JPH06119205A - Quality analyzing device for software and supporting device for providing the same with high quality - Google Patents

Abstract

PURPOSE:To contrive to conduct objective analyzing, evaluating and advice- imparting to a program structure by using the metrics, and to provide the device with a high quality. CONSTITUTION:An analytic processing part 8 measures the metrics of a software stored in an analytic object system store part 1 by using the metrics relating to the operation structure of variable data, in addition to the metrics relating to the program quantity and the metrics relating to a control structure which are used up to the present, and stores the result of measurement in a metrics storage part 10 at every program and at every module for constituting a program, executes the analytic processing of the software, based on this stored metrics, and outputs the contents of the analytic processing to an analytic result store part 2 or an output part 5. In this regard, the metrics relating to the operation structure of this variable data consist of each separate use type frequency of the variable data classified into the utilization types of the variable data, and a data scope value obtained by measuring in how many modules each of the variable data is used, based on a module call relation stored in a module call relation storage part 9.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention uses a metric for analyzing / evaluating the quality of software to perform a software quality analysis / evaluation, a test support, and a quality improvement support, and a software quality analysis device for improving the quality. Regarding the support device.

[0002]

2. Description of the Related Art Generally, "Metrics" is used as a measure for quantitatively grasping the structure of programs constituting software, and the "Metrics" are used to grasp the structural characteristics of programs before or during software testing.
We are working to improve the quality of programs and improve the efficiency of testing. Here, the conventional "metrics" are classified into two types: metrics related to the program amount such as the number of steps and data of the program, and metrics related to the control structure such as the number of condition determinations in the program.

However, the structural characteristics of the program cannot be sufficiently grasped only by these two types of metrics. That is, these two types of metrics can grasp the outline of the program, but cannot analyze the structural characteristics of the program to the extent that they can point out the points for improving the program and the important test points.

Therefore, when the manager who oversees software development evaluates the quality of the program, it often depends on the performance of the programmer, and it is difficult to evaluate the program itself.

On the other hand, the programmer knows how to improve the program created by himself, since it is difficult to objectively evaluate the created program only with the above metrics. It was difficult to set test items that can efficiently detect errors in the integration test work, or to find an efficient error cause when debugging.

As a result, in reality, the manager judges whether or not the program making up the software is in a state in which the integration test can be carried out, based on the experience of the programmer and the past record of the manager itself. Also, the programmer
The programmer himself evaluated the program objectively from the structure of the program, and found the improvement point of the program.

Further, the quality analysis / evaluation of software, the change for reviewing / improving the program before the software integration test, and the test execution work have been carried out separately.

Particularly, since the quality of software is evaluated by measuring the amount of error at the time of the integration test,
Many defects were found during the integration test, making it difficult to find errors efficiently and effectively.

That is, the conventional software analysis / evaluation relies heavily on the subjective experience of managers and programmers, and has been reluctant to develop errors by discovering errors through integration tests. However, there is a problem that it is difficult to create high quality software.

[0010]

As described above, in the conventional software analysis / evaluation, it is difficult to sufficiently grasp the structural characteristics of the program only by the conventional metrics. It often takes a long time to develop software, and it is difficult to create high-quality software because the software development is reluctant to find errors by integration tests. There was a problem.

Therefore, the present invention eliminates such a problem, presents a metric that can grasp the program structure sufficiently, and objectively analyzes and evaluates the sufficient program structure by this metric, and as a result, Due to
A software quality analyzer that can objectively analyze / evaluate programs and give advice to managers or programmers by pointing out points for program improvement and important test points, and informing points to be noted when debugging. It is also an object of the present invention to provide a quality improvement support device for the same.

[0012]

According to the present invention, there is provided a software quality analysis device for performing software quality analysis, comprising first metric storage means for storing metrics relating to a program amount and metrics relating to a control structure. Counting the metrics stored in the first and second metrics storage means for each program that constitutes the software, and for each module that constitutes the program. However, it is characterized by comprising an analysis processing means for performing quality analysis processing for each program and each module, and an analysis result storage means for storing an analysis processing result by the analysis processing means.

Further, a module call relation storage means for storing a call connection relation between modules in a program forming the software as a tree structure, and a metric relating to an operation structure of variable data counted by the analysis processing means is , The use type frequency of the variable data counted according to a predetermined use type classified according to how the variable data is used, and the predetermined variable data is used in the tree structure stored in the module call relation storage means. The data scope value is given by the sum of the number of existing modules and the minimum number of connection branches for call connection between the modules.

Further, the analysis processing means is characterized by performing graphing processing for each program and each module based on the metrics stored in the first and second storage means.

