generate a semi-finished product of a source program from a specification document that defines items and attributes) and a program specification document that is described based on definition information such as a data item dictionary. (Hitachi, “Software Engineering Workbench / SEW”
B3 COBOL Program Synthesis / Handling Guide / Document No. 3000-7-465-20 ”, pages 2-6, 18-31).

FIG. 4 is a diagram showing an example of a general configuration of a program generation system. In the figure, a component 11, a pattern 12, a definition specification 13, and a data item dictionary 15 are created and prepared in advance. When creating a source program, an operator first selects an optimum one for realizing desired processing from many types of parts 11, patterns 12, and definition specifications 13 and selects it from the program specifications 14
Write inside. Next, the program generation unit 16 determines that the component 11 and the pattern 1 specified in the program specification 14
2, the source program 17 based on the definition specification 13
Automatically generate semi-finished products. At this time, the definition specifications 13
If any definition is described in the data item dictionary 15 for the data item used therein, the definition is expanded to the source program 17 using the corresponding definition. Finally, the operator manually creates a portion that could not be automatically generated from the above-mentioned parts 11, patterns 12, definition specifications 13, and data item dictionary 15, and uses this as user-added coding in the source program 17. Supplemented to make a finished product. In the program generation system,
A target source program is generated according to the above-described work and processing flow.

The data item dictionary 15 in the above-described program generation system is configured so that the program generation unit 16 automatically performs the input item check process and the item edit process, which are always required to be supplemented by user-added coding in the old system, in the source program 17. It is intended to be able to be expanded.

In the input item check processing, it is determined whether or not the value of the input item given from the outside is valid. For example, when a value used as "number of people" or "number of pieces" in the program is input from the outside, if the value of the input item is a negative value that is not a natural number or an integer value including a decimal point, it is necessary to branch to error processing. Therefore, specific branch conditions and branch destinations of such input item check processing are registered in the data item dictionary 15 in advance, and the input item check processing is expanded in the definition specification 13 for each worker. Have only the name of the data item you want to have written. The program generation unit 16 refers to the data item dictionary 15 for the data items in the definition specification 13 and develops the coding of the corresponding input item check processing in the source program 17.

In the item editing process, a new data item is generated based on the given data item. For example, “m ² ” is generally used as a unit to express the area quantitatively, but when the object whose area is to be grasped is land, it is often exceptionally expressed in the unit of “tsubo”. , "Tsubo"
It is often necessary to convert the value of the land area given in units to the value in the unit of "m ² " in advance.
Therefore, similar to the above, the specific contents of the item editing process are registered in the data item dictionary 15 in advance, and each worker only has the name of the data item to which the item editing process is to be expanded in the definition specification 13. Let me describe. The program generation unit 16 refers to the data item dictionary 15 for the data items in the definition specification 13 and develops the coding of the corresponding item editing process in the source program 17.

[0007]

The data item dictionary 15 used in the conventional program generation system is
It is structured assuming that only rules for each single data item will be registered. That is, as in the above-mentioned example, the input item check process that only compares and determines one data item value and a certain numerical value can be registered, but an input item that requires a comparison and determination of a certain data item value and another data item. The check process cannot always be registered. For example, if an employee's "assignment date" and "hire date" are to be entered from the outside as input items, if "hire date" is later than "hire date", error handling will be performed. Need to branch. However, if the rule is related to a plurality of data items, the conventional system cannot perform such registration in the data item dictionary 15. Similarly, although the item edit process of simply obtaining another value from one data item value can be registered, the process of obtaining the third data item from the first data item value and the second data item can be registered. Can't always be registered. For example, the process of calculating the "sales amount" by multiplying the "sales quantity" and the "unit price" cannot be defined and registered in the data item dictionary 15 in the conventional system.

Therefore, the automatic generation by the program generation unit 16 cannot be applied to the rules relating to a plurality of types of data items, and as in the case of the conventional system, the worker always performs user addition coding on the incomplete source program 17. There was a problem that supplements had to be made. The quality of such user-added coding is easily influenced by the degree of proficiency of each worker, and a plurality of source programs 1 including user-added coding are included.
There is a problem in that it is difficult to perform process control so that the quality of all 7 reaches a certain level. Also, if it becomes necessary to modify the rules related to multiple types of data items due to specification changes or defect corrections, it is necessary to revise the user-added coding after specifying all the target source programs. However, there is a problem that it is difficult to efficiently perform such process control.

