JPH0346849B2 - - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to an improvement of a system structure evaluation method for software systems.

In order to develop and maintain a software system, it is necessary to accurately understand the structure of the software system. However, as the scale of software systems increases, it is becoming difficult for designers of the software systems to understand the structure of the entire system.

As a result, it is desired to realize a means for automatically analyzing the structure of a software system from various angles and displaying the result in charts or the like that are easy for evaluators to understand.

[Conventional technology]

Conventional system structure evaluation methods target software systems created in a limited language and automatically count the usage frequency of predetermined instructions used in the software system, or The calling and reference relationships between multiple program modules and multiple data modules structuring the system have only been extracted based on limited key information.

[Problem that the invention seeks to solve]

That is, in conventional system structure evaluation methods, only limited analyzes have been performed on software systems created using limited languages. Therefore, it has been impossible to evaluate the structure of any software system from multiple angles.

[Means for solving problems]

The problem is that the software system to be evaluated, which consists of multiple program modules and multiple data modules, cannot make pre-specified calls.
a first means for extracting key information indicating a reference relationship; a second means for identifying the hierarchical structure of the program modules based on the extraction result of the means; a third means for automatically performing clustering processing for grouping interdependencies between the program modules and data modules based on similarity based on the identification results obtained by the second means; The present invention is characterized in that a fourth means for confirming the normality of the calling relationship between the program modules based on the extraction result is provided, and the processing results of the first to fourth means are displayed. Ru.

[Effect]

That is, according to the present invention, any software system can be analyzed based on specified key information, hierarchical structure analysis, cluster analysis, etc. can be automatically performed, and the normality of the calling relationship between program modules can be confirmed. and display the processing results in charts that are easy for evaluators to understand.

〔Example〕

An embodiment of the present invention will be described below with reference to the drawings. FIG. 1 is a diagram showing a system structure evaluation method according to an embodiment of the present invention, FIG. 2 is a diagram showing an example of a processing process for analyzing the hierarchical structure in FIG. FIG. 4 is a diagram illustrating an example of the processing process of cluster analysis in FIG. 1, and FIGS. 5a to 5e explain the processing process of FIG. It is a diagram that
FIG. 6 is a diagram showing an example of a process for determining the normality of the calling relationship in FIG. 1.

In FIG. 1, a plurality of program modules and a plurality of data modules constituting a software system to be evaluated are stored in a storage unit 1, and an analysis unit 3 is used to perform calls between each program module, or Commands that indicate references from each program module to each data module, hierarchical keys that indicate a hierarchical structure, etc. (for example, "Data called... is defined in the...th snartment called..., and the data is...") When inputting various evaluation key information 2 such as "referenced in The results are stored as a key management table 4 and intermediate data 5. Next, the evaluation control unit 6 sequentially starts the evaluation units 71, 72, 73, etc. for the intermediate data 5,
A required analysis as described below is executed, and the analysis results are displayed on the display unit 8.

First, the processing steps of the evaluation section 71 that analyzes the hierarchical structure of the software system will be explained with reference to FIG.

In FIG. 2, the evaluation unit 71 selects the designated call key from the key management table 4 (step S11), and searches for a program module containing the call key as a call source module (step S12). , and searches for a program module called from the calling module as a called module (step S13).
Based on the search results, the evaluation unit 71 accumulates related information such as the attributes of the calling module and the calling module (step S14). Next, the evaluation section 71
selects a predetermined hierarchical key from the key management table 4 (step S15). The call key is used as the hierarchy key. Based on the hierarchical key, the evaluation unit 71 classifies related information such as the attributes of the calling module and the calling module that are being stored (step S16), and analyzes the hierarchical structure based on the calling relationships obtained as a result of the classification. Hierarchical structure information is accumulated (step S17).

As a result of the above, a hierarchical structure as shown in FIG. 3 is obtained and displayed on the display unit 8 (FIG. 1).

As a hierarchy key, for example, the name given to each program module is Xi for the program module belonging to the first hierarchy, and Yj for the program module belonging to the second hierarchy, as shown in Figure 3. , Zk for program modules belonging to the third hierarchy, etc. When indicating the hierarchy to which it belongs,
It is also possible to use the name of each program module.

Next, the processing steps of the evaluation section 72 that performs cluster analysis of the software system will be explained with reference to FIG.

The purpose of cluster analysis is to reconstruct and display program structures with the aim of reusing software systems, which cannot be achieved by simply analyzing control structures or hierarchical structures. program module,
Additionally, combinations (clusters) of data modules referenced by each program module are automatically grouped and displayed.