Further, a metric related to the program amount,
A metric storage means for storing a metric for a control structure and a metric for an operation structure of variable data, and a metric stored in the metric storage means are counted, and quality analysis processing is performed for each program and each module constituting the program. Analysis processing means, analysis result storage means for storing the analysis processing result by the analysis processing means, rule storage means for storing the evaluation and advice rules of the program, and quality analysis processing stored in the analysis result storage means It is characterized by comprising an evaluation / advice processing means for evaluating the program and giving an advice based on the result and the rule stored in the rule storage means.

[0016]

According to the present invention, in addition to the metrics related to the amount of programs and the metrics related to the control structure that have been conventionally used, the metrics related to the operation structure of variable data are measured, and software is analyzed and processed based on this metric. The analysis processing contents are output as a file or displayed and further displayed as a graph. Metrics related to the operation structure of this variable data are defined and assigned.
It consists of frequency by usage type of variable data classified into usage type of variable data such as "reference" and "reference", and data scope value that measures how each variable data is used beyond the module. , The usage of variable data allows us to see the operation method of various variable data in a program in detail, and the scope of variable data can be evaluated from the data scope value.

Further, based on the analysis processing result by the metric including the metric relating to the operation structure of the variable data, the processing of software evaluation and advice provision based on the evaluation and advice rules stored in the rule storage means in advance. I do. At this time, the software evaluation is presented for each program that constitutes the software, and the advice is
It is possible to select the advice to the manager or the programmer, and the reason for the advice content and the point for improving the quality are presented by the selection.

[0018]

Embodiments of the present invention will be described below with reference to the drawings.

FIG. 1 is a block diagram showing the configuration of a software quality analysis apparatus according to an embodiment of the present invention. In FIG. 1, a software quality analysis device includes an analysis target system storage unit 1 that stores an analysis target system that is analysis target software configured of a single program or a plurality of programs, and an analysis result storage unit 2 that stores analysis results as files. , A file management unit 3 for managing file input / output of the analysis target system storage unit 1 and the analysis result storage unit 2, an input unit 4 for inputting with an input device such as a keyboard or a mouse, and an output including an output device such as a CRT or a printer An input / output control unit 6 that controls input / output of the unit 5, the input unit 4, and the output unit 5, an analysis processing unit 8 that performs analysis processing of the analysis target system, and a work area necessary for analysis processing of the analysis processing unit 8 are secured. The module call relationship of each program forming the analysis target system, which is obtained by the temporary storage unit 7 and the analysis processing unit 8; Module call relation storage unit 9 for storing ie tree structure, and a metrics memory unit 10 for storing the metrics analysis systems obtained by the analysis processing unit 8.

Further, the metrics storage unit 10 stores the number of steps storage unit 11 for storing the number of steps of each program and module constituting the analysis target system, and the number of variable data (hereinafter referred to as "data") of each program and module. Data number storage unit 12, condition determination number storage unit 13 that stores the condition determination number of each program and module, data usage type frequency storage unit 1 that stores the usage type frequency of data for each program and module
4 and a data scope value storage unit 15 that stores a data scope value for each data.

Next, the operation of the software quality analysis device will be described with reference to the flow chart. In the following description, an office work program as an analysis target system, particularly an office work program described in the programming language "COBOL" will be described as an example. This business processing program has a data section (DATA DIVISION) and a procedure section (PROCEDURE DIVISION), and the data section is composed of several modules.

FIG. 2 is a flowchart of software quality analysis processing. In FIG. 2, first, the analysis target system stored in the analysis target system storage unit 1 is read (step 101). Then, the analysis processing unit 8 obtains the module call relationship for each program constituting the read analysis target system and stores it in the module call relationship storage unit 9 (step 102). This module call relationship is stored in a tree structure. Further, the number of steps for each program and each module of each program is obtained and stored in the step number storage unit 11 (step 10
3).

Next, the analysis processing unit 8 takes out one program which constitutes the read analysis target system (step 104) and analyzes the data part of the taken-out program (step 105). Further, the procedure section of the fetched program is analyzed (step 106). Next, the data scope value for each data is calculated from the module call relationship and the retrieved data (step 10).
7). Then, the analysis result of this program is stored in the analysis result storage unit 2 as an output file. Then, it is determined whether or not there is another program that constitutes the analysis target system (step 109), and if there is another program, the process proceeds to step 104, and another program is analyzed. On the other hand, when there is no other program that constitutes the analysis target system, the analysis processing unit 8 performs analysis processing for each program that constitutes the analysis target system, and the analysis result storage unit 2 outputs the result as an output file of the analysis target system.
(Step 110), and this processing ends.

FIG. 3 is a flow chart of the analysis process of the data part in step 105 of FIG. In FIG. 3, it is first determined whether or not there is a data part to be analyzed (step 201). If there is no data part, this process is terminated, and if there is still a data part, another program is quoted. It is determined whether or not (step 202),
If no other program is cited, the data name is extracted (step 205) and the number of data is counted (step 206). Then, the data name extracted in step 205 is output to the tree structure of the module call relationship stored in the module call relationship storage unit 9 (step 207). Then, it transfers to step 201 and repeats a process.

