WO2006038303A1 - Software creating method - Google Patents

Abstract

A word-unit program constituting a system and adapted for crating output data is processed with a minimum number of implementations by eliminating useless repetition. Even a path action element is assumed as one word, and a requirement is defined as a word-unit program where the relation between words is defined. By placing the condition of implementation of the definition formula of the path action element in the condition of implementation of the definition formula of a word belonging to the basic structure specified by the path action element when the path action element definition formula implementation condition is met, the path action element is eliminated, and the structures of the whole system are integrated into one structure. The word-unit programs (not including the path action element) of the thus unified whole system are topologically sorted to re-arrange the word-unit programs in an optimal order. With this, for example, useless repetition can be avoided.

Description

 Software generation method

 Technical field

 TECHNICAL FIELD [0001] The present invention relates to a software generation method based on a mathematical concept, and more particularly to a software generation method using an adjacency matrix and a topological 'sort in Lyee methodology.

 Background art

[0002] (1) Issues in software development

 Making high-quality software easily and quickly is a fundamental concern in the field of software development research. Various methodologies and techniques have been devised and proposed to improve the productivity, maintainability, and quality of software development, but none of these issues have been solved essentially. One reason is that the software itself is complex and difficult to catch, and another reason is that current methodologies are limited. In fact, almost all the proposed methodologies are not only clearly understood and fail to produce modifiable systems, but they are also used by professionals with a very wide range of capabilities, skills and knowledge. It is still considered impossible. As a result, labor costs and maintenance costs are high, and extensive checks are required to run on software. The summary and issues of the conventional software development method are summarized as follows.

 (2) Procedural programming

 In general, procedural programming has been effective in business system development and implementation because of the ease of handling of its description language. However, there are still important issues such as difficulty in accurately implementing requirements, and improvement in productivity and maintainability.

 (3) Declarative programming

 On the other hand, declarative programming, in which programming is performed by declarations that are requirement definitions themselves, includes artificial intelligence, expert systems, logical reasoning, inference, user information gathering, and user intelligence gathering.

) And other non-computational application fields. Declarative programming Requirement defined in the control procedure Declares the relationship between defined data and programming is performed based on the declaration, so the advantage is different from procedural programming that requires the procedure such as the control procedure to be explicitly specified. This means that requirements can be implemented accurately.

 [0003] However, declarative programming also has a problem that the efficiency of execution is not good due to the structure of execution control. Another disadvantage is that it is difficult to use a language for declarative programming. Languages for declarative programming include logical languages that use logical expressions (for example, PROLOG) and functional languages that describe target states as mathematical functions (for example, LISP). Declarative languages have a procedural logical mechanism for executing declarations, because values are not generated just by declaring them. The logical mechanisms of these procedures have a number of conditions in order to be able to execute all declarations, require knowledge of mathematics and logic, and are difficult to master. For this reason, it is not popular as a programming language. This is why procedural programming using a procedural language that has a structure similar to that of human natural language is becoming more prevalent. (4) Object-oriented programming

 Object-oriented programming is a programming technique in which data and procedures (called methods) that operate on it are combined as a unit called an object, and a program is written as a combination of objects. Since the program is composed of highly independent modules called objects, the scope of the impact of changes and modifications is limited, and there is an advantage that reuse can be reduced in units of objects.

 [0004] However, there is no strict rule on how to implement the requirements by what units of objects, and how the components are assembled. There are no restrictions. In actual development situations, the relationships between objects are not applied completely consistently, and in many cases a lot of objects are generated in the development artifacts, even if the creator itself is excluded. No one understands it. In other words, the completed software becomes a group of non-regular functions, and it tends to be difficult to finally disassemble and reuse.

Patent Document 1: Pamphlet of International Publication No. 97Z16784 Patent Document 2: Pamphlet of International Publication No. 98Z19232

 Patent Document 3: Pamphlet of International Publication No. 99Z49387

 Patent Document 4: Pamphlet of International Publication No.OOZ79385

 Patent Document 5: Pamphlet of International Publication No. 02Z42904

 Patent Document 6: International Publication No. 2004Z68342 Pamphlet

 Patent Document 7: JP 2002-202883 A

 Non-Patent Document 1: Dallin 'Winskel, “Official Semantics of Programming Languages”, MIT Press, 1993

 Non-Patent Document 2: Hanne 'Rees' Neilson, Fleming's Nielson, “Semantics of Application Semantics”, John “Wylie'and” Sands, 1992 Of the Information Society in the 21st Century (IS2000), Japan, Aizu, November 5-8, 2000, pp441-446

 Non-Patent Document 4: Fumio Negori, “High-Density Processing Method for Source Code Generation”, 5th World Multipolar Conference Report on Systemitas, Cybernetics and Informatics, (SCI2001), USA, July 22-25, 2001

 Disclosure of the invention

 Problems to be solved by the invention

[0005] (l) Lyee Development Methodology

 Recently, a new and very promising method called Lyee (governmentaL methodologY for so satwarE providencE §¾ · tailed. Read as “Lee. The inventor is Fumio Nerai.”) Was proposed. Lyee is a new way to automatically develop software with requirements as well.

[0006] In the case of Lyee, although it is declarative, it does not require a special logic mechanism for executing the declaration, and execution of the declaration is a procedure of assigning a value as simple as procedural programming. Therefore, the language to be used may be a procedural language that is widely used. In other words, Lyee has the declarative advantage of being able to implement the requirements correctly and the procedural advantage of being able to use a procedural language that is easy to describe. [0007] For an overview of the Lyee development method, reference is made to Japanese Patent Application No. 2004-272400 and Patent Document 6 by the applicant of the present application, which are incorporated in the disclosure. An outline is given below.

 (2) What is the declaration in Lyee?

 In Lyee, the object of the abstraction that you declare is a variable. Declared variables are called words in Lyee. In general, if a specification is declarative, the unit of abstraction in which it is declared excludes the order of description, and it is expressed as an equation that has the right-hand side power and the bidirectionality from the left-hand side to the right-hand side. That is. In Lyee, words have no power bi-directionality that eliminates the ordering of descriptions, which simplifies the framework.

[0008] The main elements of a word declaration are a word name, a definition formula, a calculation condition for the definition formula, input / output attributes, a value attribute, and the like. A definition expression is a relational expression that indicates a relationship with another word. For example, the word a can be declared as the defining expression a = b + c. And the execution of this declaration only executes the relational expression as an assignment expression. For example, to execute the declaration a = b + c of the word a (to find the value of a), just assign a value to the variables b and c on the right side.

 (3) Mechanism for executing declarations in a procedural framework

 It is the program structure of the fixed structure provided by Lyee that achieves the elimination of the order of declaration descriptions while implementing it in the framework of a procedural language. As long as the computer that processes the program is a Neumann type that processes instructions sequentially, if it is a procedural language, the ordering problem must be solved in some way. Other declarative programming is solved by a declarative language. Lyee, which uses a procedural language, is solved by the program structure! /

 [0009] The program structure is composed of fixed-format modules (hereinafter referred to as "declaration execution modules") that execute word-by-word declarations, and these modules are aggregated in units called palettes. Yes. The execution order of declaration execution modules in the palette can be any array. Execution control module power in the pallet Controls execution so that execution of the declaration execution module is repeated until execution of all the declaration execution modules in the pallet is completed (ie, values can be generated).

[0010] The palette consists of the W02 palette for executing the input word declaration, the W03 palette for executing the calculation condition in the output word declaration, and the definition expression execution and output in the output word declaration. There is a W04 palette to execute. These three types of pallets form a set, forming a collective unit called the basic structure. In other words, the Lyee structure program is divided into basic structure units, the basic structure is divided into three types of palettes, and the palettes are divided into declaration execution modules. The basic structure is created in units of output logic (ie, groups of words that are simultaneously output to the same medium). This division structure of the Lyee program is illustrated as a processing path diagram (PRD:

 Process Route Diagram; fcOo

[0011] The entire program by Lyee is composed of a fixed-structure declaration execution module that executes word-by-word declarations, and the method of configuring the entire program with these declaration execution modules is also standardized. Therefore, it is possible to automatically generate a program by substituting the declaration of requirements in units of words into this fixed structure declaration execution module.

 [0012] Actually, automatic generation of a program by Lyee is realized by a tool called "LyeeALL". In other words, LyeeALL automatically assigns the requirement declarations to a given blank in the code of the boilerplate declaration execution module (called a “template”. Template is a template). It is a tool that is constructed according to a predetermined method and builds one program.

 [0013] In software development using the Lyee methodology, engineers are divided into two groups, each performing two specialized tasks. One team is familiar with technologies related to the external environment in which the Lyee structure program runs, such as the OS, middleware, and database, and sets parameters for each external environment in the template of the Lyee structure program. The other team defines the user requirements and declares the requirements according to the Lyee methodology to implement the requirements in the Lyee structured program template. Since this kind of division of labor can be developed, the development of Lyee methodology can also solve the problem of shortage of human resources with such skills, which does not require all programmers to have technologies related to a wide range of external environments.

 (4) Issues with conventional Lyee structure

As mentioned above, there are significant requirements such as accurate implementation of requirements, automation of code generation, and ease of maintenance. This is a Lyee methodology that has many advantages, but there are also problems with its repetition and division structure. It is said that the program size is large due to the division structure with slow execution speed due to the repeated processing.

 On the other hand, in Lyee, requirements are divided into units called basic structures to capture specifications. Understand the clear information about the execution order such as screen transitions, database key relationships, etc., using word duplication as a unit, and grasp the execution conditions of words belonging to the group. In the current Lyee structure, this split structure is coded as it is, so the program size is large and the execution speed is also slow.

 [0015] Improving the Lyee structure to an efficient processing speed program that is smaller in size than unnecessary iterations, while maintaining its declarative programming benefits without changing the requirement semantics. Is an important issue.

 [0016] As a method to improve the execution speed by eliminating unnecessary repetition, the topological sort automatically optimizes the placement order of declaration execution modules (declaration execution is completed in one round) Ordering) method has already been proposed (PCTZJP03Z09591). You can topologically sort the modules in the palette to eliminate wasted repetition in the palette.

 [0017] However, if the topological sort is performed with the Lyee division structure as it is, the effect is limited. This is because there is repetition between pallets and basic structures, and the optimal ordering by topological 'sorting within the pallet minimizes the number of repetitions, but generation of all values is not completed in one cycle. . In addition, the split structure itself is one factor that increases the program size.

 [0018] Therefore, in order to further improve the problem of the size and processing speed of a Lyee program having a repetition and division structure, it is necessary to eliminate the division unit of the palette and the basic structure and implement it as a single program. It seems to be useful to perform topological 'sorting' on modules that execute the range declaration.