Therefore, it is an object of the present invention to solve the above problems and provide a program generation system capable of easily producing a high quality software product with more efficient process management than before. .

[0010]

In order to achieve the above object, the program generation system of the present invention stores various types of definition information such as a large number of centrally managed parts, patterns, definition specifications and data item dictionaries. The program specification described based on the above is analyzed, and the definition regarding the input / output, the definition regarding the data item, and the program transition condition specified by the definition specification, and the part have the standard logical structure in the pattern. In a program generation system for automatically generating a desired source program by expanding it into a predetermined number, a plurality of types of rule definition parts in which rules relating to a plurality of data items are defined and the various definition information are associated with each other in advance. Based on the information, the rule definition parts supplemented by the description in the program specification Expand the serial pattern in is obtained by a configuration having a program generating section for automatically generating a desired source program.

Therefore, by defining rules relating to a plurality of types of data items in advance as rule definition parts, the parts to be supplemented by the user-added coding can be greatly reduced compared to the conventional method, and many work steps can be performed. Can be limited to the description work of the program specification. Therefore, most of the target source programs can be automatically generated by the program generation unit, and the process control can be easily realized such that the quality of each source program reaches a certain level.

In addition to the above-mentioned configuration, a repository for collectively registering and managing all the related information in a unified manner is provided, and the program generation unit is a specific program described in the program specification. For each definition information item, a specific rule definition component corresponding to the definition information item is obtained by searching the repository. Therefore, if it is necessary to modify the rules related to multiple types of data items due to specification changes or defect correction, refer to the relevant information registered and managed in the repository, and then add the rule definition parts. By performing the automatic generation of the source program again by the program generation unit only for the program specifications that utilize the, it is possible to perform the process management more efficiently than before.

Further, in addition to the above configuration, the program generation unit replaces each of the data items in the definition information previously expressed by a tentative name with the designated name described in the program specification. After that, a desired source program is automatically generated.
As a result, the degree of freedom in designing the software product and the process control can be improved, and the process control described above can be performed more easily and efficiently.

As described above, in the program generation system of the present invention, it is possible to easily produce a high quality software product with more efficient process control than before.

[0015]

BEST MODE FOR CARRYING OUT THE INVENTION An embodiment of a program generation system of the present invention will be described in detail below with reference to the drawings.

FIG. 1 is a diagram showing a schematic configuration of an embodiment of a program generation system of the present invention. 4, the same components as those in FIG. 4 are denoted by the same reference numerals, and the description thereof will be omitted. Further, 21 is a rule definition component that respectively represents rules relating to a plurality of types of data items that are shared between individually generated programs, 22 is a related definition unit that defines related information to be described later, and 23 is related information and A repository for centrally registering and managing rule definition parts, 23a is related information directly managed by the repository, 2
Reference numeral 4 is a part of the source program generated from the rule definition part described in the program specification. Hereinafter, FIG.
A brief description will be given of a case where a program is automatically generated by the system.

First, various data items that may be shared between the individually generated programs are represented by pseudo names, and the pseudo names are used to define rules between a plurality of data items. Many types of rule definition parts 2
Prepare 1 in advance. Then, the relation definition part 22
Defines the related information 23a for associating the rule definition component 21 with the definition specification 13, and stores the related information 23a.
To be stored. The operator describes the general part 11, the pattern 12, and the definition specification 13 which have been conventionally given in the program specification 14a. The program generation unit 16 reads out the processing registered in the data item dictionary 15 for the data items in the general part 11, the pattern 12, and the definition specification 13 described in the given program specification 14a. Next, the related information 23a in the repository 23 is searched based on the name of the definition specification 13, the associated rule definition component 21 is read, and the source program 17a is automatically generated. At this time, there are sufficient types of rule definition parts 21 suitable for the target program.
If the related information 23a and the related information 23a are prepared in advance, the operator can appropriately write the program specification 14a to greatly improve the completeness of the source program 17a, and the user-added coding after the automatic generation. Reduces the need for workers to manually create and supplement

FIG. 2 is a diagram for specifically explaining the processing by the program generation unit in the system of FIG. In the figure, the same components as those in FIG.
The description is omitted. An operator in charge of developing a software product, for example, specifies a temporary input / output specification "A :? New file" as one of the definition specifications and develops rule definition parts related to the temporary input / output specifications. The position designation “PARTS01 A” is described in the pattern 12.
The temporary input / output specification is a temporary name to be written instead of directly writing the corresponding input / output specification name in the pattern 12.
Pattern 1 is created by associating with the actual input / output specification in the program specification 14a created later.
This is for improving the reusability of No. 2. Further, with respect to the data items that may be shared between the individually generated programs, a large number of types of rule definition parts 21 that define the rules between the respective data items are prepared. Also in the rule definition part, the reusability is improved by describing such a data item with a tentative name. The temporary name is described as “@ PREFIX # [Temporary I / O specification name] [Shared data item name] @ SUFFIX # [Temporary I / O specification name]”. In addition, in the related information 23a, various definition specifications and rule definition parts are associated with each other.