On the other hand, if another program is cited in step 202, it is further judged whether or not there is a cited program (step 203). If there is no cited program, it means that there is no cited program. Message is displayed and output (step 204), then step 2
The processing is repeated by shifting to 01. On the other hand, step 20
In step 3, if there is a citation program, it is further determined whether or not the citation program has ended (step 20).
8) Unless the quoting program ends, the data name is fetched (step 209), the number of data is counted (step 210), and the data name fetched in step 209 is stored in the module call relation storage unit 9. Output to a tree structure related to module call (step 21)
1). Then, in step 208, when the quotation program ends, the process proceeds to step 201, and the process is continued as long as there is a data section.

FIG. 4 is a flow chart of the analysis processing of the procedure part in step 106 of FIG. In FIG. 4, first, it is determined whether or not there is a procedure section (step 301), and if there is no procedure section, this processing ends.
On the other hand, if there is still a procedure section, the module name that constitutes the program is taken out (step 302), and the number of modules is counted (step 303). Then, it is judged whether or not the module is completed (step 3
04), if the module is finished, step 301
And the process is repeated.

On the other hand, in step 304, when the module is not completed, each part constituting the module is
It is judged whether or not it matches the usage type (pattern) of the data of "definition / substitution" (step 305). If they do not match, the process proceeds to step 307. If they match, the data of "definition / substitution" is determined. After counting as (step 30
6) and shifts to step 307. In step 307, it is determined whether or not it matches the usage type of the “reference with condition determination” data. If they do not match, the process proceeds to step 310. If they match, the data is referred to as “reference with condition determination”. After counting (step 308) and counting as the condition determination number (step 309), the process proceeds to step 310. In step 310, it is determined whether or not it matches the usage type of the "argument" data. If they do not match, the process proceeds to step 313, and if they match, they are counted as "argument" data (step 311). After counting the number of program calls (step 312), the process proceeds to step 313. In step 313, it is determined whether or not it matches the usage type of the "other" data. If they do not match, the process proceeds to step 315, and if they match,
After counting as "other" data (step 31
4) and shifts to step 315. And step 3
At 15, the usage types of the data that do not apply to steps 305, 307, 310 and 313 are counted as the “reference” data. Then, the process proceeds to step 304, and the process is repeated unless the module ends.

Next, the calculation of the data scope value in step 107 of FIG. 2 will be described.

FIG. 5 is a diagram for explaining the calculation of the data scope value. First, FIG. 5A shows a tree structure showing the module call relationship in one program. The tree structure shown in FIG. 5A corresponds to step 102 in FIG.
It is obtained and stored in (1) and is composed of nodes n1 to n7 and arcs a1 to a7. Each node represents a program root and each module, and each arc represents a program root or a connection relationship between each module.

In FIG. 5B, attention is paid to one data X in the root of the tree-structured program or the data name used in a predetermined module in step 207 or step 211 of FIG. The parts used are indicated by diagonal lines.

FIG. 5C shows a newly created tree structure including the nodes using the data X in the tree structure of FIG. 5B and having the minimum number of arcs. here,
All nodes n1 to n7 in FIG.
Let all arcs a1 to a7 be an arc set A. In addition, the shaded nodes n1, n3, n4, and n6 in which the data X in FIG. 5C is used are set as a node set NN, and arcs a1 to a1 indicating the connection relationship regarding the data X.
Let 3, a5 be the arc set AA. In this case, the number of node sets NN is 4, and the number of arc sets AA is 4.

Here, the data scope value DS (X) regarding the data X is defined as follows. That is, DS (X) = | NN | + | AA | Therefore, the data scope value for the data X in the tree structure in FIG. 5 is as follows. That is, DS (X) = 4 + 4 = 8. Since this data scope value indicates a calling relationship between modules in which the same data is used, a large data scope value means that an error is likely to occur.

Next, a display example of the output file for each module and each program will be shown.

FIG. 6 is a diagram showing the output file display processing result for each module. In FIG. 6, FIG.
FIG. 6 is a diagram showing an example of display output of metric values of the number of steps, the number of data, and the number of condition judgments for each module for one program, and FIG. 6B shows the data for each module for each data for one program. It is a figure which shows the example of a display output of the frequency according to usage type, and FIG.6 (c) is a figure which shows the example of a display output of the data scope value with respect to each data of one program.

In FIG. 6A, for example, the module M
The number of steps for 1 is 48, the number of data is 32, and the number of condition determinations is 5. And the total number of steps of this program is 2425,
The total number of data is 110, and the total number of condition determinations is 35. This metric value is a conventionally used metric value.