The present invention relates to a method of performing topological sort on the entire program by eliminating the division unit of the conventional Lyee structure program. This aims to improve software execution speed and program size with the Lyee methodology. Means for solving the problem

 [0020] In order to solve the problem, the present invention provides a user requirement to be implemented as one program for each logical body with an access condition, for each word on the logical body, a word name, a definition formula, From the word unit declaration declared by the execution condition, input / output attribute, and word value attribute of the definition expression, all L2 processing (input word attribute check processing) and L processing (output) required for satisfying the requirements are performed. The first step that defines the execution unit of one of the declarations of word value generation processing, 12 processing (logical body input processing), and 04 processing (logical body output processing), and all the L2 defined above Define (semi) order relations of processing (input word attribute check processing), L processing (output word value generation processing), 12 processing (logical input processing), 04 processing (logical output processing) Specified by the step and the (semi) order relationship defined by the step The L2 process, the L process, the 12 process, and the 04 process are subjected to a third step of performing topological sort, and according to the sequence of execution units of the declarations rearranged by the third step. And a fourth step of arranging a given code sequence based on the Lyee methodology corresponding to the declaration execution unit.

 [0021] According to the present invention having a powerful configuration, the word unit declaration defined in the first step is subjected to L2 processing (input word attribute check processing), L processing (output word value generation processing). ), 12 processing (Logical body input processing), 04 processing (Logical body output processing), the program to be executed in the shortest order is created, and software generation using Lyee methodology becomes more efficient. In other words, all the functions of these steps can be automated, and eventually automatic processing can be performed consistently from the elimination of repeated processing to software generation, and the software production speed and efficiency are improved in each stage.

[0022] Here, “all L2 processing (input word attribute check processing), L processing (output word value generation processing), 12 processing (logical input processing), 04 processing (logical output processing)” “Defining (semi-) order relation” means expressing the mutual order relation in a directed graph, focusing on the words used in the declaration. As the order relationship, a partial order and a full order can be considered. `` Directed graph '' means each declaration execution unit included in the requirements issued by the user who is a software development requester by a node, and it is a symbol that means that a≤b (a force b is created. Define “≤”. same as below. ) Is a partial order relation (or full order relation) between the start and end points. For example, software in which such a function is programmed may be used, or a code obtained by coding the software may be written in a ROM (Read Only Memory).

 [0023] In addition, "topological 'sorting" is a method in which a depth-first search (described later) is performed on the directed graph obtained above, and nodes are acquired when the force search path is reached. This is a function that arranges node sequences (numerical sequences). For example, it may be software that has programmed such functions, or may be software that has been coded into ROM (Read Only Memory). In this topological sort, the directed graph obtained above is represented by an adjacency matrix, and a depth-first search is performed on this to generate a zero-throw matrix.

 Here, “depth-first search” means

 Depart from the starting node and visit nodes that are connected to each other so that visits do not overlap,

 When there is no place to go, go back to the node where the place is going (recursion),

 Proceed in the same way (1) from the returned node,

 End when there is no more way forward,

 This is an algorithm.

[0025] In addition to the above-mentioned depth-first search, “search” includes breadth-first search, iterative deeping, heuristic search, Various algorithms including Hill Climbing, Best- First Search, Euler stroke, Dijkstra method and so on. For example, it may be software that programs such functions, or may be software that has been coded in ROM (ReadOnlyMemory).

 [0026] "Software" used here has a broad meaning. In other words, it is a concept that includes both development requests from users and target programs that meet the needs.

[0027] By providing the configuration as described above, in the present invention, since repetition is avoided, higher Software development with high efficiency, high speed and high performance can be realized. Furthermore, preprocessing that eliminates repetition is automatically performed, and a target program is automatically generated by the Lyee (registered trademark) methodology based on the preprocessed declaration execution unit program obtained in this way. In other words, it is possible to automate everything from the refinement (arrangement) of user requirements to the generation of the target program. This will have a significant effect on the software industry, such as significantly improving the efficiency of software production, improving productivity, and improving quality.

[0028] A program (software), a program generation device, a program processing device, and a tool (including both as a device and software) for producing "software to be developed" as different embodiments of the present invention The software development device, the software development support device, and the software development management device, the user requirements to be implemented as a single program, for each logical body with access conditions, for each word on the logical body, word name, definition formula, From the execution condition of the definition expression, input / output attributes, and word unit declarations declared by the word value attribute, all L2 processing (input word attribute check processing) and L processing ( Output word value generation process), 12 process (logical body input process), 04 process (logical body output process) All of the specified L2 processing (input word attribute check processing), L processing (output word value generation processing), 12 processing (logical input processing), 04 processing (logical output processing) (half) Means for defining an order relationship, means for performing a topological sort on the L2 processing, L processing, 12 processing, and 04 processing defined by the (semi-) order relationship defined by the step, and the third And a means for arranging a given code string based on the Lyee methodology corresponding to the declaration execution unit according to the sequence of the execution units of the declarations rearranged by the above steps.

[0029] The present invention further includes software produced by the above-described "method for producing software to be developed" and a recording medium on which the software is installed or an apparatus on which the software is installed. In this case, the present invention implements the user requirements implemented as one program for each logical body with access conditions, for each word on the logical body, a word name, a definition formula, Declaration power in units of words declared by execution conditions of definition formulas, input / output attributes, and word value attributes Everything necessary to satisfy the requirements The execution units of declarations of L2 processing (input word attribute check processing), L processing (output word value generation processing), 12 processing (logical input processing), and 04 processing (logical output processing) are single words. It is constructed as a given code sequence based on the relevant Lyee methodology according to the sequence sequence sorted by the topological sort performed on the (semi) order relationship that also defines the requirement power declared in the You can also.

 [0030] Furthermore, the present invention relates to software as a model of software code used for producing software by the above-mentioned "method for producing software to be developed", and a recording medium on which the software is mounted. Alternatively, it can be realized as a device (hardware) on which the software is installed.

 [0031] Further, the present invention is an extraction method of information (document (paper, data)) extracted from the requirements by the above-mentioned "method for producing software to be developed", and extraction by a powerful extraction method. Information (document (paper, data)), as a method of using the extracted information, or as an information recording medium on which such information is mounted, or as a method of extracting information Z It may be realized as information extracted from the software development requirements, including coded software and related information that can be realized as a recording medium Z device (hardware) on which the software is installed.

 [0032] Regarding the logical body, refer to Japanese Patent Application No. 2004-272400 by the same applicant, which is cited and made a part of the disclosure.

 The invention's effect

 [0033] According to the present invention, the processing of the declaration execution module for generating output data in the system program that also has the word-by-word declaration execution module power is completed with the minimum number of executions by eliminating unnecessary repetition. In addition, the program size can be reduced. More specifically, the following points are possible.

 [0034] From the declarations in which the program requirements are declared in word units, it is possible to determine the word execution units to be executed and to define the partial order relationship between the word execution units.

[0035] A word-based declaration execution unit, for example, a path action element, which becomes unnecessary by the optimal ordering is deleted. The order information that must be implemented to maintain the meaning of the requirement. The execution condition of the road action element can retain the meaning of the requirement by reflecting it in the definition of the partial order relation of the declaration execution unit of other word units.

 [0036] Topological sorting is performed on the word-wise declaration execution unit module group (not including the path action element) obtained in this way, and the word-unit declaration execution unit group is rearranged in the optimal order. This avoids unnecessary repetition and allows the program to be executed with the minimum number of executions.

 BEST MODE FOR CARRYING OUT THE INVENTION

[0037] The present invention improves the above-mentioned problems of the prior art by eliminating the division unit of the conventional Lyee structure program and performing topological sort on the entire program. According to the present invention, improvement in the execution speed of software developed by the Lyee methodology and improvement in program size are realized. Hereinafter, specific embodiments of the present invention will be described with reference to the drawings. For the technique of topological 'sorting for Lyee structure program, reference is made to Patent Document 6 by the same applicant, which is a part of the disclosure.

 [0038] Chapter 1 explains the outline and outline of the division structure of the Lyee software that is the object of integration of the present invention. Chapter 2 describes a specific method for integrating the basic structure. Section 3 describes examples of programs that integrate basic structures using the present invention and their effects. Chapter 1 Structure of Lyee software division structure>

In order to integrate a conventional Lyee structure program with a divided structure, it is first necessary to understand how the divided structure works. The division into basic structures is declared based on the requirement that the word should be output in what group and under what conditions. There is an input / output attribute item in the element of the word unit declaration. In addition to the input / output distinction, what input or output group (logical body) each word belongs to, and what conditions the input / output group has The power to be input / output in is also a requirement to be declared. For details on how to accurately extract the requirement information related to this input / output from the requirement definitions representing various processes, refer to Japanese Patent Application No. 2004-272400 by the same applicant, which is a part of the disclosure. . Here, only the outline considered necessary for the explanation of the present invention will be described with respect to the meaning of the division into the basic structure. 1. Meaning of division into basic structures

Meaning of division structure

 The division structure of Lyee software is a structure in which declaration execution modules, which are the smallest unit modules, are grouped according to certain rules. The structure is shown in Fig. 1 (a) and (b).

Sync range

 The modules that make up the program are first grouped into units shown as “synchronization range” in Fig. 1 (a). The synchronization range is a set of requirements defined by the computer to be executed by an event (for example, a command such as pressing a button or selecting a menu) that is a process execution instruction to the program by the user. It is the scope of processing. In other words, a synchronization range is a group of modules that are executed using the same event as an execution condition. In Lyee, this synchronization range processing is called “synchronization”. Completing the synchronization range processing means completing the generation and output of all output words specified in the requirements for the event.

 This structure

 Next, modules within the synchronization range are grouped into basic structural units as shown at 101 in FIG. 1 (a). The input / output processing of a computer is performed on a set of data rather than on a data unit. A set of words that are simultaneously input or output to the same definition is called a logical entity in Lyee. A logical unit is a unit that holds word values such as screens, files, forms, and messages. The basic structure is a grouping of modules involved in the same output logic. The role of the program is to generate and output the value of the output word specified by the event by the user, but the output to be performed within the synchronization range by one event is not necessarily one. In many cases, two or more output logics within one synchronization range, such as saving (outputting) the generation results to a file, in addition to displaying (outputting) the word value generation results on the screen Field is output, and the synchronization range is grouped into multiple basic structures.

Knowlet

Finally, the modules in the basic structure are grouped into three groups by type. Basic A group of modules in a structure is called a palette in Lyee. The three pallets are called W02 pallet, W03 pallet, and W04 pallet, respectively. The W02 palette is a collection of modules related to input word input. The W03 palette is a set of modules related to the determination of the calculation condition of the output word definition formula. The W04 palette is a set of modules related to the generation of values by the execution of output word definition formulas and the output of values.

Next, the execution order of the basic structure will be described.

 Within the same synchronization range

 There is an end point starting point relationship between the basic structures. Therefore, the basic structure to be executed next in the depth-first search is determined. If the topological sort is executed for the declaration execution modules within the same synchronization range, the execution order is determined.

 Execution order of synchronization range

 Which sync range is executed depends on which event is fired by the user. For example, there are a button for executing process A and a button for executing process B on the screen, and the contents of the process executed by the program when the user presses one of these buttons (that is, the synchronization range) ) Is determined. In this way, the entire program is composed of a plurality of synchronization ranges with different events as execution conditions, so it is not possible to simply perform topological sort on the declaration execution module of the entire program. Chapter 2 explains how to perform topological sorting by integrating the declaration execution modules with different synchronization ranges, using a specification with two synchronization ranges due to two events as an example.

