JP5687122B2 - Software evaluation device, software evaluation method, and system evaluation device - Google Patents

Description

  The present invention relates to a software evaluation apparatus and software evaluation method that can be used for software structure analysis and influence range analysis, and a system evaluation apparatus that can be used for system structure analysis and influence range analysis.

  As one of software (S / W) architecture evaluation techniques, there is a method using DSM (Design Structure Matrix). In this DSM, dependency that occurs between the service provider and the user is obtained from the source code by “calling a function”, “referencing a variable”, and the like. Such DSM is suitable for concisely expressing dependencies between tasks of complex processes and analyzing task relevance.

  For example, by applying DSM to S / W, the dependency between program components is weighted by the frequency of occurrence (dependency) and displayed in a tabular format, thereby simplifying the complex S / W architecture structure Can be expressed in

  In the information flow model creation method described in Patent Document 1, the DSM dependency is weighted, and the weighted DSM is mapped to a pre-patterned information flow model. An information flow model used for simulation of various processes is easily created.

  Further, in the project management method described in Patent Document 2, a dependency (influence on a project) indicating dependency between a plurality of tasks included in a project is created with a DSM expressed in a matrix format, and the DSM is weighted. The risk of backtracking and parallel processes in project management is quantified.

JP 2003-141314 A JP 2003-030395 A

  However, there are multiple types of dependency relationships such as “function call” and “variable reference” in the S / W dependency relationship, meaning that each has a different meaning, and the strength of dependency depends on the dependency type even with the same dependency frequency. I know empirically that it is different. For example, it has been found from experience that “depending on two times of function calls” and “reference of variables two times” are more dependent on “two times of function calls” even if they are the same two times of dependence.

  For this reason, there is a problem that the strength of dependence cannot be determined from the dependence strength calculated only from the frequency of occurrence of the dependence, as in the former and latter prior arts. For example, it is difficult to accurately grasp the degree of coupling between a certain module (provider) and another module (user) and to mechanically specify the range of the influence of source code modification. As described above, when there are a plurality of types of dependency relationships and the strengths of the dependencies differ from one another, it is difficult to determine the strength of the dependencies.

  The present invention has been made in view of the above, and obtains a software evaluation apparatus and software evaluation method capable of accurately and easily evaluating software, and a system evaluation apparatus capable of accurately and easily evaluating a system. For the purpose.

In order to solve the above-described problems and achieve the object, the present invention provides a weight that is the importance of the dependency type for each dependency type that is the type of dependency between software source code and the software components. An input unit that accepts input of weight information, a storage unit that stores source code and weight information input to the input unit, a combination of the components from the source code in the storage unit, and the configuration A dependency extraction unit that extracts a correspondence relationship between the dependency frequency, which is the number of dependency relationships between elements, and the dependency type, as dependency relationship information for each combination of the component elements, the dependency relationship information, and the weight information Is used to weight the dependence frequency according to the weight, thereby depending on the combination of the combination of the component elements and the dependency type between the component elements. Calculating a dependent intensity calculation unit for calculating a strength, and dependent intensity between the components, the correspondence between the frequency of occurrence within the combination of the dependent intensity, based on the calculated relationship, the appearance of more than a predetermined value extracting a threshold setting unit that sets the value of dependent intensity corresponding to the frequency to the threshold of the dependent intensity, what the dependent intensity from the combinations of the components is equal to or greater than the threshold value, the combination of the extracted components The range of components that can be traced from the correction target of the source code that directly or indirectly depends on the correction target of the source code among the components included in the source code is defined as the influence range when the source code is corrected comprising a influence range extraction unit for extracting, wherein the input unit inputs the range of components used for calculating the threshold as calculated range information, wherein the storage unit Storing the calculated range information, the threshold setting unit, the correspondence between calculated for a range of the specified components in the calculating range information, set the threshold value based on the calculated correspondence Rukoto It is characterized by.

  According to the present invention, it is possible to evaluate software accurately and easily.

FIG. 1 is a block diagram showing the configuration of the influence analysis device according to the first embodiment. FIG. 2 is a diagram illustrating a configuration example of the dependency relationship list. FIG. 3 is a diagram illustrating a configuration example of the dependency type table. FIG. 4 is a diagram illustrating a configuration example of a weighted dependency relationship list. FIG. 5 is a diagram illustrating a configuration example of the weighted DSM according to the first embodiment. FIG. 6 is a diagram showing the weighted DSM shown in FIG. 5 in units of files. FIG. 7 is a flowchart showing a calculation processing procedure of the weighted dependence strength. FIG. 8 is a diagram illustrating a configuration example of the DSM created without considering the dependency type. FIG. 9 is a diagram showing the DSM created without considering the dependency type in units of files. FIG. 10 is a block diagram illustrating a configuration of the influence analysis device according to the second embodiment. FIG. 11 is a diagram illustrating the appearance frequency with respect to the dependency strength. FIG. 12 is a block diagram showing the configuration of the influence analysis device according to the third embodiment. FIG. 13 is a block diagram showing the configuration of the influence analysis device according to the fourth embodiment. FIG. 14 is a block diagram showing the configuration of the influence analysis device according to the fifth embodiment. FIG. 15 is a diagram for explaining an example of a weight calculation method. FIG. 16 is a block diagram showing the configuration of the influence analysis device according to the sixth embodiment. FIG. 17 is a diagram illustrating a configuration example of the weighted DSM according to the sixth embodiment. FIG. 18 is a diagram illustrating a hardware configuration of the influence analysis apparatus.

  Hereinafter, a software evaluation device, a software evaluation method, and a system evaluation device according to embodiments of the present invention will be described in detail with reference to the drawings. Note that the present invention is not limited to these embodiments.