In FIG. 6B, for example, for the data D1, the module M1 is "1D2R0C0P0E".
The module M2 is
The data usage type frequency is “0D1R0C1P0E”, and this program indicates that the data usage type frequency is “3D5R2C1P0E”. here,
The data usage type frequency is expressed in the following format. That is, “αDβRγCδPεE” where α: number used for “definition / substitution” β: number used for “reference” γ: number used for “reference with conditional judgment” δ: number used for “argument” ε : It is the number used for "Others" and is classified into five data usage types, and the number used for each classification is shown. In addition, D, R, C, P, E
Are symbols for identifying the numbers α, β, γ, δ, ε, respectively.

The types of data used classified into five types of D, R, C, P and E are easy to understand by referring to a part of the following program. That is, if IF data 1> 0 THEN MOVE data 2 TO data 3 CALL "program name" USING data 4 INSPECT data 5 TALLING ........., data 3 is a data usage type D, data 2 is data usage type R, data 1
Is a data usage type C, data 4 is a data usage type P, and data 5 is a data usage type E.

That is, the "definition / assignment" of the data usage type D is the data when the value is rewritten, and the "reference" of the data usage type R is only referenced, and the value is rewritten. There is no data. However, it is excluded when it is used for condition judgment. The “reference with condition determination” of the usage type C of data is data that is referred to in the condition determination, and the value is data that cannot be rewritten. Furthermore, the "argument" of the usage type P of data is data when used as an argument when calling another program,
The "others" of the data usage type E are D, R, C, P above.
It is data when it is used for other than the usage type of the above four data.

FIG. 6C shows that the data scope value for the data D1 is 8, for example.

In this way, the metric value for each module for one program can be known.

Next, FIG. 7 is a diagram showing the output file display processing result for each program. In FIG.
7A shows the metric values of the number of steps, the number of data, the number of condition judgments, the number of modules, and the number of program calls for each program, and FIG. 7B shows the data usage type for each program with respect to the data. The frequency is displayed in terms of the number of uses, and FIG. 7C shows the frequency by type of data use for each program in percentage.

In FIG. 7A, for example, the program P1 shown in FIG. 6 has 2425 steps, 110 data, 35 condition determinations, 53 modules, and 53 programs. It indicates that the number of calls is two.

In FIG. 7B, for example, regarding the data D1, the program P1 is "1D2R0C0P0E", the program P2 is "0D1R0C1P0E", and the analysis target system is "3D5R2C1P0".
The frequency of each data usage type "E" is displayed.

In FIG. 7 (c), since the data usage types are classified into five, the frequency by data usage type for each program for each of these five data usage types is displayed as a percentage, for example, "definition / substitution". It is shown that the program P1 uses 54%, the program P2 uses 58%, and the analysis target system uses 56% regarding the data usage type of “.

As will be described later, by displaying these metric values in a graph, the characteristics of the analysis target system in general, each program configuring the analysis target system, and the configuration of the modules configuring each program can be clearly seen. You can know

In particular, in addition to the metrics of the program amount such as the number of steps and the number of data and the metrics of the control structure such as the number of condition judgments, the metrics of the operation structure of the variable data according to the frequency by the data usage type and the data scope value. By using, the characteristics of the analysis target program can be known in detail.

Next, a detailed description will be given based on a specific example of the system to be analyzed.

FIG. 8 is a diagram showing the structure of the program ASCENDANT, and FIG. 9 is a program MINIMU.
It is a figure which shows the structure of M. Here, the analysis target system is composed of the two programs shown in FIGS. 8 and 9. Note that the analysis target system is an office processing program written in the programming language "COBOL".

In FIG. 8, the program ASCENDA
NT (PG1) has a data part DD1 and a procedure part PD
1 and the procedure unit PD1 has GETDATA, M
It is composed of four modules ma1 to ma4 of AIN, PREPARE, and PRINT.

Further, in FIG. 9, the program MINI
The MUM (PG2) has a data section DD2 and a procedure section P.
D2, and the procedure unit PD2 has START, CO
It is composed of four modules mb1 to mb4 of MPARE, EXCHANGE, and FINISH.

Next, the above two programs PG1 and PG
2 shows a display example of an output file by the quality analysis device of this software for 2 and an analysis target program.

FIG. 10 is a diagram showing a display example of an output file for the program PG1. In FIG. 10, FIG. 10A shows a metric value for each module that constitutes the program PG1, and FIG. 10B shows a data usage type frequency for each module of the program PG1 in terms of the number of uses. Yes, FIG. 10C shows the program PG
2 shows the data scope value of each data item 1.

In FIG. 10A, as in the conventional case, for example, the number of steps of the module name "GETDATA" is 9, the number of data is 9, and the number of condition determinations is 0. .