2. Topological sort target

 The target of optimal ordering by topological sort is the processing executed by the declaration execution module that is a component of Lyee software. There are two types of declaration execution modules, one is a logical element and is a module for establishing the value of a word. The other is an action element, a module that performs processing other than value establishment, such as input / output processing.

[0040] Patent document 6 by the same applicant treats logical elements that establish values as elements of a directed graph and an adjacency matrix, expresses them by a mathematical model, and implements a topological sort. Embodiments are disclosed. Action elements can also be handled as elements of the directed graph and adjacency matrix that are the same as logical elements, and can be the target of topological sort. The only major difference between a logical element and an action element is that the logic element acts on only one word (acting as a value generation), but the action element acts on two or more words at the same time. For example, a path action element (which specifies the basic structure to be executed next) operates on a set of words called the basic structure, and an output action element (outputs an output word) is a word called an output logic. Act on sets.

[0041] In the following, a method for subjecting topological sorts will be described while considering the declaration execution modules sequentially.

 "Logical elements"

 "L2"

 L2 is a declaration execution module that checks whether the attribute of the input word meets the requirements. Under any condition, the condition under which L2 processing can be executed is that the input word's value / medium power is also taken into the memory area (already entered). Therefore, it can be considered that the L2 processing that is the target of topological sorting is the input 12 of the L2 value at the start point.

 [0042] Event commands (buttons and menus) are also treated as one of the input words.

[0043] “: L3 and L4”

 In the current Lyee software structure, the output word generation processing is realized by two modules, L3 (execution of calculation condition expressions) and L4 (execution of definition expressions). Furthermore, when the output word a has a plurality of definition expressions (the word in such a case is called an equivalent word), L3 and L4 are provided as many as the number of combinations of the definition expression and the calculation condition.

[0044] Assume that the output word a has a plurality of defining expressions (the word in such a case is called an equivalent word) and is declared as follows.

[0045] <Definition formula> <Calculation conditions>

 (1) b + c e> 10

 (2) b-d e≤10

Equivalent words have multiple definitions, but only one value is assigned to the output word. There must be. Therefore, the calculation conditions for the definition of equivalent words must be exclusive and complete.

 [0046] Under any conditions, the condition for generating the word a is sufficient if all the definition words of the word a and all the start words used in the calculation condition are determined. Yes. In other words, the topological sort target only needs to execute the definition expressions for the words b, c, d, and e first.

 [0047] Therefore, the topological sort targets are all of the L3 and L4 processing pairs for each output word, and the starting point is the input word L2 used as the starting point in their definition and calculation conditions. It is a set of processing and output word L4 and L3 processing.

 `` Functional element ''

 `` Input action element 12 ''

 Reference numeral 12 denotes a module for performing (input) processing for fetching values from the medium into the memory. Therefore, the condition under which 12 processes can be executed under any condition is unconditional (that is, it does not depend on other processes). The 12 processes subject to topological sort do not have a starting point. (However, in the case of input from a file, the value of the key word, which is the key for accessing the file, must be established)

 "Output action element 04"

 04 is a module that writes (outputs) word values to the output buffer in the Lyee software domain memory. Therefore, the condition under which the process 04 can be executed under any condition is that the value of the output word to be output has been established (that is, one L4 has been executed). The 04 process that is the target of the topological sort can be regarded as the output word that is the output target.

[0048] "Structure action element S4"

S4 initializes the area to record the word value and the area to record the processing result (for example, output completed successfully, output not completed normally) (record the initial value) It is a module that performs processing. In Lyee with a conventional repeat structure, overwriting prohibition and initialization of word value areas were means for the repeat structure mechanism. Therefore, the topological sort is performed on the entire program, and the Lyee structure is optimally ordered. Therefore, it is not necessary to initialize the word value area.

[0049] On the other hand, the initialization of the processing result recording area is not a process for the mechanism of the repetitive structure! /, So it is also necessary for the integrated Lyee structure.

[0050] "Path action element"

 The path action element is a module that performs processing for designating a palette to be executed next. The path action element plays a role of linking the pallet which is a division unit of the program and the basic structure which is a set of the pallet.

[0051] <Route action element that performs transition between palettes in the basic structure>

 <Route action element that makes transition between basic structures>

 The meaning of the path action element that specifies the execution of another basic structure from the basic structure is explained below with an example.

 (Example 1)

 Fig. 2 shows a system in which a user obtains a value for data item g by inputting data into data item c and data item d and pressing a or b of the execute button on a screen of a certain system. 701 is an input data field of data item c, 702 is an input data field of data item d, 703 is an output data field of data item g, 704 is execute button a, and 705 is execute button b. When the execution button a is pressed, the value of g is calculated by the definition expression c + d, and when the execution button b is pressed, the value of g is calculated by the definition expression cXd. Since the execution buttons a and b are not pressed at the same time, it is a condition that one of the two defining expressions for generating the value of g is executed.

 [0052] The program of the screen in FIG. 2 is represented in FIG. 3 as divided units of the basic structure. BS1, BS2, and BS3 in Fig. 3 are basic structures. Basic structure BS1 is the basic structure of the output logic to the screen shown in Fig.2. BS1 includes an input word a and an input word b that are execution buttons, an input word b and an input word c that are input data items, an output word g that is an output data item, and a path action element X.

[0053] The basic structure BS2 is a basic structure (medium is a file) for calculating the value of the output data item g by the expression c + d when the execute button a is pressed. Log the result of the expression c + d. It contains the word e. Basic structure BS3 is output data item g when execute button b is pressed Is the basic structure (medium is a file) for calculating the value of x by the formula c X d. It contains the word f that records the calculation result of the expression c X d.

 [0054] After the user enters data on the screen (BS1 is executed), the basic structure for calculating the value of the output word g is BS2 or BS 3 Specifies which force should be executed It is the pathway action element X that does. The execution of the words belonging to the two basic structures depends on whether the button a or b is pressed, so that the path action element X can be used when the execution button a is pressed (word a = true, value ) Specifies BS2, and if execution button b is pressed (word b = true, has value). Specifies BS3 as the next basic structure to execute. Also, since execution moves to the next basic structure when input data is ready, the fact that word c and word d have a value is also an execution condition for path action element X.

 [0055] By executing the basic structure BS2 or BS3, the value of the output word g is the word e or the word. Basic structure for outputting output data to the screen The output word g of BS1 is g = e when a value is established for the word e (e = true), and when a value is established for the word f (f = true) It is defined as g = f.

 [0056] The definitions of the above words are summarized in Table 1 below.

[0057] [Table 1]

Control module

 The conventional Lyee structure has a division structure, and iterative processing is performed for each division unit. Two types of control modules were required: a palette chain function (one program) that calls and executes the palette function itself.

After the topological sort, the entire program is integrated and sequenced, so it is sufficient if there is one control module that collects the modules in the program! /. Replacement paper (Rule 26) 2. Basic structure integration and topological sorting

 Using the requirements of Example 1, the specific procedure for topological sorting is explained.

 Adjacency matrix of basic structure

 FIG. 4 is a diagram showing the relationship of the words in FIG. 3 with a directed graph. Words a, b, and d, which are input words belonging to the basic structure BS1, are the end points of the word network represented by the directed graph. Similarly, as shown in Table 1 above, the word X, which is a path action element belonging to the basic structure BS1, must be either the word a or the word Since b, word c, and word d are needed, they are linked to the four words by arrows starting from these four words. The word e belonging to the basic structure B S2 and the word f belonging to the basic structure BS3 require the word c, the word d, and the word X in order for the value to be satisfied. Each arrow is linked to three words. Since the output word g belonging to the basic structure BS1 requires the word e or the intermediate word f in order for the value to be established, it is linked to the two words by arrows starting from these two words.

Next, the system of Example 1 shown by the directed graph in FIG. 4 is represented using an adjacency matrix. First, the adjacency matrix F1 that represents the basic structure BS1 (901) that handles the input word in Fig. 4 is shown in Equation 1 below.

 Adjacency matrix of basic structure BS1 (input word)

[0060] [Equation 1]

abcdx _r

 β `` 0 0 0 0 0

 ft 0 0 0 0 0

 FJ = c 0 0 0 0 0

 0 0 0 0 0

x _r [l 1 1 1 0

Since the elements of BS1 (input word) are word a, word b, word word d, and word ^, the matrix of F1 is composed of these elements. Since the input word word a, word b, and word word d do not have a start word, each row of end words a, b, c, d (start word a, b, c, d, x “Intersection points” are all “0” indicating that they are not used as the start point words. As shown in the directed graph in Fig. 4, word X, which is a route creation element, has word a, word b, Since it has the word d, the row of the end word X _f is “1” indicating that the intersection with these start words is the start point.

 [0061] Next, the adjacency matrix F2 representing the combination of the basic structures BS2 and BS3 (hereinafter referred to as the basic structures BS2 & BS3) is shown in Equation 2 below.

 Adjacency matrix of basic structure BS 2 & BS 3

[0062] [Equation 2]

 e f

Since the elements of BS2 & BS3 are only word e and word f, the matrix of F2 is composed of these two elements. Since the word e and the word f are independent from each other (that is, both words use the other word as the starting word), the misalignment intersection is also “0”.

 [0063] The adjacency matrix F3 of the set that handles the output words of the basic structure BS1 is shown in the following Equation 3.

 Adjacency matrix of basic structure BS1 (output word)

 [0064] [Equation 3]

 8

 F3 = g [0]

 Since the element of BS3 is only word g, the matrix of F3 is composed of only gl words. The word g does not use itself as the starting point, so the intersection is “0”.

 As described above, in the basic structural unit constituting the system of Example 1, the relationship between elements in the same basic structure is represented in the adjacency matrix.

 2. Integration of basic structure by joint adjacency matrix

 Next, in order to represent them as a single adjacency matrix for the entire system, a combined adjacency matrix that links the basic structures by showing the relationship between the words that are the elements of each basic structure and the words that are the elements of other basic structures (Connection matrix) will be described.

 [0066] The relationship between a word having an arbitrary basic structure and a word having another basic structure is represented by the relationship between the end point word and the start point word as shown in FIG. Basic structures can be said to be linked by the relationship of the words belonging to them. Figure 4 shows the following.

When the word of basic structure BS1 (input) is the end word, any other basic structure word Since it appears after the word BS1 (input), it cannot be the starting word.

 BS1 appears before the word of BS2 & 3 when the word of the basic structure BS2 & 3 is the end word.

(Input) word can be the starting word BS1 (Output) word cannot be the starting word Basic structure When the BS1 (output) word is the end word, either BS2 & 3 or BS1 (input) Since the word also appears before the BS1 (output) word, it can be the starting word.

[0067] As a result of the above consideration, what is important as a joint adjacency matrix for linking basic structures is a joint adjacency that represents the relationship between an arbitrary basic structure and a basic structure that includes a word that can be the starting word of a word belonging to it. It turns out that it is a matrix. Because, as shown in the directed graph, there is only a directional relationship from the starting word to the ending word for the relationship between words, so the relationship between the basic structures is also changed from the basic structure including the starting word to the ending word. There is only a relationship toward the basic structure including, and there is no reverse relationship (the joint adjacency matrix that represents the relationship between such basic structures has all intersection points of 0).