Embodiment 1 FIG.
FIG. 1 is a block diagram showing the configuration of the influence analysis device according to the first embodiment. The impact analysis apparatus 101 is a computer that analyzes the impact of S / W components according to the weight (importance) set for each type of software (S / W) components (dependency). Device. The influence degree analysis apparatus 101 visualizes the S / W structure so that it can be used for S / W structure analysis and influence range analysis. The impact analysis apparatus 101 uses a DSM (Design Structure Matrix or Dependency Structure Matrix) (design structure matrix) when evaluating the S / W structure.

  The influence analysis device 101 includes an input unit 201, a storage unit M, a dependency relationship extraction unit 2, a weighted dependency strength calculation unit 5, a weighted DSM generation unit 7, an influence analysis unit (an influence range extraction unit). ) 9, an output unit 202, and a display unit 203.

  The input unit 201 inputs the source code 1 and the S / W component 11 from an external device and sends them to the storage unit M. Further, the input unit 201 sends the dependency type table 4 and the threshold value T input from the user to the storage unit M. The source code 1 is an enumeration of characters (a series of instructions to the computer) that is the source of S / W (computer program).

  The threshold value T is a value that serves as a criterion for determining whether or not the S / W component has an influence level (weighted dependence strength described later) of a predetermined value or more, and is set by the user for each S / W component. . The weighted DSM 8 is information indicating the weighted dependency strength calculated for each combination (between S / W components) of the dependency source S / W component and the dependency destination S / W component. The S / W component 11 is an S / W component (a user-desired S / W component) that is a dependency determination target among the S / W components, and is specified by the user.

  The dependency relationship extraction unit 2 extracts a dependency relationship between S / W components from the source code 1 and creates a dependency relationship list 3. The dependency relationship extraction unit 2 stores the created dependency relationship list 3 in the storage unit M. FIG. 2 is a diagram illustrating a configuration example of the dependency relationship list. The dependency relationship list 3 is information (dependency relationship information) in which a dependency source S / W component, a dependency destination S / W component, a dependency type, and a dependency frequency are associated with each other.

  The dependency source S / W component is a function (S / W component) that is a dependency source among functions having dependency relationships, and the dependency destination S / W component is a dependency among functions having dependency relationships. This is the previous function (S / W component).

  The dependency type is a dependency type, for example, “include”, “macro”, “function call”, and the like. The dependency frequency is the number of dependency relationships. For example, line 13 of dependency list 3 indicates that there is a dependency frequency of “function call” from the func function of file1.cpp to the exp1 function of file2.cpp four times.

  The weighted dependency strength calculating unit 5 converts the dependency relationship list 3 into a weighted dependency relationship list 6 using the dependency relationship list 3 created by the dependency relationship extracting unit 2 and the dependency type table 4. The weighted dependency strength calculating unit 5 stores the converted weighted dependency relationship list 6 in the storage unit M.

  FIG. 3 is a diagram illustrating a configuration example of the dependency type table. The dependency type table 4 is information (weight information) in which the dependency type is associated with the weight, and is set by the user. The weight is a weight set for the dependency type. In this embodiment, a weight is set for each dependency type. For example, the row 15 of the dependency type table 4 indicates that the weight of the dependency type “function call” is 0.9.

  FIG. 4 is a diagram illustrating a configuration example of a weighted dependency relationship list. FIG. 4 shows a configuration example of a weighted dependency relationship list 6 created using the dependency relationship list 3 in FIG. 2 and the dependency type table 4 in FIG. The weighted dependency relationship list 6 is information in which a dependency source S / W component, a dependency destination S / W component, a dependency type, and a weighted dependency strength are associated with each other.

  The weighted dependence strength is a dependence frequency considering the weight of the dependence type. The weighted dependence strength is calculated, for example, by multiplying the dependence frequency in the dependence relation list 3 and the weight in the dependence type table 4.

  In line 13 of the dependency list 3 shown in FIG. 2, “function call” is set from the func function of file1.cpp to the exp1 function of file2.cpp. The dependency frequency on the line 13 in the dependency relationship list 3 is four times, and the weight of the “function call” on the line 15 in the dependency type table 4 shown in FIG. 3 is 0.9. Therefore, the weighted dependency strength calculation unit 5 calculates weighted dependency strength = 3.6 by the dependency frequency (4 times) × weight (0.9) = weighted dependency strength. As a result, in the row 16 of the weighted dependency relationship list 6, the weighted dependency strength of the dependency “function call” from the func function of file1.cpp to the exp1 function of file2.cpp becomes 3.6. Thus, in the present embodiment, the weighted dependence strength is calculated for each dependence type.

  The weighted DSM generation unit 7 generates a weighted DSM 8 based on the weighted dependency relationship list 6 (processing result) created by the weighted dependency strength calculation unit 5. In addition, the weighted DSM generation unit 7 generates a file unit weighted DSM 40 to be described later using the weighted DSM 8. The weighted DSM generation unit 7 stores the generated weighted DSM 8 and file unit weighted DSM 40 in the storage unit M (the file unit weighted DSM 40 is not shown in FIG. 1).

  FIG. 5 is a diagram illustrating a configuration example of the weighted DSM according to the first embodiment. FIG. 5 shows a configuration example of the weighted DSM 8 created using the weighted dependency relationship list 6 of FIG. The weighted DSM 8 is information indicating the weighted dependence strength calculated for each S / W component. In the weighted DSM 8 of FIG. 5, the dependency source S / W components are arranged in the column direction, and the dependency destination S / W components are arranged in the row direction. Then, at the position of the intersection of the dependency source S / W component and the dependency destination S / W component, which is a combination of the dependency source S / W component and the dependency destination S / W component , The value of weighted dependence strength is entered. For example, in the cell 17 within the range 24A, the weighted dependence strength from the func function (dependence source) of file1.cpp to the exp1 function (dependence destination) of file2.cpp is 3.6.