In FIG. 10B, for example, the data name “DATAA” is used once as the “definition / substitution” data in the module name “GETDATA”.
The module name "PREPARE" indicates that the data is used once as "reference" data, and is used once as "definition / substitution" and "reference" data in the entire program.

In FIG. 10C, the data scope values of the data names "I" and "N" are 1, and the data scope values of the data names "DATAA" to "DATAI" and "ARRAY" are the same. It shows that each is 5.

FIG. 11 is a diagram showing a display example of an output file for the program PG2. In FIG. 11, FIG. 11A shows a metric value for each module that constitutes the program PG2, and FIG. 11B shows a data usage type frequency for each module of the program PG2 in terms of the number of uses. Yes, FIG. 11C, program PG2
The data scope value of each data is shown.

In FIG. 11A, as in the conventional case, for example, the number of steps of the module name "START" is 2, the number of data is 2, and the number of condition determinations is 2. .

In FIG. 11B, for example, the data name "I" is used once in the module name "START" as the data "definition / substitution" and the data "reference with conditional judgment", respectively. Module name "C
Used once as "reference" data in OMPARE, once as "definition / substitution" data in module name "EXCHANGE", and twice as "definition / substitution" data in the entire program. , Is used once as the data of “reference” and once as the data of “reference with condition determination”.

Further, in FIG. 11C, the data scope value of the data name "I" is 5, and the data name "TM"
The data scope values of "P" and "IMAXN" are 1 respectively, and the data names "IMIN" and "ARRA"
The data scope values of “Y” are 3, respectively.

FIG. 12 is a diagram showing a display example of an output file for each program of the analysis target program. 12
12A, the metric value for each program is shown, and FIG. 12B shows the data usage type frequency for each program in terms of the number of uses.
(C) shows the frequency of each usage type for each combination of various data types of the analysis target system for each program.

In FIG. 12A, as in the conventional case, for example, the number of steps of the program name "ASCENDANT" is 49, the number of data is 13, and the number of condition determinations is 2. .

In FIG. 12B, for example, the data name "DATAA" is the program name "ASCENDAN".
In "T", "definition / substitution" data and "reference"
One piece of data is used for each, and the program name "MI
NIMUM ”indicates that it is not used.

Further, in FIG. 12 (c), for example, the program name "ASCENDANT", "definition / substitution"
Data and data used as “reference” data, that is, there are 10 data types “D & R”.

Here, the program name "ASCENDAN
The data scope will be described by taking concrete data as an example for each of "T" and the program name "MINIMUM".

FIG. 13 shows the program name "ASCENDA
It is a figure explaining the data scope value with respect to the data DATAA of "NT". Here, FIG. 13A is a tree structure diagram showing the module configuration. In FIG. 13B, the module in which the data DATAA is used in the module configuration shown in FIG. FIG. 13C includes a module using the data DATAA indicated by the hatched lines in FIG. 13B,
The module connection relationship with the minimum number of arcs is shown. Therefore, based on the tree structure for the data DATAA in FIG. 13C, the number of modules including the data DATAA is 2, and the required number of arcs is 3. Therefore, the data scope value DS (DATAA) of the data DATAA is Is 5
It turns out that

Further, FIG. 14 shows the program name "MINI
It is a figure explaining the data scope value with respect to the data IMIN of "MUM". Here, FIG. 14A is a tree structure diagram showing the module configuration. In FIG. 14B, the module in which the data IMIN is used in the module configuration shown in FIG. 14A is indicated by diagonal lines. FIG. 14C shows a module connection relationship in which the number of arcs is minimum, including the module using the data IMIN shown by the shaded area in FIG. 14B. Therefore, based on the tree structure regarding the data IMIN in FIG. 14C, the number of modules including the data IMIN is 2, and the required number of arcs is 1, so that the data IMI
It can be seen that the data scope value DS (IMIN) of N is 3.

The output files for each of the two programs that make up the analysis target system and the output files for each of the modules that make up this program are the analysis result storage unit 2.
Will be stored in.

Next, an example of graphing processing based on these output files will be shown.

FIG. 15 shows FIG. 10 (c) and FIG.
It is a figure which shows the example which graph-processed the data scope value of (c). Here, FIG. 15A shows the program name “A
This is an example in which the data scope value of “SCENDANT” is bar-graphed, and FIG.
It is a figure which shows the example which processed the data scope value of "NIMUM" into a bar graph. When the graphing process is performed and displayed in this way, the distribution of the data scope values can be understood, and the characteristics of the program can be easily understood.