[0068] The following considers one connected adjacency matrix that links the basic structure of Example 1 (representing a directional relationship from the basic structure including the start word to the basic structure including the end word).

[0069] Equation 4 is a joint adjacency matrix FC1 indicating the relationship between the basic structure BS1 (input) (start word) force and the basic structure BS2 & BS3 (end word).

[0070] drawings

abcdx _r

 e `` 0 0 1 1 Γ

 FC1 =

/ | _0 0 1 1 1 The adjacency matrix FC1 is the word a, word b, word c, word that are elements of BS1 (input) when word e and word f, which are elements of the basic structure BS2 & BS3, are placed at the end word. d, word x _f represents the state used as the starting word, and shows the relationship between the two basic structures. As shown in Fig. 4, word e and word f use word c, word d, and word X as the starting word. Therefore, in FC1, the intersection of the line of end word e and end word f and start point words c, d, and X is “1”, and the intersection point with other start point words is “0”.

Equation 5 represents the adjacency matrix FC2 indicating the relationship from the basic structure BS1 (input) and the basic structure BS2 & BS3 (start point word) to the basic structure BS1 (output) (end point word). [0072] [Equation 5]

 a b c d e f

 FC2 = g [0 0 0 0 1 l]

 The adjacency matrix FC2 is the word a, word b, word word that is an element of BS1 (input), BS2, BS3 that appears before in the directed graph when the word g, which is an element of the basic structure BSl (output), is placed at the end word. d, word X word e, and word f are used as starting words, and the relationship between the two groups is shown. As shown in Fig. 4, word g uses word e and word f as the starting word. Therefore, in FC2, the intersection of the line of the end point word g and the start point words e and f is “1”, and the intersection point of the other start point words is “0”.

 [0073] The above is the joint adjacency matrix that is meaningful for the links between the basic structures constituting the system of Example 1. To simplify the description of the process of integrating the entire system into one adjacency matrix, the following is also shown the adjacency matrix between basic structures that do not have a relationship. Equation 6 is a joint adjacency matrix FC3 indicating that there is no relationship between the basic structure BS2 & BS3 (start word) and the basic structure BS1 (input) (end word).

 [0074] [Equation 6]

 Equation 7 is a connection adjacency matrix FC4 indicating that there is no relationship between the basic structure BSl (output) (start word) and the basic structure BSl (input) (end word).

[0075] [Equation 7]

Equation 8 is a joint adjacency matrix FC5 indicating that there is no relationship between the basic structure BS1 (output) (start word) force basic structure BS2 & BS3 (end word).

[Equation 8]

 S

 1

Using the adjacency matrices F1 to F3 and the combined adjacency matrices FC1 to FC5 shown above, the system of Example 1 is expressed as one adjacency matrix F.

[Equation 9]

abcdx _r efg

In the above formula, with Fl, F2, F3, FC1, FC2 and 0 shown on the right side of “F =” (adjacent matrix with all matrix intersections 0, corresponding to FC3, FC4, FC5) The constructed matrix shows how the basic structural unit adjacency matrix and the combined adjacency matrix are integrated to represent the entire system. Next, the power of changing those elements into words is an adjacency matrix whose elements are a, b, c, d, x, e, f, and g on the right side. Fl, F2, F3, FC1, FC2, FC3, FC4, FC5 It is the power of others. 5 1001 repulsive force SF1, 1002 force SF2, 1003 force F3, 1004 force FC1, 1005 are FC2. A joint adjacency matrix with all matrix intersections of 0 has FC3 force 006, FC4 force 1007, and FC5 force 1008.

[0078] As described above, one system composed of two or more basic structural forces can be represented by one adjacency matrix.

 [0079] Since the entire system can be represented by one adjacency matrix as described above, the topological sort disclosed in Patent Document 6 by the same applicant is performed on the adjacency matrix of this system, and all words are displayed. The program is rearranged in the optimal order so that the state of the value of is fixed with the minimum number of executions (ie, the generation of all output words has been completed).

 [0080] Looking at the adjacency matrix F of the system of Example 1, as shown in Fig. 6, all the intersections with the starting word whose value is 1 are in the lower left triangle (1301). As mentioned in Chapter 1, this means that the words are in the optimal order. Therefore, the adjacency matrix F in Equation 9 has already been topologically sorted and is in the order of words that can complete the generation of output words with the minimum number of executions, given the initial word state vector. It can be said.

 Here, a supplementary explanation will be given on the effect of topological sorting. A word-by-word program arranged in the optimal execution order by topological sorting can generate all output word values in one round. However, when referring to the entire system, it is appropriate that it can be realized more accurately with the “minimum number of executions”. This is because the system as a whole may include processes that require repeated use of the same formula by placing a temporary recording area, such as tabulation processing. Of course, “repetition” in this case is not useless repetition. The number of data to be minus minus 1 is the minimum number of repetitions. Such an aggregation process can also be expressed in one adjacency matrix using a combined adjacency matrix.

It should be noted that here, as an example of the search technique, the power and other search techniques described using the topological sort as an example may be used. In addition, the adjacency matrix definition is essentially an indispensable step in this paper to help ease understanding. In other words, it is within the scope of the idea of the present invention to reach a search such as topological 'sorting' directly from the definition of the relationship between declarations without defining adjacent matrices or through directed graph creation. It can realize the intended purpose of Ming.

 3. Verification of program adjacency matrix

 This section functions as a function that generates the value of the adjacency matrix F force output word of the system shown in the previous section, and the force is also topologically sorted, so all output words are generated with the minimum number of executions. Is verified by a calculation operation that multiplies the state vector of the word disclosed in Patent Document 6 by the same applicant. If the value of the output word g is determined, the adjacency matrix F can be said to function as a function. In addition, since the system of Example 5 does not include the repetitive processing such as the above-described aggregation processing, it can be said that the processing is completed with the minimum number of executions if the state of the value of the output word g is determined in one round processing.

 [0083] In the system of Example 1, the state of the value of the word before the user inputs, that is, the initial value of the state vector X of the word is all elements as disclosed in Patent Document 6 by the same applicant. Is “null (undecided)”, and it becomes like the following number 10.

 [0084] [Equation 10] a

 b null

 c null

 d

 null

 e null

 f null

 s noil Performs a calculation process that multiplies the adjacency matrix F of number 9 (ie, the system of Example 1) by the state vector X of number 10 (ie, inputs to the system of Example 1). The state of the value of the word that is changed by the FX calculation process will be explained for each terminal word.

 (1) Calculation of end word a during the first round

 The calculation of the end word a during the first round is as follows.

 a = 0-null + 0-null + 0-null + 0-null + 0-null + 0 · null + 0 · null + 0-null

 = (+1) + (+ 1) + (+ 1) + (+ 1) + (+ 1) + (+ 1)

+ (+1) + (+ 1) = (+ 1)

 Since the input word a has no starting word, the value state is fixed regardless of the value state, and changes from nuU to (+1). In other words, this indicates that the value has been determined by the input. As a result, the state of the value of each word after completion of the word a line during one round is as shown in Table 2 below.

[Table 2]

(2) Calculation of end point single feb during the first round processing

 The calculation of the end point word b during the first round process is as follows because it uses the state of the word value in Table 2.

b = 0-(+ l) + 0-null + 0 -null + 0-null + 0-null + 0 · null + 0 · null + 0 · null

 = (+ 1) + (+ 1) + (+ 1) + (+ 1) + (+ 1)

+ (+1) + (+ 1) + (+ 1)

 = (+ 1)

 Since the input word b also has no starting word, the value state is fixed regardless of the value state, and changes from null to (+1). As a result, the state of the value of each word after completion of the word b line during one round is as shown in Table 3 below.

[Table 3]

 In the system of Example 1, word a and word b are alternative buttons on the screen, so when one of them is selected, the state of both values is confirmed. (3) Calculation of end word c during the first round

 The calculation of the end point word c during the first round process is as follows because it uses the state of the word value in Table 3.

c = 0-(+ l) +0-(+ l) + 0-null + 0-null + 0-null + 0-null + 0-null + 0-null

 = (+ 1) + (+ 1) + (+ 1) + (+ 1) + (+ 1)

+ (+1) + (+ 1) + (+ 1)

Replacement paper (Rule 26) = (+ 1)

 Since the input word c does not have a starting word, the value state is fixed regardless of the value state and changes to null force (+1). As a result, the state of the value of each word after completion of the word c line in one round is as shown in Table 4 below.

[Table 4]

 (4) Calculation of end point word d during the first round processing

 The calculation of the end point word d during the first round process is as follows because the state of the word value in Table 4 is used.

d = 0 (+1) + 0 (+1) + 0 (+1) + 0-null + 0-null + 0-null + 0-null + 0-null

 = (+ 1) + (+ 1) + (+ 1) + (+ 1) + (+ 1)

+ (+1) + (+ 1) + (+ 1)

 = (+ 1)

 Since the input word d also has no starting word, the value state is fixed regardless of the value state, and changes from nuU to (+1). As a result, the state of the value of each word after completion of the word d line during one round is as shown in Table 5 below.

[Table 5]

(5) Calculation of end point word during the first round

 The calculation of the end-point word ^ during the first round process is as follows, using the word value states in Table 5.

 = 1 · (+1) + 1 · (+1) + 1 · (+1) + 1 · (+1) + 0 · null + 0 · null + 0 · null + 0-null

 = (+1) + (+ 1) + (+ 1) + (+ 1) + (+ 1)

+ (+1) + (+ 1) + (+ 1)

 = (+1)

 The state of the value of the word a b c d that is the starting word of the word ^ has already been determined, that is, the definition formula execution conditions (1) and (2) for the word X that is the path action element shown in Table 1 are satisfied.

Replacement paper (thin IJ26) As a result, the value state of the word ^ also changes from null to the confirmed (+1). As a result, the state of the value of each word after completion of the word _x line during one round processing is as shown in Table 6 below.

[Table 6]

(6) Calculation of end word e during the first round

 The calculation of the end point word e during the first round process is as follows because it uses the state of the word values in Table 6.

e = 0 '(+ 1) + 0 · (+1) + 1 · (+1)' -t-l-(+ l) + l-(+ l) + 0'null + 0-null + 0-null

 = (+1) + (+ 1) + (+ 1) + (+ 1) + (+ 1)

+ (+1) + (+ 1) + (+ 1)

 = (+ 1)

 Since the state of the values of the words cd and X that are the starting words of the word e have already been determined, that is, the values of the starting words of the definition expression c + d have been determined, and the path action element X executes next. Since the basic structure is specified, the value of the word e changes from null to confirmed (+1). As a result, the state of the habit of each word after completion of the word e line during one round is shown in Table 7 below.