  The weighted DSM generating unit 7 generates the weighted DSM8 by adding each dependency from the weighted dependency relationship list 6 to the corresponding cell in the weighted DSM8. For example, the cell 17 has the value of the row 16 shown in FIG. 4 which is the func function whose dependency source is file1.cpp and the exp1 function whose dependency destination is file2.cpp.

  FIG. 6 is a diagram showing the weighted DSM shown in FIG. 5 in units of files. The file unit weighted DSM 40 is information indicating the weighted DSM 8 shown in FIG. 5 in file units. In the file unit weighted DSM 40, the weighted dependency strength calculated for each combination of the dependency source S / W component and the dependency destination S / W component is expressed in file units. Therefore, the file unit weighted DSM 40 is information indicating the weighted dependence strength calculated for each combination of the dependence source file and the dependence destination file.

  The weighted DSM 8 shown in FIG. 5 includes file1.cpp, file2.cpp, file2.h, file3.h, and file4.h as the dependency source files, and file1.cpp and file2 as the dependency destination files. There are .cpp, file2.h, file3.h, and file4.h files. Therefore, in the file unit weighted DSM 40, file1.cpp, file2.cpp, file2.h, file3.h, file4.h, file1.cpp, file2.cpp, file2.h, file3.h, file4.h , And the weighted dependence intensity calculated for each combination.

  For example, in the weighted DSM 8 shown in FIG. 5, there are file1.cpp, file1.cpp / main (), and file1.cpp / func () as the file1.cpp of the dependence source. In addition, there are file2.h and file2.h / exp1 (int) as files of file2.h that depend on. Of these combinations, 0.3 is set as the weighted dependency strength for the combination of the dependency source file1.cpp and the dependency destination file2.h, and the dependency source file1.cpp / func () and 0.4 is set as the weighted dependence strength for the combination of the dependency destination file2.h / exp1 (int). For this reason, in the file unit weighted DSM 40, 0.3 + 0.4 = 0.7 is set as the weighted dependence strength for the combination of the dependence source file 1.cpp and the dependence destination file 2.h. Has been.

  The influence analysis unit 9 is an S / W component that depends on the S / W component 11 among the S / W components in the file unit weighted DSM 40 (weighted DSM 8), and the weighted dependency strength is a threshold value. S / W components exceeding T are extracted, and the influence range list 12 is created using the extracted S / W components. The influence analysis unit 9 stores the created influence range list 12 in the storage unit M. The influence range list 12 is a list of S / W components that are dependent on the S / W component 11 and whose weighted dependency strength exceeds the threshold T. In the influence range list 12, an S / W component having a weighted dependency strength exceeding the threshold T is registered as an influence range when the source code is modified.

  The storage unit M is a memory that stores the source code 1, the dependency type table 4, the threshold T, the S / W component 11, the dependency relationship list 3, the weighted dependency relationship list 6, the weighted DSM 8, and the influence range list 12. is there.

  The output unit 202 outputs the dependency relationship list 3, the weighted dependency relationship list 6, the weighted DSM 8, the file unit weighted DSM 40, and the influence range list 12 to an external device or the like. Note that the output unit 202 may output the source code 1, the dependency type table 4, the threshold T, and the S / W component 11.

  The display unit 203 includes a dependency relationship list 3, a weighted dependency relationship list 6, a weighted DSM 8, a file unit weighted DSM 40, an influence range list 12, a source code 1, a dependency type table 4, a threshold T, and an S / W component 11. A device such as a liquid crystal monitor that displays the above.

  Next, an analysis process procedure of the S / W influence level (a process procedure for creating the influence range list 12) will be described. Source code 1, dependency type table 4, and threshold value T are input to input unit 201 in advance. The input unit 201 stores the source code 1, the dependency type table 4, and the threshold T in the storage unit M.

  The dependency relationship extraction unit 2 extracts a dependency relationship from the source code 1 and creates a dependency relationship list 3. Then, the dependency relationship extraction unit 2 stores the created dependency relationship list 3 in the storage unit M. Further, the weighted dependency strength calculating unit 5 converts the dependency relationship list 3 into the weighted dependency relationship list 6 by using the dependency relationship list 3 created by the dependency relationship extracting unit 2 and the dependency type table 4. . The weighted dependency strength calculating unit 5 stores the converted weighted dependency relationship list 6 in the storage unit M.

  Here, the calculation processing procedure of the weighted dependency strength by the weighted dependency strength calculation unit 5 will be described. FIG. 7 is a flowchart showing a calculation processing procedure of the weighted dependence strength. The weighted dependence strength calculation unit 5 weights all S / W components in the dependency relationship list 3 (a set of dependence source S / W components and dependence destination S / W components). It is determined whether or not conversion to the relationship list 6 has been performed (step S101).

  When all the S / W components in the dependency relationship list 3 are converted to the weighted dependency relationship list 6 (Yes in step S101), the calculation processing of the weighted dependency strength of all dependency relationships is performed. Finish because it is complete.

  On the other hand, when the conversion to the weighted dependency relationship list 6 is not performed for all the S / W components in the dependency relationship list 3 (No in step S101), the calculation processing of the weighted dependency strength is finished. Since there is no dependency, weighted dependency strength calculation processing is performed. That is, the processes of steps S102 to S105 are executed.