The program generation unit 16 includes a pattern 12,
When the source program 17a is expanded based on the program specification 14a, rule definition part 1, related information 23a, etc., '@ PREFIX # [temporary input / output specification name]' and '@ SUF
FIX # [temporary input / output specification name] 'is replaced with the prefix and suffix specified in the program specification 14a to generate the complemented rule definition part. In the example of FIG. 2, in the description in the program specification 14a, the provisional input / output specification name "? New file" is the actual input / output specification name "Definition specification A", the prefix "P1" and the suffix ". S1 ”is associated, so it is complemented as“ @ PREFIX # ADATAX @ SUFFIX # A ”→“ P1-DATAX-S1 ”“ @ PREFIX # ADATAY @ SUFFIX # A ”→“ P1-DATAY-S1 ” The defined rule definition component 1 is the pattern 12
It is expanded to the specified expansion position “PARTS01 A” inside.

FIG. 3 is a diagram showing an example of a detailed configuration of the repository in the system of FIG. In the figure, various rule definition parts 21 that define rules between data items that may be shared between the individually created programs 17a, and related information 23a ₁ that associates the definition specifications 13 with each, Related information 23a ₂ for associating programs 17a and define specifications 13 centrally registered and managed in the repository 23 shown in FIG. Therefore, by specifying the rule definition parts to be modified to change any of the rules, all the definition specifications and programs affected by the rule change can be specified. That is, in the example of FIG. 3, when the rule definition component 1 to be modified is specified, the definition specification A that is affected is specified, and the programs P1 and P2 that are affected are also specified. Therefore, when the rule definition parts are revised due to the change of the rule, the rule is generated by re-creating the source program automatically by the program generation section only for the program specifications associated with the related information. The result of the change can be reflected in the program almost automatically.

[0021]

As described in detail above, according to the program generation system of the present invention, most of the target source programs can be automatically generated by the program generation section, and the quality of each source program is kept at a constant level. The effect that the process control that reaches can be easily realized is obtained. Further, even if the specification is changed or the defect is corrected, it is sufficient to automatically generate the source program by the program generation unit only for the related program specifications, so that the process management can be performed more efficiently than before. Is obtained. Therefore, it is possible to easily produce a high-quality software product with more efficient process management than before.

[Brief description of drawings]

FIG. 1 is a diagram showing a schematic configuration of an embodiment of a program generation system of the present invention.

FIG. 2 is a diagram for specifically explaining a process by a program generation unit in the system of FIG.

FIG. 3 is a diagram showing an example of a detailed configuration of a repository in the system of FIG.

FIG. 4 is a diagram showing an example of a general configuration of a program generation system.

[Explanation of symbols]

11 parts (general parts) 12 patterns 13 definition specifications 14, 14a program specifications 15 data item dictionary 16 program generator 17, 17a source program 21 rule definition parts 22 related definition part 23 repositories 23a, 23a ₁ , 23a ₂ related information

Claims (3)

[Claims] 1. A program specification written on the basis of various definition information such as a large number of types of parts, patterns, definition specifications, and data item dictionaries that are centrally managed is analyzed and designated by the definition specifications. Input / output definition,
A program generation system that automatically generates a desired source program by expanding the definition of data items, program transition conditions, and the parts in the pattern having a standard logical structure, and relates to a plurality of data items. The rule definition parts complemented by the description in the program specification are developed in the pattern based on predetermined related information that individually associates a large number of types of rule definition parts in which rules are defined with the various definition information. A program generation system characterized by comprising a program generation section for automatically generating a desired source program. 2. A repository for collectively registering and managing all the related information in a centralized manner is provided, and the program generation unit, for each specific definition information described in the program specification, The program generation system according to claim 1, wherein a specific rule definition component corresponding to the definition information is obtained by searching the repository. 3. The program generation unit replaces each of the data items in the definition information previously expressed by a tentative name with a designated name described in the program specification, and then a desired source program is created. The program generation system according to claim 1, wherein the program generation system is automatically generated.