FIG. 16 is a diagram showing an example of graphing the usage type-based frequency for each program of FIG. 12 (c). Here, FIG. 16A shows the program name “ASCE
It is a figure which shows the example which processed the frequency according to the usage type of "NDANT" into the pie chart, and FIG.16 (b) is a program name "MIN.
It is a figure which shows the example which carried out the pie chart process about the frequency according to the usage type of "IMUM." This pie chart is a pie chart processing of what kind of combination of data usage type data the data of each program is used, so it is possible to duplicate the usage of different data usage types. It is possible to easily understand the program structure that exists.

Next, based on the output file analyzed in this way, evaluation of the system to be analyzed and high quality of software for giving advice to the manager or programmer to support high quality of software The conversion support device will be described. FIG. 17 is a configuration block diagram of a software quality improvement support device. 17, the analysis result storage unit 22 stores an output file which is the analysis result of the software quality analysis apparatus shown in FIG. Further, the evaluation / advice standard 21 stores rules for evaluation and advice giving. Then, the evaluation / advice processing unit 20 reads the target output file for evaluation / advice, and refers to the evaluation / advice criteria 21 for the analysis result of this output file to perform the evaluation / advice addition processing. Incidentally. Input unit 24, output unit 25
The input / output control unit 26 is the same as the input unit 4, the output unit 5, and the input / output control unit 6 in FIG.

Next, the operation of the software quality improvement support apparatus will be described with reference to a flowchart.

FIG. 18 is a flowchart of the evaluation / advice giving process of the software quality improvement support device. FIG.
First, the output file of the evaluation target program is extracted from the analysis result storage unit 22 (step 40).
1). Then, the program structure relating to the module call is displayed and output from the output unit 25 as a tree structure indicating the program structure (step 402). Next, based on the analysis result of the extracted output file, the evaluation / advice processing unit 20 refers to the evaluation rule 410 of the evaluation / advice criteria 21 and outputs and outputs the evaluation result (step 403).

Then, it is judged whether or not advice is given to the manager (step 404). If advice is not given, the process proceeds to step 406. If advice is given, advice of the evaluation / advice standard 21 is given. After giving advice to the manager by referring to the rule 411 (step 405), the process proceeds to step 406. Step 406
Then, it is determined whether or not the advice is given to the programmer. If the advice is not given, this process is terminated, and if the advice is given, the advice rule 4 of the evaluation / advice standard 21 is given.
After giving advice to the programmer with reference to step 12 (step 407), this processing is ended.

Next, the advice giving in steps 405 and 407 in FIG. 18 will be described.

FIG. 19 is a flow chart for giving advice to the manager in step 405. In FIG. 19, first, the first rule concerning the manager of the advisory rule is applied (step 501). Next, it is determined whether or not the condition part of this advice rule is true (step 502). If it is not true, the process proceeds to step 504. If it is true, the advice part is displayed and output by the advice rule execution part. After that (step 503), the process proceeds to step 504. Then, in step 504, it is determined whether or not there is a next advice rule. If there is an advice rule, this new advice rule is applied (step 505), the process proceeds to step 502, and the number of advice rules is repeated. on the other hand,
In step 504, if there is no next advice rule, this processing ends.

FIG. 20 is a flow chart for giving advice to the programmer in step 407. In FIG. 20, first, the first rule relating to the advisory rule programmer is applied (step 601). Next, it is judged whether or not the condition part of this advice rule is true (step 602), and if it is not true, the process moves to step 604, and if it is true, the advice part is displayed and output by the execution part of this advice rule. After that (step 603), the process proceeds to step 604. Then, in step 604, it is determined whether or not there is a next advice rule, and if there is an advice rule, this new advice rule is applied (step 605), the process proceeds to step 602, and the number of advice rules is repeated. on the other hand,
If there is no next advisory rule in step 604, this process ends.

Here, the evaluation or advisory rule is as follows. For example, RULE1: “Definition / substitution” and “reference” data is large. IF Y = ((D AND R) / DATA × 1
00) ≧ X THEN RULE1 = true BECAUSE: The data used for “definition / substitution” and “reference” not only change the value due to definition / substitution and increase the number of internal states, but also the timing of reference Problem (a new value has already been assigned when you want to refer to it), and if X is large, it tends to be corrected or changed. ) This rule is the number of data (D AND R) used for “definition / assignment” and “reference” in the “IF” statement.
Of the total number of data (DATA) (Y%) is X%
If the ratio Y is equal to or more than X, it is determined that the rule is true and the program has a large number of “definition / substitution” and “reference” data for the target program. To do. Then, if necessary, "BEC
Refer to the contents of the reason for evaluation judgment of "AUSE".