[]

 (7) Calculation of end word f during the first round

 The calculation of the end-point word f during the first round process is as follows because it uses the word value states in Table 7.

f = 0 · (+1) + 0 · (+1) + 1 · (+1) + l '(+ l) + l' (+ l) +0 '(+ l) + 0'mill + 0' null

 = (+ 1) + (+ 1) + (+ 1) + (+ 1) + (+ 1)

+ (+1) + (+ 1) + (+ 1)

 = (+ 1)

 Since the values of the words c d and X that are the starting words of the word f have already been determined, the value of the word f also changes from null to (+1). As a result, the word f line being processed in one round is completed.

Difference paper (mm The status of each word after completion is shown in Table 8 below.

[0091] [Table 8]

In the system of Example 1, the execution conditions of word e and word f are alternative conditions, so if one of the conditions is determined to be true, the other condition is determined to be false. The state of the value will be confirmed. For example, if BS2 is specified, the definition expression execution condition of word e = true is confirmed and the value state is determined by giving a value to word _e . on the other hand

The condition of the value of the word f is determined when the definition formula execution condition = false is confirmed and no value is given to the word f. Therefore, when the path action element specifies one of the basic structures, the values of both word e and word f are already determined.

 (8) Calculation of end point word g during the first round processing

 The calculation of the end point word g during the first round process uses the word value states in Table 8 and is as follows.

 g = 0 · (+1) + 0 · (+1) + 0 · (+1) +0-(+ l) +0-(+ l) + l-(+ l) + l-(+ l) + 0-null

 = (+ 1) + (+ 1) + (+ 1) + (+ 1) + (+ 1)

 + (+1) + (+ 1) + (+ 1)

 = (+ 1)

 Since the state of the values of words e and f, which are the starting words of word g, have already been determined, the state of the value of word g also changes to null force (+1). As a result, the state of the value of each word after completion of the word g line during one round is as shown in Table 9 below.

[Table 9]

 This completes the calculation of all the word lines of JF (one word in JF. Therefore, Table 9 is the result of FX. The state vector X in Equation 11 below is the result of FX.

[0093] [Equation 11]

Replacement bowl (Rule 2

 As described above, since the adjacency matrix F is arranged so that the word unit programs are executed in the optimal order by topological sorting, the state of the value of all words is determined by executing one round of calculation processing. The value of the output word g could be established.

 4. Elimination of pathway action elements

 The execution condition of the path action element is equal to the condition for executing the defining expression of the set of all words (one or more powers) belonging to the basic structure specified by the path action element. This is because all the words of the basic structure are executed when the condition of the path action element is satisfied. Therefore, the execution condition of the path action element can be shifted to the condition of the word belonging to the basic structure specified by the path action element.

 [0094] The case of the system of Example 1 is specifically described as follows. First, Table 10 shows the definition formula and definition formula execution conditions for word X, word e, and word f, which are path action elements.

[0095] [Table 10]

 Therefore, the condition of X (1) “a = true and c = true and d = true” is specified when this condition is satisfied. The execution condition “Xr specifies BS2” for the word e belonging to the basic structure BS2. Replace condition (2) Replace “b = true, c = true and d == true” with the execution condition “Xr force ¾S3 specified” for the word f belonging to the basic structure BS3 specified when this condition is satisfied. It turns out that. As a result, the definition formula and definition formula execution conditions for word e and word f are as shown in Table 11 below.

Differences (Rule 26) Become.

[0096] [Table 11]

As shown in Table 11, if the execution condition of the path action element is replaced with the execution condition of the word of the designated basic structure, the link between BS1, BS2, and BS3 is established by the execution condition of the word of BS2 and BS3. The path action element simply indicates that the condition to move to the next basic structure is ready (the input rate word c: ^ and d have a value and the button a or b has been pressed). It only has meaning. This means that the meaning of the system does not change even if the path action element is removed. Therefore, after performing the operation to replace the execution condition of the path action element with the execution condition of the basic structure word specified by the establishment of the condition, the path action element can be removed from the adjacency matrix of the entire system.

[0097] FIG. 7 is a directed graph showing the relationship of words in the system of Example 1 with the path action element removed. In FIG. 7, the word X, which is a path action element, disappears, and the relationship between the word e and the word f and the starting word is shown according to the definition formula and the definition formula execution condition defined in Table 11 above. ing. That is, the word e is a simple word “a” (definition formula execution condition “execution button a is pressed”), word c and word d (definition formula “c + d” and definition formula execution condition “ c = true and d = true "). The word f is the starting point word b (definition formula execution condition “execution button b was pressed”), word c and word d (definition expression “c + d” and definition expression execution condition “ _c = true and d” = True "). Word e and word f force The relationship to word g is the same as in Figure 4.

 [0098] When the system of Example 1 with the path action element removed is represented in the adjacency matrix according to the directed graph of Fig. 7, the following is obtained. F1 'in Equation 12 is the adjacency matrix of the basic structure BS1 (input). The word action element X has been deleted.

 [0099] [Equation 12]

Replacement paper it & ca

F2 in Equation 13 is the adjacency matrix of basic structure BS2 & BS3 _c

[0100] [Equation 13]

^ /

F3 'in Equation 14 is the adjacency matrix of basic structure BS1 (output).

[0101] [Equation 14]

 S

FC1 'in Equation 15 is a combined adjacency matrix indicating the relationship from basic structure BS1 (input) to basic structure BS2 & BS3. Since the path action element has been deleted, the word X is deleted as the element of the starting word. As shown in Table 11, word e and word f have new word a and word b as starting words, respectively, by changing the execution conditions of the path action element. Therefore, the end point word e has an intersection point with the word a in addition to the start point words c and d. The end word f has an intersection point with the word b in addition to the start words c and d.

[0102] [Equation 15]

 a b c d

 e f l 0 1 f

 FCl '=

f [0 1 1 1 FC2 of number 16 is a connected adjacency matrix that shows the relationship from basic structure BS2 & BS3 to basic structure BS1 (output). Since the path action element has been deleted, the word as the element of the starting word It has been deleted.

[0103] [Equation 16] a h € ά e f

F € 2 ^f = ^> Q 0 1 l] FC3 in Equation 17 is a combined adjacency matrix indicating that there is no relationship from basic structure BS2 & BS3 to basic structure BS1 (input). Since the path action element has been deleted, the word X is deleted as the element of the starting word.

[0104] [Equation 17] e /

 0 0

 b 0 0

 war c u

 d 4 0 FC4 'of number 18 is a joint adjacency matrix indicating that there is no relationship from basic structure BS1 (output) to basic structure BS1 (input). Since the path action element has been deleted, the word X is deleted as the element of the starting word.

[0105] [Equation 18] g

 a 0

 b 0

 0

 d 0 FC5 'of number 19 is a joint adjacency matrix indicating that there is no relationship from basic structure BS1 (output) to basic structure BS2 & BS3.

[0106] [Equation 19] S

The adjacency matrix F of the system can be expressed by the adjacency matrix F ′ shown in the following equation 20 when the condition of the path action element is replaced and the path action element is deleted.

 [Equation 20]

 a b c d e f g

 In the above formula, Fl ', F2', F3 ', FC1', FC2 'and 0 (the adjacency matrix where all matrix intersections are 0, FC3', FC4 (Corresponding to ', FC5') shows how the basic structural unit adjacency matrix and combined adjacency matrix are integrated to represent the entire system. Next, it is an adjacency matrix whose elements are a, b, c, d, e, f, and g on the right side. Fl ', F2', F3 ', FC 1', FC2 ', FC3', FC4 ', FC5' are integrated to form F ', and even more power ^) 8. In FIG. 8, 1201 has a proportional force F1 ', 1202 force SF2', 1203 is F3 ', 1204 is FC1', and 1205 is FC2 '. A joint adjacency matrix with all matrix intersections of 0 constitutes FC3, force 206, FC4 'force 207, same 5, and ^ 208.

 [0108] As described above, one system having two or more basic structural forces can be represented by one adjacency matrix by removing the path action element.

[0109] Next, topological sorting is performed on the adjacency matrix of this system so that the state of all word values is determined with the minimum number of executions (that is, generation of all output words is completed). Sort programs by word in the optimal order. Looking at the adjacency matrix F ', as shown in Fig. 9, all the intersections with the starting word whose value is 1 are in the lower left triangle (1601). As shown in the same applicant's patent document 5, this is the word in the optimal order Means that. Therefore, the adjacency matrix F ′ of the number 20 has already been topologically sorted, and if an initial value word state vector is given, the order of the words that can complete the generation of the output word with the minimum number of executions becomes. It can be said that.

 5.Verification of the structure excluding the path action element

 In this section, the adjacency matrix F ′, which shows the structure of the system excluding the path action element, shown in the previous section, has the same function function as the adjacency matrix F, which is the structure before eliminating the path action element. To verify. The process for obtaining the value of the output word g by giving the word state vector to the adjacency matrix F 'is shown below. As a result of the m-th calculation, if all the values of the state state vector X of the word are confirmed, F ′ can be said to be a function that functions in the same way as F. It is also verified that the generation of output words can be completed with the minimum number of executions by performing topological sorting. It can be said that F 'which is the system of Example 1 could be executed with the minimum number of executions if the state of the value of the word g is determined in one round.

[0110] As described above, since all elements of the initial value of the word state vector are “null (undecided)”, the state vector X of the word to be given to F ′ first is Like 21.

[0111] [Equation 21]

 Apply the calculation processing to multiply the adjacency matrix F in Equation 20 (ie, the system in Example 1) by the state vector X in Equation 20 (ie, input to the system in Example 1). The state of the value of the word that is changed by the calculation process of F'X 'will be explained for each terminal word.

 F 'Calculation of end word a during the first round

 Since the end word a uses the state of the value of the word X, in the calculation process, the calculation of the end word a during the first round process is as follows.

a = 0-null + 0-null + 0-null + 0-null + 0-null + 0 · null + 0 · null = (+ 1) + (+ 1) + (+ 1) + (+ 1) + (+ 1) '

+ (+1) + (+ 1)

 = (+ 1)

 Since the input word a has no starting word, the state of the value is fixed and changes from null to (+1). That is, this indicates that the value has been determined by the input. As a result, the state of the value of each word after completion of the word a line during one round is as shown in Table 12 below.

[Table 12]

(2) F 'Calculation of end word b during the first round process

 Since the end point word b uses the value state of the word in Table 12 for the calculation process, the end point word b during the first round process is calculated as follows.

b = 0-(+ l) + 0-null + 0-null + 0-null + 0-null -I- 0-null + 0-null

 = (+ 1) + (+ 1) + (+ 1) + (+ 1)

+ (+1) + (+ 1) + (+ 1)

 = (+ 1)

 Since the input word b does not have a starting word, the value state is fixed and the nuU force changes to (+1). As a result, the state of the value of each word after completion of the word a line during one cycle is as shown in Table 13 below.

[Table 13]

(3) F 'Calculation of end word c during the first round process

 Since the end point word c uses the state of the word value in Table 13 in the calculation process, the calculation of the end point word c during the first round process is as follows.

c = 0-(+ l) + 0-(+ l) + 0-null + 0 · null + 0 · null + 0-null + 0-null

 = (+ 1) + (+ 1) + (+ 1) + (+ 1)

 + (+ 1) + (+ 1) + (+ 1)

 = (+ 1)

Replacement paper (Rule 26) Since the input word c also has no starting word, the state of the value is fixed and changes from null to (+1). As a result, the state of the value of each word after completion of the word c line during one round is as shown in Table 14 below.