  Specifically, the weighted dependency strength calculation unit 5 determines that one dependency relationship (dependence source S / W component element and dependency destination S / W component element) from the dependency relationship in which the weighted dependency strength calculation process has not been performed. ) Is selected (step S102). Then, the weighted dependency strength calculation unit 5 searches the dependency type table 4 for a dependency type that matches the dependency type of the selected dependency relationship, and obtains a weight corresponding to the dependency type (step S103). Further, the weighted dependency strength calculation unit 5 calculates the weighted dependency strength by multiplying the dependency frequency of the selected dependency relationship by the acquired weight (step S104).

  Then, the weighted dependence strength calculation unit 5 calculates the weighted dependence strength and the dependence relationship selected in the process of step S102 (a combination of the dependence source S / W component and the dependence destination S / W component, Dependency type) is added to the weighted dependency relationship list 6 (step S105). Thereafter, the processing of steps S101 to S105 is repeated until all the S / W components in the dependency relationship list 3 are converted into the weighted dependency relationship list 6.

  After the weighted dependency relationship list 6 is created, the weighted DSM generation unit 7 generates a weighted DSM 8 based on the weighted dependency relationship list 6. Further, the weighted DSM generation unit 7 generates a file unit weighted DSM 40 using the weighted DSM 8. The weighted DSM generation unit 7 stores the generated weighted DSM 8 and file unit weighted DSM 40 in the storage unit M.

  The display unit 203 displays the weighted DSM 8, the file unit weighted DSM 40, the influence range list 12, and the like. Thereafter, the S / W component 11 is input from the input unit 201 by the user and stored in the storage unit M.

  The influence analysis unit 9 is an S / W component that depends on the S / W component 11 among the S / W components in the file unit weighted DSM 40 and has a weighted dependency strength exceeding the threshold T. The W component is extracted, and the influence range list 12 is created using the extracted S / W component.

  In the S / W component 11, for example, the S / W component “file4.h” of the file unit weighted DSM 40 of FIG. 6 is designated by the user. In this case, the influence analysis unit 9 is an S / W component that directly or indirectly depends on the S / W component “file4.h”, and has an S / W configuration that exceeds the threshold T = 3.0. Extract elements. First, the influence analysis unit 9 extracts S / W components that depend on “file4.h” and that exceed the threshold T = 3.0. Among the S / W components registered in the file unit weighted DSM 40, those that depend on the S / W component “file4.h” are “file1.cpp” (dependency strength: 2. 7) and “file2.cpp” (dependence intensity: 3.9). For this reason. For example, when the influence range of the S / W component “file4.h” is the threshold T = 3.0, the influence analysis unit 9 extracts “file2.cpp” exceeding the threshold T = 3.0.

  Further, the influence analysis unit 9 extracts S / W components that depend on the extracted “file2.cpp” and that exceed the threshold T = 3.0. What depends on the S / W component “file2.cpp” registered in the file unit weighted DSM 40 is “file1.cpp” (dependence strength: 3.6) from the line 19. Therefore, the influence analysis unit 9 extracts “file1.cpp” that depends on “file2.cpp”. Since no S / W component that depends on the extracted “file1.cpp” exceeds the threshold T = 3.0, the influence range list 12 includes “file1.cpp” and “file2.cpp”. Will be registered.

  As described above, the influence analysis unit 9 extracts the S / W component that depends on the S / W component specified by the user, and extracts the S / W component that depends on the extracted S / W component. Repeat the process. In other words, the influence analysis unit 9 sequentially traces the S / W components that depend on the S / W components. Then, the influence degree analysis unit 9 creates the influence range list 12 by registering the extracted S / W components in the influence range list 12.

  The influence analysis unit 9 stores the created influence range list 12 in the storage unit M. The display unit 203 displays the weighted dependency relationship list 6, the weighted DSM 8, the file unit weighted DSM 40, the influence range list 12, and the like. Further, the output unit 202 outputs the weighted dependency relationship list 6, the weighted DSM 8, the influence range list 12 and the like to an external device or the like as necessary.

  Here, the difference between the weighted DSM 8 of the present embodiment and the DSM generated by the conventional method will be described. FIG. 8 is a diagram illustrating a configuration example of the DSM created without considering the dependency type. The DSM 51 is a DSM created by a conventional method, and is a DSM that expresses S / W components such as functions and variables with a minimum granularity. In the DSM 51, dependency source S / W components are arranged in the column direction, and dependency destination S / W components are arranged in the row direction. Then, at the position of the intersection of the dependency source S / W component and the dependency destination S / W component, which is a combination of the dependency source S / W component and the dependency destination S / W component The value of dependency strength (dependence frequency) is entered. The dependency frequency in the DSM 51 is a dependency frequency calculated without considering the dependency type and weighting. For example, cell 14 indicates that the dependency strength from the func function of file1.cpp to the exp1 function of file2.cpp is 4.

  FIG. 9 is a diagram showing the DSM created without considering the dependency type in units of files. The file unit DSM 52 is a file unit of the DSM 51 shown in FIG. As described above, since the conventional DSM 51 and the file unit DSM 52 are created without considering the dependency type and weighting, it is difficult to determine the strength of the dependency.

  As described above, according to the first embodiment, the DSM (weighted DSM8, file unit weighted DSM40) is created in consideration of the weight for each dependency type. For this reason, it is possible to grasp the strength of dependence by numerical values, automatically detect the range of influence of modification, and to improve the architecture analysis technology and efficiency of analysis work by eliminating personality. Therefore, there is an effect that S / W can be accurately and easily evaluated.