Further, as an example of the rule, RULE2: a large number of global data IF Y = SCOPE2 / DATA × 100) ≧
X THEN RULE2 = true BECAUSE: (Omitted) This rule is that in the "IF" statement, the total number of data (D) whose data scope value is 2 or more (SCOPE2)
ATA) ratio (Y%) is greater than or equal to X%. If the ratio Y is greater than or equal to X, this rule is set to true and a program that has a large number of global data in the target program. Evaluate to be. Then, if necessary, the evaluation determination reason content of “BECAUSE” omitted is referred to.

Note that a plurality of such rules are
It is held in advance in the advice standard 21.

Next, the processing results of the program evaluation and the advice giving processing by the quality improvement support device will be described.

FIG. 21 shows the program "AS" shown in FIG.
It is a figure which shows the display output example of the evaluation result of "CENDANT." In FIG. 21, the features of the program structure are: Most of the data is used for "definition / substitution" and "reference". 2. The ratio of data with a data scope value of 2 or more is large. The evaluation result is obtained. This is because the rule that a large amount of data is used for “definition / assignment” and “reference” is true, and the rule that a ratio of data having a data scope value of 2 or more is large is true. This means that from the graphs of FIG. 15 (a) and FIG. 16 (a), there are 10 data whose data scope value is 5,
It can be understood from the fact that the data used for “substitution” and “reference” is 83.3%. Further, as the predicted value of the correction / change amount, the value of 31% of the program is displayed, which is predicted and calculated from the past data.

FIG. 22 shows the program "ASCENDAN".
It is a figure which shows the example of a display output of the advice result of "T". here,
FIG. 22A is an example of display output of advice to the manager, and FIG. 22B is an example of display output of advice to the programmer.

In FIG. 22 (a), the rule of the advice M1 "It is necessary to consider again how to divide the module" and the advice M2 "Definition / substitution" and "reference".
This is because the rule that "a detailed review is necessary focusing on data" is true.

Further, in FIG. 22B, similarly to the advice to the manager, the rule of the advice P1 "It is necessary to consider again how to divide the module" and the advice P2 of "" Definition / substitution ". Moreover, the rule that "a detailed review is necessary centering on" reference "data" was true. Further, in the case of advice to the programmer, the analysis result which is the basis of the advice is also displayed in detail.

FIG. 23 shows the program "ASCENDAN".
It is a figure which shows the example of a display output of the advice to quality improvement of "T".
Here, FIG. 23A shows an example of display output of advice for high quality to the manager, and FIG. 23B shows an example of display output of advice for high quality to the programmer. ing.

In FIG. 23 (a), "It is necessary to reconsider how to divide modules" of the advice HM1.
The result is displayed because the rule for improving quality and the rule for improving quality called “Need a detailed review centered on the data“ ARRAY ”” of Advice HM2 are true. There is.

Further, in FIG. 23 (b), the advice H
This is because the high quality rule of "Let's review how to divide modules again" of P1 and the rule of "Let's be careful about data" ARRAY "" of advice HP2 were true, and the result is displayed. There is.

Here, in the case of advice for improving the quality to the programmer, the "reason" and "point" selection sections S1 to S4 for the advice, that is, icons are provided on the display screen of the display section, and this icon is displayed. By selecting with the mouse, a new window will open, displaying the reason for the selection or the contents of point.

FIG. 24 is a diagram showing a display when the reason for the advice HP1 on the display output screen of FIG. 23 (b) is selected, and FIG. 25 is shown for the advice HP2 of the display output screen of FIG. 23 (b). It is a figure which shows the display when a point is selected. That is, the content of the reason for the advice HP1 is displayed in the window W1 of FIG. 24, and the content of the points for the advice HP2 is displayed in the window W2 of FIG.

As described above, by using the analysis result of the program, it is possible to obtain general evaluation of the program and advice to the manager or the programmer based on the rule.

In the above embodiment, the software quality analysis device and the software quality improvement support device have been described separately, but these two devices may be combined into one device. Rather, the process of evaluation and advice from the program quality analysis is a continuous process, so 1
Two devices are more efficient.

[0093]

As described above, according to the present invention, in addition to the metrics related to the program amount and the metrics related to the control structure which have been conventionally used, the metrics related to the operation structure of variable data are measured, and this metric is also measured. The software is subjected to analysis processing, and the analysis processing contents are output or displayed as a file, and further displayed and output as a graph. The metrics related to the operation structure of the variable data are classified into the usage types of the variable data such as "definition / assignment" and "reference", and the frequency by type of usage of the variable data and the number of modules used for each variable data. It is made up of the data scope value that measures whether or not the variable data is used, and the operation method of various variable data in the program can be seen in detail from the frequency of usage of variable data, and the global degree of variable data is evaluated from the data scope value. Therefore, there is an advantage that the quality of the program that constitutes the software can be essentially and objectively analyzed and evaluated.