[Table U]

(4) F 'Calculation of end word d during the first round process

 Calculation processing of the end point word d uses the value state of the word in Table 14. The calculation of the end point word d during the first round processing is as follows.

 d = 0 · (+1) + 0 · (+1) + 0 · (+1) Η-0-null + O-null + O-null + O-null

 = (+ 1) + (+ 1) + (+ 1) + (+ 1)

 + (+1) + (+ 1) + (+ 1)

 = (+ 1)

 Since the input word d also has no starting word, the state of the value is confirmed and changes to null force (+1). As a result, the state of the value of each word after completion of the word d line during one round is shown in Table 15 below.

[Table 15]

(5) F 'Calculation of end word e during the first round processing

Since the end point word e uses the value state of the word in Table 15 for the calculation process, the calculation of the end point word _e during the first round process is as follows.

 e = 1 (+1) +0 (+1) +1 (+1) -t- l-(+ l) -t-0-nulH-0-null + 0-null

 = (+ 1) + (+ 1) + (+ 1) + (+ 1)

 + (+1) + (+ 1) + (+ 1)

 = (+ 1)

Since the values of the words a, c, and d, which are the starting words of the word e, are fixed, both the value of the word e and the nuU force are changed to (+1) and fixed. As a result, the state of the value of each word after completion of the word e line during one round is as shown in Table 16 below. Replacement paper (Rule 26) [Table 16]

 (6) F 'Calculation of end word f during the first round processing

 Since the end point word f uses the state of the word value in Table 16 for the calculation process, the calculation of the end point word f during the first round process is as follows.

f = 0 (+1) + 1 (+1) + 1 (+1) + l-(+ l) +0-(+ l) + 0-null + 0-null

 = (+ 1) + (+ 1) + (+ 1) + (+ 1)

+ (+1) + (+ 1) + (+ 1)

 = (+ 1)

 Since the values of the words b, c, and d, which are the starting words of the word f, are fixed, the value of the word f is also changed from null to (+1) and fixed. As a result, the state of the value of each word after completion of the word f line during one round is as shown in Table 17 below.

[Table 17]

 (7) F, total of end word g during the first round processing

 Since the end point word g uses the state of the word value in Table 17 in the calculation process, the end point word g during the first round process is calculated as follows.

g = 0 '(+ l) +0' (+ l) +0 '(+ l) +0' (+ l) + l '(+ l) + l' (+ l) + 0'null

 = (+ 1) + (+ 1) + (+ 1) + (+ 1)

 + (+1) + (+ 1) + (+ 1)

 = (+ 1)

 Since the values of the words e and f, which are the starting words of the word g, are fixed, the value of the word g is also changed to null force (+1) and fixed. As a result, the state of the value of each word after completion of the word g line during one round is as shown in Table 18 below.

[Table 18]

This completes one round of calculation processing for all word lines in the adjacency matrix F '. Therefore, Table 1 Replacement bowl (Rule 26) The state of the 8 word value results in F'X '. X

 Is the result of

[0119] [Equation 22]

As described above, the adjacency matrix F ′ is arranged in such a way that the word unit programs are executed in the optimal order by topological sorting! /, So all word values are determined by executing one round of calculation processing. And the value of the output word g was established.

 [0120] Also, the adjacency matrix F 'excluding the word of the path action element also gives the initial value of the state vector (input), so that the value of all the words is the same as the adjacency matrix F. It can be seen that the state is fixed and the value of the output word g can be generated. Therefore, it was verified that the adjacency matrix F, which is a system with a path action element, and the adjacency matrix F ', which is a system without a path action element, have equivalent processing functions as a system.

 [0121] As described above, when the entire system has a plurality of basic structures, these are integrated, and a search (for example, topological sort) is performed on the integrated system. , Lyee (registered trademark), etc., provide performance at each stage when applying software autonomous development technology. For example, if there are a plurality of basic structures as they are without being integrated, performance as high as that of the invention of the present application cannot be obtained, and the search technique such as topological sort has been used in order. In that case, the performance of the invention according to the present application cannot be obtained! /.

 6. Reuse of programs

 Processing that requires program reuse is repeated even when topological sorting is performed.

Any program (algorithm) can be written only in a recursive way . Any program written in a procedural programming language (eg C, java, etc.) can be written in a while program that plays the same role as a recursive program. This shows that the while program can theoretically do anything. However, the word-based programming method cannot use itself as the starting word. For example, a = a + b

 It cannot be used in a word-based program. This is because the value of a does not change once it is decided. Therefore, for recursive use, an area that temporarily holds the value of a and a clearing method are required. Use the stored value as the starting point. For this purpose it is necessary to have an area where content is temporarily held. This area is called a temporary word. A structuring element (Vector of Clear) is defined to clear.

 In Lyee, the action element is generally a kind of word with a predicate structure. The structure of the program code is the same as a word. The main difference from a word is that it handles two or more words at the same time. The meaning is to clear the path and area for the external interface and control structure.

[0122] The program's adjacency matrix (the entire program's matrix) is merged into a single matrix, using a split adjacency matrix and a connection matrix that only moves state. The connection matrix indicates that words from other groups are used as the starting word. It is therefore possible to use a mathematical model of a word-based program.

 As a result, all normal algorithms that can be expressed in procedural languages can be expressed in word-based program structures.

[0123] Next, the integration of adjacency matrices by connection matrices will be described using (Example 2) which is an example of aggregation processing that requires temporary words such as a = a + b described above.

 (Example 2) The record X and X values of each word in the file are summed in order of the leading force.

 2 6

 , Total the results into words X and X.

 3 7

 [0124] In Lyee, the value once decided cannot be overwritten, so write x = x + x

 3 1 2

 Will be reused. X X

 1 is a word for substituting the previous value of 3. X

The value of 3 is transferred to the temporary area X for assignment to X at the next reuse. Call it a “temporary word” . Figure 2 shows Example 2 in a directed graph. Words X, X,

 1 2

 X is cleared when X is transferred to X. Word X, Word X, and Word X are simply

3 3 4 5 6 7

 Reuse of word X, word X, and word X.

 1 2 7

 [0125] The adjacency matrix (one path matrix) of the portion of the reused program in Figure 10 is

F1.

[0126] [Equation 23] 1,5 ^x 2fi ^x 3, l

 ^ 1,5 0 0 0

 Fl = Χ2, β 0 0 0 one path matrix

 X3 1 1 0

In order to reuse Fl, the connection matrices that integrate F1 and reused F1 are F2 and F3. These have the function of only transmitting the state. Each end point word indicates which word is used as the start point word.

 [Number 24]

F2 ₄ [θ 0 1]: connection matnx irom

「'― I

5 _& I 0 I: connection matrix from x ₄ to x _s> x ₆ ,

 Four. "

 F2 is a connection matrix of x 1, x 2, x force x. F3 is from X to X, X, X,

 2 3 Connection matrix. F2 is the output of x force, F3 is the action element for X force and input to X.

 4 5

 Understood. The adjacency matrix of the entire program that reuses F1 is F.

 [Equation 25]

χ \ ^χ 2 ^χ 3 ^χ 4 Χ5 ^χ β

entire system matrix

The adjacency matrix of the entire program is expressed using a circular matrix and a connection matrix. Gatsutsu It is possible to use a mathematical model of a word-based program. As a result, all the usual algorithms that can be expressed in a procedural language can be realized in that structure by word-based programming.

Chapter 3 Specific examples using the present invention

 <Mathematical algorithm for topological sort>

Before showing the concrete method, we will show an algorithm that topologically sorts an array with elements of partial ordered sets as elements.

 When Lyee's specification is expressed in terms of words, the specification in Lyee is expressed as follows: X = F (x) (Equation 2- 1-3)

Here, the functional F is defined as follows.

F (x) (i) = f (x (l), x (2), &fracl4;, x (n)), i = l, 2, &fracl4;, n

(Formula 2-1-2)

Ordered set (A,

A non-empty subset of B) is a fully ordered set under, B is called a chain, and B is called an anti-chain when any binary of B is incomparable. F is monotonous because one iteration will define more endpoints than the number of defined start points on the right side. Therefore, iterative sequences generated one after another by repeating F

[Equation 26]

(F ^k ) _{k N} is a chain. Where k is the number of iterations. The minimum number of iterations is the maximum chain size. In general, a partially ordered set can be divided into antichains by the following theorem. [theorem]

Partially ordered set (A,

) Where n is the largest chain size, A is divided into n disjoint antichains.

This theorem is proved by mathematical induction. Applying this theorem to all words that are partially ordered sets, we can say the following. That is, all words are divided into a set of n words that are not in any order. The length (size) n of the repetitive sequence (Equation 26) indicates the maximum number of chains that can be obtained, ie, the minimum number of iterations required before synchronization.

The proof of this theorem gives an algorithm that topologically sorts an array whose elements are partial order sets. That is, if An is the opposite chain, A1, A2, and A2 are obtained one after another until A = 0 from the input side, the elements of A are arranged in an appropriate order, and n + 1 n

After the elements of A in the proper order and finally the elements of A in the appropriate order, the ascending n-1 1

Topological sort can be done in order. Where A is the set of local maxima. Here, it is the set of words closest to the output side. The word arranged in this way is A

 If executed from 1 (input side), all the words are established one after another in a single process.

1. How to integrate split structures

The basic structure is one, and the words are rearranged. The program generated by this method has the following structure.

 1) The W04 logic element for value generation is placed in W03.

 2) A word that combines the W03 logic element and the W04 logic element. Equivalent words are combined into one word.

 3) Put all basic structural words on W04 on the start screen. At this time, the execution condition attached to the action element of the path is expanded to each word. Path action elements do not expand.

 4) For all words and input / output action elements, the relationship between the start point and basic elements

(Graph), and the topological sort and the order are determined and arranged by the above method. Parez It is not installed because it is not necessary to repeat it. Clearing the pallet area is also unnecessary and will not be installed.

 5) The Control Area flag and the input / output action element flag must be cleared and left as they are.

As a result,

 1) Words and input / output action elements that belonged to the same basic structure have the same execution conditions because the execution conditions for the path action elements of the basic structure are expanded equally. Create a group of groups.

 2) As a result, the basic elements are regrouped according to the route execution conditions, and the basic elements are arranged in the order of execution from the top.

 2.Specification of case

 The requirement of this example is an example of a student grade management system. In this system, the general manager can register the student's name, the teacher can register the score for each test of the student, and register the overall evaluation at the end of the semester. In this system, the initial screen shown in Fig. 11, the "Student Registration Screen" shown in Fig. 12 and the "Grade Management Screen" shown in Fig. 13, and the administrator and teacher IDs and passwords are registered. “ID & password file”, “Student name file” where the student's name and ID are registered! / “Student name file” where the score for each test is registered, It consists of four files: “Comprehensive evaluation file” that is registered at the end of the period.