In FIG. 4, the evaluation section 72 arranges reference information indicating reference relationships between all program modules and data modules based on the intermediate data 5 (step S21). Figure 5a shows program modules P1 to P6 and data module D1.
It is a figure which shows the reference relationship between thru|or D6. Fifth
In figure a, program module P1 refers to data modules D1, D4, D5 and D6, program module P2 refers to data modules D2, D4 and D5, and in the same way program module P6 Indicates that data modules D2 and D5 are referenced. For the obtained array, the evaluation unit 72 sorts the data modules in the order of the number of reference program modules (step S22). FIG. 5b is a diagram showing the result of classification in step S22 for FIG. 5a.
In FIG. 5b, data module D5 is first referenced from program module 5, then data module D2 is referenced from program module 4, and finally data module D2 is referenced from program module P3 only. They are classified in the order of data modules D3. Next, the evaluation section 72
calculates the reference density σ for each program module. The reference density σ is for each data module D
Binary weighting is performed in the order classified in step S22 from 1 to D6, and the resultant binary number is shown. For example, in FIG. 5b, 2 ⁰ is applied to the data module D3, 2 ¹ is applied to the data module D6,
Similarly, the data module D5 is weighted by ²⁵ , and as a result, the reference density σ of the program module P1 becomes 101110 (binary representation), the reference density σ of the program module P2 becomes 111000, and so on. Therefore, the reference density σ of the program module P6 becomes 110,000. The obtained reference density σ
Based on this, the evaluation unit 72 classifies the program modules in order of reference density σ (step S23). FIG. 5c is a diagram showing the classification results obtained in step S23 with respect to FIG. 5b. In Figure 5c,
Program modules P2 and P5 with maximum reference density σ=111000 first, then program module P3 with reference density σ=110011
However, in the same manner, the program module P4 having the minimum reference density σ=000100 is classified last. Based on the obtained classification results, the evaluation unit 7
2 is the cluster density δ when considering each program module along with adjacent program modules.
(the ratio of the number of combinations of program modules and data modules that are actually referenced to the total number of combinations of program modules and data modules in a cluster) is calculated, and the cluster density δ
Cluster processing is performed within a range that does not divide the specified cluster density δ ₀ to generate an initial cluster (step S24). FIG. 5d is a diagram showing the initial cluster obtained in step S24 with respect to FIG. 5c.
Based on the obtained initial cluster, the evaluation unit 72 adjusts the classification of data modules having reference information not included in the initial cluster, and performs cluster optimization (cluster density δ of all clusters included in the software system (Step S25). Figure 5 e
5 is a diagram showing the optimization performed in step S25 for FIG. 5d. The optimized cluster obtained in the above manner is displayed on the display section 8 (FIG. 1).

Next, the processing steps of the evaluation section 73 for checking the normality of the calling relationships between program modules will be explained with reference to FIG.

In FIG. 6, the evaluation unit 73 extracts the calling program module to be checked (step S31), and extracts information related to the specified key (1) in the extracted calling program module (step S31). S32). Next, the evaluation unit 73 extracts the program module that is the call destination to be checked (step S33), and selects the designated key (2) in the extracted call destination program module.
related information is extracted (step S34). Next, the evaluation unit 73 collates the extracted related information of key (1) and key (2) (step S35), and determines the set relationship between the related information of key (1) and key (2). (step S36), and the normality of the calling relationship between both program modules is determined based on the presence or absence of a contradiction. The determination result is also displayed on the display section 8 (FIG. 1).

As is clear from the above description, according to this embodiment, by setting the predetermined evaluation key information 2 for any software system stored in the storage unit 1, the evaluation units 71, 72, 73, . . . This makes it possible to analyze the software structure from multiple angles, and the analysis results are displayed on the display unit 8. Note that the display unit 8 displays the analysis results by the evaluation units 71, 72, 73, etc. in a three-dimensional diagram, thereby enabling the evaluator to make an accurate evaluation.

Note that FIGS. 1 to 6 are only one embodiment of the present invention, and the hierarchical structure obtained by the evaluation unit 71, for example, is not limited to that shown in the figures, and may be modified in many other ways. However, the effects of the present invention remain the same in either case. Also, the evaluation section 72
The clusters obtained by this method are not limited to those shown in the drawings, and many other modifications may be considered, but the effects of the present invention remain the same in any case.

〔Effect of the invention〕

As described above, according to the present invention, by specifying the required evaluation key information for any software system, it is possible to analyze the hierarchical structure, cluster analysis, and call relationship normality from multiple angles, and the software Accurate evaluation of the system becomes possible.

[Brief explanation of drawings]

Fig. 1 is a diagram showing a system structure evaluation method according to an embodiment of the present invention, Fig. 2 is a diagram showing an example of the processing process for analyzing the hierarchical structure in Fig. 1, and Fig. 3 is a diagram showing the processing process in Fig. 2. Figure 4 is a diagram explaining
Figures 5a to 5e are diagrams illustrating the processing process of Figure 4, and Figure 6 is a diagram illustrating the normality of the calling relationship in Figure 1. It is a figure showing an example of a processing process. In the figure, 1 is a storage unit, 2 is evaluation key information,
3 is an analysis section, 4 is a key management table, 5 is intermediate data, 6 is an evaluation control section, 71, 72, 73, ... are evaluation sections, 8 is a display section, D1 to D6 are data modules, P1 to P6 are program module,
S is step, X, Y1 to Yn and Z1 to
Zm indicates the program name.

Claims (1)

[Claims] 1. A first means for extracting key information indicating a pre-specified interdependence relationship from a software system to be evaluated consisting of a plurality of program modules and a plurality of data modules, and based on the extraction result of the first means. a second means for identifying the hierarchical structure of the program modules; and a mutual dependence relationship between the program module and the data module based on the extraction result by the first means and the identification result by the second means. a third means for automatically performing a cluster analysis to group the program modules based on similarity; and a fourth means for confirming the normality of the calling relationship between the program modules based on the extraction result by the first means. and displaying the processing results of the first to fourth means.