Further, based on the analysis processing result by the metric including the metric relating to the operation structure of the variable data, the processing of software evaluation and advice provision based on the evaluation and advice rules stored in the rule storage means in advance. I do. At this time, the software evaluation is presented for each program that constitutes the software, and the advice is
It is possible to select the advice to the manager or the programmer, and the reason for the advice content and the point for improving the quality can be presented by the selection, so that effective program modification / correction before and during the software integration test This has the advantages that changes can be made, the time required for software development can be shortened, and high-quality software can be created.

[Brief description of drawings]

FIG. 1 is a block diagram showing the configuration of a software quality analysis device according to an embodiment of the present invention.

FIG. 2 is a software quality analysis processing flowchart.

FIG. 3 is an analysis processing flowchart of a data part in step 105 of FIG.

FIG. 4 is an analysis processing flowchart of a procedure section in step 106 of FIG.

FIG. 5 is a diagram illustrating calculation of a data scope value.

FIG. 6 is a diagram showing a display processing result of an output file for each module.

FIG. 7 is a diagram showing a display processing result of an output file for each program.

FIG. 8 is a diagram showing a configuration of a program ASCENDANT.

FIG. 9 is a diagram showing a configuration of a program MINIMUM.

FIG. 10 is a diagram showing a display example of an output file for a program PG1.

FIG. 11 is a diagram showing a display example of an output file for a program PG2.

FIG. 12 is a diagram showing a display example of an output file for each program of an analysis target program.

FIG. 13 is a diagram illustrating a data scope value for data DATAA having a program name “ASCENDANT”.

FIG. 14 is a program name “MINIMUM”.
For explaining the data scope value for the data IMIN of FIG.

FIG. 15 is a diagram showing an example of graphing the data scope values of FIGS. 10 (c) and 11 (c).

FIG. 16 is a diagram showing an example of a graphing process of the frequency of each usage type for each program of FIG.

FIG. 17 is a configuration block diagram of a software quality improvement support device.

FIG. 18 is a flowchart of an evaluation / advice giving process of the software quality improvement support device.

FIG. 19 is a flowchart for giving advice to a manager in step 405.

FIG. 20 is a flowchart of giving advice to a programmer in step 407.

FIG. 21 shows the program “ASCENDAN” shown in FIG.
The figure which shows the display output example of the evaluation result of "T".

FIG. 22 is a diagram showing a display output example of the advice result of the program “ASCENDANT”.

FIG. 23 shows the program “ASCENDAN”
The figure which shows the example of a display output of the advice to quality improvement of "T".

FIG. 24 is a diagram showing a display when the reason for the advice HP1 on the display output screen of FIG. 23 (b) is selected.

FIG. 25 is a diagram showing a display when a point for the advice HP2 on the display output screen of FIG. 23 (b) is selected.

[Explanation of symbols]

 1 analysis target system storage unit 2 analysis result storage unit 3 file management unit 4 input unit 5 output unit 6 input / output control unit 7 temporary storage unit 8 analysis processing unit 9 module call relation storage unit 10 metrics storage unit 11 step number storage unit 12 Number of data storage unit 13 Number of condition judgment storage unit 14 Frequency storage unit by data usage type 15 Data scope value storage unit 20 Evaluation / advice processing unit 21 Evaluation / advice standard

Claims (4)

[Claims] 1. A software quality analysis device for performing software quality analysis, comprising first metric storage means for storing metrics relating to a program amount and metrics relating to a control structure, and storing metrics relating to an operation structure of variable data. Second metric storage means, each of the programs constituting the software, and each of the first and second modules for each module constituting the program.
The analysis processing means for counting the metrics stored in the metrics storage means and performing the quality analysis processing for each program and each module, and the analysis result storage means for storing the analysis processing result by the analysis processing means. A software quality analysis device characterized in that 2. A metric relating to an operation structure of variable data counted by the analysis processing means, comprising module call relation storing means for storing a call connection relation between modules in a program constituting the software as a tree structure, The usage type frequency of the variable data counted by the predetermined usage type classified according to the usage of the variable data, and the predetermined variable data is used in the tree structure stored in the module call relation storage means. 2. The software quality analysis device according to claim 1, wherein the data scope value is given by the sum of the number of existing modules and the minimum number of connection branches for call connection between the modules. 3. The analysis processing means performs graphing processing for each program and each module based on the metrics stored in the first and second storage means. 1. A software quality analysis device according to 1. 4. A metric storage unit that stores a metric related to a program amount, a metric related to a control structure, and a metric related to an operation structure of variable data; Analysis processing means for performing quality analysis processing for each of the constituent modules; analysis result storage means for storing analysis processing results by the analysis processing means; rule storage means for storing rules for evaluation and advice of the program; A quality analysis processing result stored in the result storage means; and an evaluation / advice processing means for evaluating and advising the program based on the rules stored in the rule storage means. Characteristic software high quality support device.