The initial screen is a screen displayed when the grade management system is activated. Press the Exit button to exit the system. If you enter the ID and password fields and press the OK button, the system will refer to (read) the entered data and the data registered in the ID & password file, and the entered data will be sent to the general affairs staff. Whether it is a teacher's, a teacher's, or neither. If the result of the judgment is that the ID and password are for the administrator, the student registration screen is displayed. If the ID and password are for the teacher, the grade management screen is displayed. Otherwise, the initial screen is displayed.

The student registration screen is a screen in which the general affairs officer registers the student name and student ID number. ID and password determined on the initial screen At the time of display, the system refers to the student name file and displays the data of already registered students in the field of the student list. After the screen is displayed, enter the student name and student ID number in the new registration field and press the registration button. The entered data will be registered (written) in the student name file. Press the back button to return to the initial screen.

 The grade management screen is a screen for registering and referencing the score and overall evaluation for each student test. Displayed when the ID and password determined on the initial screen belong to the teacher. When the screen is displayed, the system refers to the student name file and prepares a list of student names and student ID numbers assigned by the accessing teacher as a selection candidate list that can be displayed in the student name text box. To do. When the general clerk presses the button at the right end of the text box to display the student name list and selects (enters) the student name and ID number that he / she wants to refer to, the system references Calculate and display student test results and average scores. If a comprehensive evaluation has already been registered, at the same time, refer to the comprehensive evaluation file and display the comprehensive evaluation. To register a new test result, enter the data of the selected student in the test result registration field and press the register button. The system registers new data in the test results file, and at the same time updates and displays the contents of the test result field including the new data, and recalculates and displays the average score. To register a comprehensive evaluation, enter the comprehensive evaluation of the selected student in the comprehensive evaluation registration field and press the register button. The system registers the comprehensive evaluation data in the comprehensive evaluation file, and at the same time displays the data registered in the comprehensive evaluation field. Press the back button to return to the initial screen.

Fig. 14 shows the requirements of this grade management program and the processing path diagram showing the structure when this program is divided into the conventional basic structure.

 Includes a summary calculation for partial repetition of the same process, such as displaying a list of grades by exam for one student, and calculating the average score. Covers processing.

 3. Example implementation procedure

The language used is C ++, only the screen part uses the C ++ function, and the rest is written only in the C grammar. Yes. Database 'Management system is Access 97 with DOA.

At the time of design, the requirements divided into 9 basic structures were implemented by integrating them into 1 basic structure program. The program was generated automatically using LyeeALL2 (Lyee structure program code automatic generation tool), and its templates (basic elements and control program templates) were modified for the basic structure integrated Lyee program. Figure 15 shows part of the integrated new program code. Line 7 of the new program is the execution condition of the path, that is, the execution condition in the process definition, and the basic elements are grouped again by this. W02

The basic elements that were in the palette are arranged in the sorted order. As for the word, only the generation formula remains like 11, 12, 13, 14 lines.

The Lyee structure program generated with the basic structure divided is compared with the new program based on the Lyee methodology as follows. The number of execution steps and the time are those for a program that requires the requirement of “selecting 7 students' ability and displaying the results of 5 exams”. CPU process time is a value that does not include database access time, and Total process time is the total time that includes database access time. The processing time was calculated as the difference between the start and end by adding a time stamp function to the program.

4. Results and evaluation

1) Results

[Table 19]

The decrease in executable code is mainly due to the fact that the word structure has become smaller and the basic structure has become one. The main inflection point is

 1. The predicate structure is about 21 powers. This is usually only a generation expression consisting of one line of assignment expression.

 2. The generation condition of W03 and the generation expression of W04 are integrated into one.

3. As for the action element, the action element of the path becomes one IF statement. Input / output elements Remains almost intact.

 4. 80% clearing action is removed.

 5. One basic structure requires about 24KB excluding words and action elements. In this case, nine basic structures become one.

 The decrease in the number of execution steps and the decrease in execution time is mainly due to the decrease in repetition. The main change is

 1. All pallets are repeated at least twice and the basic structure is repeated at least three times. This power is once.

 2. Program 'code is reduced to 31%.

 3. The total process time will not decrease so much. This is because the database access time is controlled.

 4. CPU time is dominant in real-time processing that does not use database processing, such as process control such as a building monitoring system. Therefore, the new method is very effective in the case of process control in terms of program size and processing time.

 2) Evaluation

The following can be said when compared with a program based on conventional structural programming. At present, conventional structural programming based on this specification remains a challenge, and the conventional method depends on the skill of the programmer, so it is limited to qualitative comparison.

 1) Program design

Word execution conditions can be viewed as execution conditions for process-defined actions, reducing design difficulties in traditional Lyee. The advantage of Lyee, declarative specification capture, is realized in the new program with process definitions and word definitions. Therefore, in the new program, the procedure for control is realized by the process definition, so no program design is required. The conventional method requires some design based on the specifications.

2) Program size

There is no code that is clearly increased in structure in the new program compared to the program using the conventional structural programming. 3) Execution speed

 In the new program, the number of iterations is limited to the minimum necessary number, so there is no reason for the execution speed to be slower than in the conventional program.

 Chapter 4 Conclusion

 At present, the program generated by the tool based on the Lyee methodology is a specification that has been divided at the time of design and expanded into program modules in a divided state. Certainly, utilizing information about the order clearly expressed in the requirements in the design will improve the accuracy of the design and help improve efficiency. However, Lyee-structured programs dynamically solve problems that can be solved statically before execution, so the code size is large and the processing speed is slow. There is room for improvement here. It is possible to generate a program without changing the meaning of the program and dividing it into the basic structure from the beginning, while taking advantage of the efficiency of design efficiency. By doing so, the whole word becomes one area. Therefore, when the correct order of words is determined in the topological sort on the directed graph created by the word relationship, it is not necessary to repeat the order to ensure the order (see Patent Document 6). Therefore, it is very useful to create a processing path diagram, that is, to divide the requirements and to grasp the system requirements accurately. Creating a processing path diagram means capturing words that have common elements in the output condition in logical units. Assigning common elements of conditions to route action elements has great advantages for efficiently defining user requirements.

[0134] On the other hand, at the time of generating and executing program code, it is useful to integrate a system divided into basic structures. This is because the execution order of the word-by-word programs in the entire system can be arranged in an order in which generation of output words can be completed with the minimum number of executions.

 In this specification, the processing of the declaration execution module for generating output data in a system that also has a word-level declaration execution module power is completed with the minimum number of executions without unnecessary repetition, and the program The issue of reducing the size was described as a solution by the following means.

1) A set of words (logical body) with the same output condition as a basic structure Capture the body with a processing path diagram. As a result, execution conditions common to execution of declarations belonging to the same basic structure (conditions for the path action element to specify the basic structure) can be accurately grasped.

 2) Path action elements are also subject to topological sorting, and word relationships are defined, and the same requirements as logical elements are defined by word relationships. In other words, it defines what conditions are satisfied (execution conditions of definition formulas) and which basic structure execution is specified (definition formulas).

 3) The route action element is eliminated by adding the execution condition of the definition expression of this route action element to the execution condition of the module belonging to the basic structure specified by the route action element when the condition is satisfied. By placing the execution condition of the route action element in the execution condition of the module of the designated basic structure, even if the route action element is deleted, a link between basic structures (that is, a link between words) is established. It is possible to combine the entire program into one.

 The declaration execution module group (not including the path action element) of the unified program obtained in this way is topologically sorted to rearrange the declaration execution module group in the optimal order. This avoids unnecessary repetition and allows the program to be executed with the minimum number of executions.

 Note that the present invention is not limited to the above-described embodiments and examples, and various modifications can be made within the scope of the technical idea of the present invention. For example, a business 'method, software development device, software development support device, software development management device, or software for realizing these functions on a computer and a recording medium equipped with the software' dedicated machine. Is possible. Further, as described above, the present invention can be applied to a method, software having a powerful function, a device on which the software is mounted (including software itself), a system Even so, it can be realized.

For example, FIG. 16 shows, as a different embodiment of the present invention, a program (software), a program generation device, a program processing device, a tool (as a device or a device) for producing “software to be developed”. Software development equipment, software FIG. 5 is a functional block diagram showing a configuration provided as a function when the present invention is implemented as either a software development support device or a software development management device. As shown in the figure, a program (software), a program generation device, a program processing device, a tool (including both as a device and software), a software development device, a software development support device, or The software development management device includes an overall control unit 1601, an execution unit definition unit 1602 for word unit declaration, a (semi) order relation definition unit 1603, a topological sort unit 1604, a code string arrangement unit 1605, and an information storage unit 1606. It is comprised with.

 The overall control unit 1601 is a program (software), a program generation device, a program processing device, a tool (including both as a device and software), a software development device, a software development support device, or a software development management device. It has functions to perform overall operation control, timing control, input / output control, etc., and a dedicated chip having this function, a dedicated circuit, or software for causing the computer to perform this function (as a tool) (Including software), or a recording medium on which the software is recorded, a processing device, a management device, or a tool on which the recording medium is mounted.

 [0139] The execution unit specifying unit 1602 for the word unit declaration has a function for specifying the execution unit for the word unit declaration as described above, and includes a dedicated chip, a dedicated circuit, Alternatively, it is realized as software (including software as a tool) for causing the computer to perform the function, or as a recording medium on which the software is recorded, a processing device / management device / tool on which the recording medium is mounted.

 [0140] The (semi) order relation definition unit 1603 performs all L2 processing (input word attribute check processing), L processing (output word value generation processing), 12 processing (logical input processing), 04 processing (logic For body output processing), it has a function to define the order relationship by expressing the mutual order relationship with a directed graph, focusing on the words used in the declaration. It is realized as a circuit or software (including software as a tool) for causing a computer to perform the function, or a recording medium storing the software, a processing device, a management device, or a tool equipped with the recording medium. The

[0141] The topological 'sorting unit 1604 is, for example, provided by the (semi-) order relation defining unit 1603. A depth-first search is performed on the directed graph, and the node sequence (number sequence) is arranged by acquiring the node when returning to the force search path when it arrives. A dedicated chip having a function, a dedicated circuit, or software (including software as a tool) for causing a computer to perform the function, or a recording medium on which the software is recorded, and a processing device equipped with the recording medium Management device · Realized as a tool.

[0142] The code string arrangement unit 1605 is a sequence of execution units of declarations rearranged by the topological 'sort unit 1604 (all L2 processing (input word attribute check processing), L processing (output word value generation processing) ), 12 processing (Logical body input processing), 04 processing (Logical body output processing))) and a function for allocating a given code string based on the Lyee methodology corresponding to the declaration execution unit A dedicated chip, a dedicated circuit, or software that allows a computer to perform the function (including software as a tool)

Or a recording medium on which the software is recorded, a processing apparatus on which the recording medium is mounted, a management apparatus / tool.

 [0143] In addition to program information, the information storage unit 1606 is a model of various control information, logical elements (L2, L3, L4) modularized as process' cells, and action elements (12, 04, S4), Code string layout part 1605 template, topological 'sort part 1604 template, (semi) order relation definition part 1603 template, execution unit specification part for word unit declaration 1602 model, target program, etc. It is also used as a memory for temporary storage.