Embodiment 2. FIG.
Next, a second embodiment of the present invention will be described with reference to FIGS. In the first embodiment described above, the threshold value T input to the influence analysis unit 9 is given by the user. In the present embodiment, the threshold value T is automatically adjusted (calculated) based on the weighted DSM 8.

  FIG. 10 is a block diagram illustrating a configuration of the influence analysis device according to the second embodiment. Of the constituent elements in FIG. 10, constituent elements that achieve the same functions as those of the influence analysis apparatus 101 of the first embodiment shown in FIG. 1 are given the same numbers, and redundant descriptions are omitted.

  The influence analysis device 102 according to the second embodiment includes a threshold automatic adjustment unit (threshold setting unit) 20 in addition to the functions of the influence analysis device 101. Specifically, the influence analysis device 102 includes an input unit 201, a storage unit M, a dependency relationship extraction unit 2, a weighted dependency strength calculation unit 5, a weighted DSM generation unit 7, and an influence analysis unit. 9, an output unit 202, a display unit 203, and a threshold automatic adjustment unit 20.

  The threshold automatic adjustment unit 20 is connected to the storage unit M, and sets the threshold T in the storage unit M using the weighted DSM 8 in the storage unit M. The threshold automatic adjustment unit 20 extracts dependency information (appearance frequency for each dependency strength) from the weighted DSM 8 and statistically calculates the threshold T based on the dependency information. The influence analysis unit 9 of the present embodiment creates the influence range list 12 using the threshold T calculated by the automatic threshold adjustment unit 20.

  Next, an example of a method for automatically adjusting the threshold value T will be described using the weighted DSM 8 shown in FIG. 5 of the first embodiment. FIG. 11 is a diagram illustrating the appearance frequency with respect to the dependency strength. FIG. 11 shows the relationship between the dependency strength and the appearance frequency, where the horizontal axis is the dependency strength and the vertical axis is the appearance frequency (series 2).

  Series 1 is the number of S / W components, and sequence 2 is the ratio of S / W components. Specifically, the series 1 is the number of S / W components included in the weighted DSM 8 included in the range of dependence strength shown on the horizontal axis, and the series 2 is included in the range of dependence strength shown on the horizontal axis. This is the ratio (number ratio) of the S / W component to be occupied in the entire weighted DSM 8. For example, the weighted DSM 8 in FIG. 5 has 10 S / W components whose dependency strength is within 0.5. Also, the weighted DSM 8 has a total of 15 S / W components for which the dependency strengths are registered. Therefore, among the S / W components for which the dependency strength is registered, the S / W component having the dependency strength within 0.5 is 10/15 = 0.67, which is about 67% of the total. As a result, 67% is shown as the series 2. The dependence strength of 0 to 1.0 has an appearance frequency of about 0.87 (87%).

  For example, in the case where the dependency strength with an appearance frequency exceeding 80% is set as the threshold T, the threshold automatic adjustment unit 20 sets the threshold T to 1.0 based on the relationship of FIG. Thus, since the threshold value T is automatically adjusted using the relationship of the appearance frequency to the dependency strength, the influence analysis unit 9 can extract S / W components having an appearance frequency equal to or higher than a predetermined value. It becomes. Note that the threshold automatic adjustment unit 20 may automatically adjust the threshold T by a method other than the method using the relationship of the appearance frequency to the dependency strength.

  As described above, according to the second embodiment, the threshold T is automatically adjusted using the relationship of the appearance frequency to the dependency strength. Therefore, the threshold T can be automatically set without belongings. Therefore, it is possible to improve the S / W architecture analysis technology and increase the efficiency of analysis work.

Embodiment 3 FIG.
Next, Embodiment 3 of the present invention will be described with reference to FIG. In the second embodiment described above, the threshold automatic adjustment unit 20 automatically sets the threshold T from the statistical data of the dependency relationship of the entire S / W. In the present embodiment, a range of S / W components used when the threshold value T is automatically adjusted is designated.

  FIG. 12 is a block diagram showing the configuration of the influence analysis device according to the third embodiment. Of the constituent elements in FIG. 12, constituent elements that achieve the same functions as those of the influence analysis apparatus 101 of the first embodiment shown in FIG. 1 are given the same numbers, and redundant descriptions are omitted.

  The influence analysis device 103 according to the third embodiment includes a threshold automatic adjustment unit 22 with a calculation range in addition to the functions of the influence analysis device 101. Specifically, the influence analysis device 103 includes an input unit 201, a storage unit M, a dependency relationship extraction unit 2, a weighted dependency strength calculation unit 5, a weighted DSM generation unit 7, and an influence analysis unit. 9, an output unit 202, a display unit 203, and a threshold automatic adjustment unit 22 with a calculation range.

  The calculation range 23 is input in advance to the input unit 201 and stored in the storage unit M. The calculation range 23 is a list (calculation range information) of S / W components used for automatic calculation of the threshold value T. Therefore, in the present embodiment, the threshold T is calculated using the dependency information (appearance frequency for each dependency strength) of the S / W component specified in the calculation range 23.

The calculation range-added threshold automatic adjustment unit 22 extracts the dependency information in the range of the S / W component specified in the calculation range 23 from the weighted DSM 8 and calculates the threshold T using the extracted dependency information. For example, when calculating the threshold value T from the weighted DSM 8 shown in FIG. 5, if the following S / W component is specified as the calculation range 23, the dependence strength of the range 24A is used for calculating the threshold value T. .
Calculation range 23:
File1.cpp, file1.cpp / main (), file1.cpp / func ()
file2.cpp, file2.cpp / exp1 (int)
file2.h, file2.h / exp1 (int)

  The calculation range-added threshold automatic adjustment unit 22 stores the calculated threshold T in the storage unit M. The influence degree analysis unit 9 creates the influence range list 12 using the threshold value T calculated by the threshold value automatic adjustment unit 22 with calculation range.

  Thus, according to the third embodiment, the S / W component used for calculating the threshold value T can be set in detail by the calculation range 23. And since the threshold value T is calculated using the dependence information in the range of the S / W component designated by the calculation range 23, the probability of the threshold value T can be improved.

Embodiment 4 FIG.
Next, a fourth embodiment of the present invention will be described with reference to FIG. In the second embodiment described above, the threshold automatic adjustment unit 20 automatically sets the threshold T by one calculation method. In the present embodiment, a threshold calculation method template that matches the S / W characteristics is prepared, and the threshold automatic adjustment unit 20 calculates a threshold T corresponding to the template.

  FIG. 13 is a block diagram showing the configuration of the influence analysis device according to the fourth embodiment. Of the constituent elements in FIG. 13, constituent elements that achieve the same functions as those of the influence analysis apparatus 101 of the first embodiment shown in FIG. 1 are given the same numbers, and redundant descriptions are omitted.

  The influence analysis device 104 according to the fourth embodiment includes a template use threshold automatic adjustment unit 25 in addition to the functions of the influence analysis device 101. Specifically, the influence analysis device 103 includes an input unit 201, a storage unit M, a dependency relationship extraction unit 2, a weighted dependency strength calculation unit 5, a weighted DSM generation unit 7, and an influence analysis unit. 9, an output unit 202, a display unit 203, and a template use threshold automatic adjustment unit 25.

  The calculation template 26 is input in advance to the input unit 201 and stored in the storage unit M. The calculation template 26 holds a plurality of threshold calculation methods (threshold T adjustment methods) used for calculating the threshold T. Therefore, the threshold value T calculation method can be switched by switching the threshold value calculation method set in the calculation template 26. Various threshold calculation methods are set according to the characteristics of the S / W language and the composition of the S / W.

  The template use threshold automatic adjustment unit 25 extracts dependency information from the weighted DSM 8 and calculates a threshold T using a threshold calculation method in a calculation template 26 designated by a user or the like.

  As described above, according to the fourth embodiment, the threshold value calculation method can be selected according to the S / W language characteristics and the S / W composition. Therefore, it is possible to set the threshold value calculation method according to the S / W. It becomes possible. Therefore, it is possible to improve the certainty of the threshold value T.

Embodiment 5 FIG.
Next, a fifth embodiment of the present invention will be described with reference to FIGS. In the first embodiment described above, the dependency type table 4 is given by the user. In the present embodiment, the dependency type table 4 is automatically adjusted based on past analysis data.

  FIG. 14 is a block diagram showing the configuration of the influence analysis device according to the fifth embodiment. Of the constituent elements shown in FIG. 14, the constituent elements that achieve the same functions as those of the influence analysis apparatus 101 of the first embodiment shown in FIG.

  The influence analysis device 105 according to the fifth embodiment includes an analysis data storage device (storage unit) 27 and an automatic weight adjustment unit (weight information generation unit) 28 in addition to the functions of the influence analysis device 101. . Specifically, the influence analysis device 105 includes an input unit 201, a storage unit M, a dependency relationship extraction unit 2, a weighted dependency strength calculation unit 5, a weighted DSM generation unit 7, and an influence analysis unit. 9, an output unit 202, a display unit 203, an analysis data storage device 27, and an automatic weight adjustment unit 28.

  The analysis data storage device 27 stores information (accumulation data) related to the dependency type table 4 every time the S / W analysis (creation of the influence range list 12) is performed. The analysis data storage device 27 may store information related to the dependency type table 4 with reference to the weighted DSM 8.

  The automatic weight adjustment unit 28 calculates the weight of each dependency type based on the accumulated data stored in the analysis data storage device 27 and generates the dependency type table 4. The automatic weight adjustment unit 28 calculates the weight from the accumulated data by a predetermined formula, and thereby generates the dependency type table 4. Then, the weight is corrected by the user as necessary.

  An example of a weight calculation method by the weight automatic adjustment unit 28 will be described. FIG. 15 is a diagram for explaining an example of a weight calculation method. Here, a case will be described in which the weight to be used next is calculated based on the average value of the weights used in the latest five S / W analyzes (creation of the weighted dependency relationship list 6). In FIG. 15, the set value of the weight of the dependency type table 4 for the latest five times and the weight calculated using this set value are shown for each dependency type.

  Weights 1 to 5 are the weights used for the latest five S / W analyses. The weight 6 is an average value of the weights 1 to 5 and is a weight set for the next S / W analysis. Therefore, the weight automatically adjusted by the weight automatic adjustment unit 28 becomes the weight of the column 29. As a result, the automatic weight adjustment unit 28 sets the weight 6 in the dependency type table 4. In the next S / W analysis, the dependency relationship list 3 is converted into the weighted dependency relationship list 6 using the dependency type table 4 in which the weight 6 is set.

  When calculating the weight to be used next based on the average value of the weights used in the most recent N times (N is a natural number) S / W analysis, when performing the first to (N-1) th S / W analysis The dependency type table 4 given by the user is used.

  As described above, according to the fifth embodiment, the dependency type weights can be automatically set from the past analysis data, so that the S / W architecture analysis technology and the efficiency of analysis work can be improved by eliminating personality. It becomes possible.

Embodiment 6 FIG.
Next, a sixth embodiment of the present invention will be described with reference to FIGS. In the first embodiment described above, all dependency information is displayed by the weighted DSM 8. In the present embodiment, the dependency information is filtered so as not to display the dependency information below the threshold T.

  FIG. 16 is a block diagram showing the configuration of the influence analysis device according to the sixth embodiment. Of the constituent elements in FIG. 16, constituent elements that achieve the same functions as those of the influence analysis apparatus 101 of the first embodiment shown in FIG. 1 are given the same numbers, and redundant descriptions are omitted.

  The influence analysis device 106 according to the sixth embodiment includes a dependency strength filtering unit (extraction unit) 30 in addition to the functions of the influence analysis device 101. Specifically, the influence analysis device 106 includes an input unit 201, a storage unit M, a dependency relationship extraction unit 2, a weighted dependency strength calculation unit 5, a weighted DSM generation unit 7, and an influence analysis unit. 9, an output unit 202, a display unit 203, and a dependency strength filtering unit 30.

  The dependency strength filtering unit 30 hides the weighted dependency strength smaller than the threshold value T from the weighted DSM 8 based on the threshold value T. In other words, the dependency strength filtering unit 30 compares the threshold T and the weighted dependency strength in the weighted DSM 8 and excludes the weighted dependency strength smaller than the threshold T from the weighted DSM 8. The display unit 203 displays the weighted dependence strength that is not excluded by the dependence strength filtering unit 30 as the weighted DSM8.

  Here, filtering of dependency information will be described. FIG. 17 is a diagram illustrating a configuration example of the weighted DSM according to the sixth embodiment. FIG. 17 shows a configuration example of the weighted DSM 8 in which the dependency information (weighted dependence strength) is filtered with respect to the weighted DSM 8 of FIG.

  For example, when the threshold T is 1.0, the weighted DSM 8 after filtering is the weighted DSM 8 shown in FIG. Specifically, from the weighted DSM 8 in FIG. 5, the weighted dependence strength with the threshold T less than 1.0 is hidden, and a new weighted DSM 8 is displayed on the display unit 203. For example, in the range 24B, only the weighted dependency strength of the cell 17 is left, and the other weighted dependency strengths are hidden.

  As described above, according to the sixth embodiment, since the weighted dependency strength smaller than the threshold value T is hidden from the weighted DSM 8, it is not necessary to see the dependency relationship having a weak weighted dependency strength. For this reason, it is possible to easily focus on the dependency relationship that is stronger than the hidden dependency relationship, and as a result, the analysis work can be highly efficient.

  In addition, you may combine the function of the influence analysis apparatuses 101-106 demonstrated in Embodiment 1-6. For example, an influence analysis device combining the functions of the influence analysis device 102 and the influence analysis device 105 includes an input unit 201, a storage unit M, a dependency relationship extraction unit 2, a weighted dependency intensity calculation unit 5, A configuration including a weighted DSM generation unit 7, an influence analysis unit 9, an output unit 202, a display unit 203, an automatic threshold adjustment unit 20, an analysis data storage device 27, and an automatic weight adjustment unit 28; Become.

  Next, the hardware configuration of the impact analysis apparatuses 101 to 106 will be described. Since the impact analysis apparatuses 101 to 106 have the same hardware configuration, the hardware configuration of the impact analysis apparatus 101 will be described here.

  FIG. 18 is a diagram illustrating a hardware configuration of the influence analysis apparatus. The influence analysis apparatus 101 includes a central processing unit (CPU) 91, a read only memory (ROM) 92, a random access memory (RAM) 93, a display unit 94, and an input unit 95. In the influence analysis apparatus 101, the CPU 91, ROM 92, RAM 93, display unit 94, and input unit 95 are connected via the bus line B.

  The CPU 91 determines a pattern using an influence analysis program 90 which is a computer program. The display unit 94 is a display device such as a liquid crystal monitor, and displays the dependency relationship list 3, the weighted dependency relationship list 6, the weighted DSM 8, the influence range list 12, and the like based on instructions from the CPU 91. The input unit 95 includes a mouse and a keyboard, and inputs instruction information (such as parameters necessary for S / W influence analysis) input from the outside. The instruction information input to the input unit 95 is sent to the CPU 91.

  The influence analysis program 90 is stored in the ROM 92 and loaded into the RAM 93 via the bus line B. The CPU 91 executes the influence analysis program 90 loaded in the RAM 93. Specifically, in the impact analysis apparatus 101, the CPU 91 reads the impact analysis program 90 from the ROM 92 and expands it in the program storage area in the RAM 93 in accordance with an instruction input from the input unit 95 by the user, and performs various processes. Run. The CPU 91 temporarily stores various data generated during the various processes in a data storage area formed in the RAM 93.

  The impact analysis program 90 executed by the impact analysis device 101 includes a dependency relationship extraction unit 2, a weighted dependency strength calculation unit 5, a weighted DSM generation unit 7, and an impact analysis unit 9. These are loaded into the main storage device and these are generated on the main storage device.

  In the first to sixth embodiments described above, the case where the S / W (S / W component) is evaluated has been described. However, the system (system component) may be evaluated. The system component is a component having a dependency relationship other than the S / W component, for example, and is a task constituting a process. When the system component is evaluated, for example, the dependency relationship extraction unit 2 extracts a dependency relationship list from the system source code.

  As described above, the software evaluation apparatus and the software evaluation method according to the present invention are suitable for software structure analysis and influence range analysis, and the system evaluation apparatus according to the present invention is suitable for system structure analysis and influence range analysis. .

DESCRIPTION OF SYMBOLS 1 Source code 2 Dependency extraction part 3 Dependency list 4 Dependency type table 5 Weighted dependence strength calculation part 6 Weighted dependence list 7 Weighted DSM production | generation part 8 Weighted DSM
DESCRIPTION OF SYMBOLS 9 Influence analysis part 12 Influence range list 20 Threshold automatic adjustment part 22 Threshold automatic adjustment part with calculation range 23 Calculation range 25 Template use threshold automatic adjustment part 26 Calculation template 27 Analysis data storage device 28 Weight automatic adjustment part 30 Dependence intensity filtering part 40 File unit weighted DSM
101-106 Influence analysis device 201 Input unit 202 Output unit 203 Display unit M Storage unit T Threshold

Claims (5)

An input unit that accepts input of software source code and weight information indicating the weight that is the importance of the dependency type for each dependency type that is the type of dependency between the software components;
A storage unit for storing source code and weight information input to the input unit;
From the source code in the storage unit, the correspondence relationship between the combination of the component elements, the dependency frequency that is the number of dependency relationships between the component elements, and the dependency type is used as dependency relationship information for each combination of the component elements. A dependency extraction unit to extract;
By calculating the dependency frequency according to the weight by using the dependency relationship information and the weight information according to the weight, the dependency strength corresponding to the combination of the combination of the component elements and the dependency type between the component elements is calculated. A dependency strength calculation unit to
Calculate the correspondence between the dependency strength between the components and the appearance frequency in the combination of the dependency strength, and based on the calculated correspondence relationship, the value of the dependency strength corresponding to the appearance frequency equal to or higher than a predetermined value is calculated. A threshold value setting unit for setting the threshold value of the dependency strength;
Extracting those the dependent intensity from the combinations of the components is equal to or greater than the threshold value, directly or indirectly dependent of the components included in the combination of the extracted component to the correction target of the source code An influence range extraction unit that extracts a range of components that can be traced from the correction target of the source code as an influence range when the source code is corrected;
Equipped with a,
The input unit inputs a range of components used for calculation of the threshold as calculation range information,
The storage unit stores the calculation range information,
The threshold setting unit, wherein for the range of the specified components in the calculating range information to calculate the correspondence between the calculated relationship software evaluation device, characterized in that you set the threshold value based on. The input unit inputs a plurality of calculation templates corresponding to the characteristics of the software as a calculation template of a threshold value calculation method used when calculating the threshold value,
The storage unit stores the calculation template,
The threshold setting unit, using the calculated template corresponding to the software subject to extraction of the influence range, software evaluation device according to claim 1, characterized in that setting the threshold value. Of the dependency strengths calculated by the dependency strength calculation unit, an extraction unit that extracts a dependency strength greater than the threshold;
A display unit for displaying the dependency intensity extracted by the extraction unit;
Software evaluation apparatus according to claim 1 or 2, further comprising a. An input step for receiving input of software source code and weight information indicating the weight that is the importance of the dependency type for each dependency type that is the type of dependency between the components of the software;
A storage step of storing the source code and the weight information;
Dependency that extracts a correspondence relationship between the combination of the component elements, a dependency frequency that is the number of dependency relationships between the component elements, and the dependency type from the source code as dependency relationship information for each combination of the component elements A relationship extraction step;
By calculating the dependency frequency according to the weight by using the dependency relationship information and the weight information according to the weight, the dependency strength corresponding to the combination of the combination of the component elements and the dependency type between the component elements is calculated. A dependency intensity calculating step,
Calculate the correspondence between the dependency strength between the components and the appearance frequency in the combination of the dependency strength, and based on the calculated correspondence relationship, the value of the dependency strength corresponding to the appearance frequency equal to or higher than a predetermined value is calculated. A threshold setting step for setting a threshold of the dependency strength;
Extracting those the dependent intensity from the combinations of the components is equal to or greater than the threshold value, directly or indirectly dependent of the components included in the combination of the extracted component to the correction target of the source code An influence range extraction step of extracting a range of components that can be traced from the correction target of the source code as an influence range when the source code is corrected;
Only including,
In the input step, a component range used for calculation of the threshold is input as calculation range information,
In the storing step, the calculation range information is stored,
In the threshold setting step, the correspondence is calculated for a range of components specified by the calculation range information, and the threshold is set based on the calculated correspondence . An input unit that accepts input of system source code and weight information indicating the weight that is the importance of the dependency type for each dependency type that is the type of dependency between the components of the system;
A storage unit for storing source code and weight information input to the input unit;
From the source code in the storage unit, the correspondence relationship between the combination of the component elements, the dependency frequency that is the number of dependency relationships between the component elements, and the dependency type is used as dependency relationship information for each combination of the component elements. A dependency extraction unit to extract;
By calculating the dependency frequency according to the weight by using the dependency relationship information and the weight information according to the weight, the dependency strength corresponding to the combination of the combination of the component elements and the dependency type between the component elements is calculated. A dependency strength calculation unit to
Calculate the correspondence between the dependency strength between the components and the appearance frequency in the combination of the dependency strength, and based on the calculated correspondence relationship, the value of the dependency strength corresponding to the appearance frequency equal to or higher than a predetermined value is calculated. A threshold value setting unit for setting the threshold value of the dependency strength;
Extracting those the dependent intensity from the combinations of the components is equal to or greater than the threshold value, directly or indirectly dependent of the components included in the combination of the extracted component to the correction target of the source code An influence range extraction unit that extracts a range of components that can be traced from the correction target of the source code as an influence range when the source code is corrected;
Equipped with a,
The input unit inputs a range of components used for calculation of the threshold as calculation range information,
The storage unit stores the calculation range information,
The threshold setting unit, wherein for the range of the specified components in the calculating range information to calculate the correspondence between the calculated relationship system evaluation apparatus characterized that you set the threshold value based on.

Images