 [0144] FIG. 17 shows a program (software), a program generation device, a program processing device, a tool (including both as a device and software), a software development device, a software development support device, or 5 is a flowchart showing the operation of the present invention implemented as one of software development management devices.

[0145] As shown in the figure, first, the execution unit specifying unit 1602 of the word unit declaration implements the user requirements implemented as one program for each logical unit with an access condition. For each word, from the word unit declaration declared by the word name, definition expression, execution condition of the definition expression, input / output attributes, and word value attributes, all L2 processing (input Specify the execution unit of one of the declarations (word attribute check processing), L processing (output word value generation processing), 12 processing (logical input processing), 04 processing (logical output processing) Step 1701). In defining the execution unit of this declaration, a template stored in the information storage unit 1606 may be used.

 [0146] Then, the (semi) order relation defining unit 1603 defines the (semi) order relations of all the declared execution units (step 1702). A template stored in the information storage unit 1606 may be used to define the (semi) order relation of the declaration execution unit.

 Next, the topological 'sorting unit 1604 performs topological' sorting on all the declaration execution units specified by the defined (semi) order relationship (step 1703). In performing this topological sort, a template stored in the information storage unit 1606 may be used.

 [0148] Next, the code string arrangement unit 1605 obtains a given code string based on the Lyee methodology corresponding to the declaration execution unit from the information storage unit 1606 according to the sequence of execution units of the sorted declarations. It arranges after reading (step 1704).

 Therefore, according to the present invention having the above-described configuration, the process of the declaration execution module for generating the output data in the program that also serves as the word-by-word declaration execution module is eliminated. It can be completed with the minimum number of executions and the program size can be reduced.

[0150] Program (software), program generation device, program processing device, tool (including both as device or software) for producing "software to be developed" as different embodiments of the present invention The software development device, the software development support device, and the software development management device, the user requirements to be implemented as a single program, for each logical body with access conditions, for each word on the logical body, a word name, a definition formula, From the execution condition of the definition expression, input / output attributes, and word unit declarations declared by the word value attribute, all L2 processing (input word attribute check processing) and L processing ( Output word value generation process), 12 process (logical input process), 04 process (logical output process) All of the specified L2 processing (input word attribute check processing), L processing (output word value generation processing), 12 processing (logical input processing), 04 processing (logical output processing) (half) Means for defining an order relationship, means for performing a topological sort on the L2 processing, L processing, 12 processing, and 04 processing defined by the (semi-) order relationship defined by the step, and the third Normal by step It is also possible to provide a means for arranging a given code string based on the Lyee methodology corresponding to the declaration execution unit according to the sequence of execution units of the replaced declaration.

 [0151] The present invention further includes software produced by the above-described "method for producing software to be developed" and a recording medium on which the software is installed or an apparatus on which the software is installed. In this case, the present invention implements the user requirements implemented as one program for each logical body with access conditions, for each word on the logical body, a word name, a definition formula, Declaration power in units of words declared by the execution conditions of the definition formula, input / output attributes, and word value attributes All L2 processing (input word attribute check processing) required for satisfying the requirements, L processing ( Output word value generation processing), 12 processing (logical body input processing), 04 processing (logical body output processing) declaration execution unit based on (semi-) order relationship that also defined the requirement power declared in word units Line It can also be configured as a given code sequence based on the appropriate Lyee methodology, according to the ordered sequence reordered by the topological sort.

 [0152] Further, the present invention relates to software as a template of software code used for producing software by the above-mentioned "method for producing software to be developed", and a recording medium on which the software is mounted. Alternatively, it can be realized as a device (hardware) on which the software is installed.

 [0153] Furthermore, the present invention is an extraction method of information (document (paper, data)) extracted from the requirements by the above-mentioned "method for producing software to be developed" and extraction by a powerful extraction method. Information (document (paper, data)), as a method of using the extracted information, or as an information recording medium on which such information is mounted, or as a method of extracting information Z It may be realized as information extracted from the software development requirements, including coded software and related information that can be realized as a recording medium Z device (hardware) on which the software is installed.

[0154] Since the present invention can be easily modified and modified by those having ordinary skill in the art, it is desirable to limit the present invention to configurations and operations that are not exactly different from those illustrated and described. And therefore any suitable modification The body and equivalents can be considered within the scope of the present invention. Although described in detail with the description and illustration of the specific embodiments of the present invention described above, it deviates from the scope of the present invention defined not only in the claims of the present application but also in the entire disclosure of the present invention. Many modifications, substitutions, and changes are possible without doing so.

 [0155] Further, the invention according to the present application is not limited to the detailed interpretation and combination of elements described in the above description or shown in the drawings. The invention is capable of other embodiments and of being practiced and practiced in various ways. Also, the terminology and terminology used here is for descriptive purposes and should not be considered as limiting.

 [0156] Accordingly, those of ordinary skill in the art may design other structures, methods, and systems to implement some of the objectives of the present disclosure based on the underlying concepts of this disclosure. It should be understood that it can be easily utilized as a basis for doing so. Accordingly, it is intended that the scope of the claims of the present application include such equivalent interpretations without departing from the spirit and scope of the invention.

 [0157] In the above detailed description, the method for realizing the idea of the present invention in the following order: directed graph notation, adjacency matrix calculation, topological sort, rearrangement, and the like has been described. However, these are not absolutely indispensable elements. For example, the idea of the present invention can be realized in the order of direct adjacency matrix calculation → topological 'sorting → rearrangement without going through the notation of directed graph notation.

 [0158] Further, for directed graphs, for example, breadth-first search, iterative deeping, heuristic search, Hill Climbing, and optimal first search (Best — Sorting may be done by combining various search techniques such as First Search, Euler, and Dijkstra.

 [0159] Alternatively, the basic idea of the present invention may be realized by combining these various search techniques after calculating the adjacency matrix.

[0160] The technical idea of the present invention includes, for example, an automatic development device for computer software, an automatic development program, a recording medium recording the automatic development program, a transmission medium, and a paper medium. However, it can also be realized and used in categories such as computer 'equipment with automatic development program', client that runs automatic development program 'server type, etc.! /.

[0161] Further, the present invention is not limited to a computer system including a single processor, a single hard disk drive, and a single local memory, and as an option of the system, any plural or combination processors or storage devices are provided. Also suitable for. Computer systems include sophisticated calculators, palm-type computers, laptop z notebook computers, minicomputers, mainframe computers, and supercomputers, and combinations of these processing system networks. It can be replaced by any suitable processing system operating in accordance with the principles of the present invention and can be used in combination.

 [0162] Furthermore, the technical idea according to the present invention can be applied to all kinds of programming languages. Furthermore, the technical idea according to the present invention can be applied to application software of all kinds of functions.

 [0163] Furthermore, the present invention allows various modifications, additions, substitutions, enlargements, reductions, etc. within the scope of the same and equivalent technical idea. In addition, even if the software power produced using the invention of the present application is listed as a secondary product and commercialized, the value of the present invention will not be reduced.

 Industrial applicability

[0164] In the present invention as defined above, the pre-processing to eliminate repetition is automatically performed, and the pre-processed word unit program thus obtained is automatically used by the Lyee (registered trademark) methodology. Generate the target program. In other words, it is possible to automate everything from the refinement (alignment) of user requirements to the generation of the target program. This will have a significant effect on the software industry, such as greatly improving the efficiency of software production, improving productivity, and improving quality.

 Brief Description of Drawings

FIG. 1 is a conceptual diagram for explaining the concept of a cell according to an embodiment of the present invention.

FIG. 2 is an explanatory diagram of a screen of the system of Example 1 according to an embodiment of the present invention. FIG. 3 is an explanatory diagram showing words in the system of Example 1 according to an embodiment of the present invention in units of basic composition.

 FIG. 4 is a directed graph showing word relationships in the system of Example 1 according to an embodiment of the present invention.

[5] FIG. 5 is a diagram for explaining a process in which the entire system of Example 1 according to the embodiment of the present invention is represented by one adjacency matrix F.

[6] FIG. 6 is a diagram for explaining that the adjacency matrix F according to an embodiment of the present invention has been topologically sorted.

[7] Structural force of the system of Example 1 according to an embodiment of the present invention is a directed graph representing the relationship between words when a path action element is deleted.

8] Structural force of system of example 1 according to an embodiment of the present invention is a diagram for explaining a process of representing the entire system by one adjacency matrix F ′ when a path action element is deleted. [9] FIG. 9 is a diagram for explaining that the adjacency matrix F ′ according to an embodiment of the present invention has been topologically sorted.

[10] Example 2 according to an embodiment of the present invention is a directed graph.

[11] FIG. 11 is a diagram showing an “initial screen” of Example 2 according to an embodiment of the present invention.

12] A diagram showing the “student registration screen” in Example 2 according to an embodiment of the present invention.

13] A diagram showing a “score management screen” in Example 2 according to an embodiment of the present invention.

 FIG. 14 is a processing path diagram showing the requirements of the grade management program of Example 2 according to an embodiment of the present invention and the structure when this program is divided into conventional basic structures.

[15] Part of the integrated new program code according to an embodiment of the present invention.

 [FIG. 16] As a different embodiment of the present invention, a program (software), a program generation device, a program processing device, a tool (including both as a device and software) for producing “software to be developed” FIG. 3 is a functional block diagram showing a configuration provided as a function when the present invention is implemented as any one of a software development device, a software development support device, and a software development management device.

圆 17] A program (software) having the above-described configuration according to an embodiment of the present invention, a program The operation of the present invention implemented as any of a program generation device, a program processing device, a tool (including both as a device and software), a software development device, a software development support device, and a software development management device is shown. It is a flowchart.

Explanation of symbols

101 Sync range

1101 Basic structure BS1 (input)

1102 Basic structure BS2

1103 Basic structure BS3

1104 Basic structure BS1 (Output)

1201 Adjacency matrix Fl,

1202 Adjacency matrix F2,

1203 Adjacency matrix F3 '

1204 Joined adjacency matrix FC1 '

1205 Joined adjacency matrix FC2 '

Claims

The scope of the claims

 User requirements to be implemented as a single program for each logical body with access conditions, for each word on the logical body, the word name, definition expression, execution condition of the definition expression, input / output attributes, and word value From the word unit declaration declared by the attribute, all the necessary L2 processing (input word attribute check processing), L processing (output word value generation processing), 12 processing (logical input processing), 04 process (logic output process), the first step to define the execution unit of any declaration,

 All the L2 processing (input word attribute check processing), L processing (output word value generation processing), 12 processing (logical input processing), 04 processing (logical output processing) specified in the above (half ) A second step for defining the order relationship, and a third step for performing topological 'sorting on the L2 processing, L processing, 12 processing, 04 processing defined by the (semi) order relationship defined by the step Steps,

 A 64th step of allocating a given code sequence based on the Lyee methodology corresponding to the declaration execution unit according to the sequence of execution units of the declaration rearranged by the third step;

 A software generation method comprising: