JP6919980B2 - Quality control method and quality control equipment - Google Patents

Description

The present invention relates to quality control of system development.

In software development projects, development is shared among each person in charge. The program developed by each person in charge is combined and tested, and if there is no defect, the program is run on the actual machine to perform the test, and if there is no defect, it is provided to the user. If there is a defect in the test, identify the defective part and correct it.

Some defects can only be seen by running the program, but some can only be seen by carefully reading the program itself. Therefore, each person in charge performs a self-check for the task he / she is in charge of by using the self-checklist to reduce defects.

The purpose of reviewing a program is mainly to find and remove errors that cannot be detected by the syntax check or grammar check performed by the compiler, and the main purpose is to manually visually check the source program. The invention disclosed in Patent Document 1 proposes a review support system that reduces manual labor by automatically detecting statements and identifiers that may cause an error that cannot be detected by a compiler.

In the invention disclosed in Patent Document 1, if a statement or identifier that matches the check item is defined in advance, the corresponding statement or identifier is searched from the source program, and the corresponding source program content is searched together with the surrounding contents. Since it is presented to the user, the user is less likely to have to look at the entire source program.

Japanese Unexamined Patent Publication No. 2005-190330

However, the skills of many developers in software development projects generally vary, and if the person has a low level of understanding of self-checking or a low quality mindset, the presence or absence of conventional checked items is present. In the checklist to fill in (check mark), it is difficult to detect the error (bug) caused by the false description when it is included.

In particular, in large-scale development and short-term software development, a large number of development personnel are gathered at once, so there are variations in the quality mindset and skills of each personnel. Development personnel with low quality mindset and skills cannot carry out the process properly, and the quality of the developed software is poor. In order to reduce the risk of delay in delivery of software development and the risk of increased cost, it is important to make the software quality after self-check of each person in charge as much as possible before the actual machine test stage of the program.

The present invention has been made in consideration of such a problem, and an object of the present invention is to control software quality by using self-check results.

In order to solve the above problems, the quality control method of the present invention is a method of controlling the quality of software by using a computer, and includes a step of acquiring a self-check result including the number of confirmation points related to the software and the self-check. The self-check result includes an analysis step of calculating the quality of the software based on the result, and the analysis step includes the software scale and the number of confirmed points. And configure to calculate software quality.
Further, as another feature of the present invention, the quality control method includes a step of acquiring a quality check result including the number of defects of the software, and in the analysis step, the scale of the software and the number of confirmed points are included. , The score is calculated based on the quality check result, and the ranking based on the score is output.

According to the present invention, software quality can be calculated using the self-check result.

This is an example of a configuration diagram of a quality control device. This is an example of a self-checklist. It is a figure which shows the data structure of the confirmation place number table. It is a figure which shows the data structure of the defective number table. It is a figure which shows the data structure of the score information table. It is a figure which shows the data structure of the viewpoint importance table. It is a figure which shows the data structure of the importance score coefficient table. This is an example of the menu screen of the tool interface. It is a flowchart of the number registration flow. It is a flowchart of the score calculation flow. It is a flowchart of point addition analysis processing. It is a flowchart of a deduction analysis process. It is a flowchart of viewpoint importance, importance score coefficient setting processing. This is an example of a screen in which the point addition analysis result is displayed on the quality control unit terminal. This is an example of a screen in which the deduction analysis result is displayed on the quality control department terminal.

In a preferred example of the invention, the number of confirmation points is given to each person in charge of software development according to the self-checklist 200 illustrated in FIG. 2, for each check item 201, for each statement of the target program corresponding to the check content, or for each identifier. It is counted as 202 and reported at the end of the self-check. This is because in the conventional self-check, when the target program is checked for each check item 201, it is marked as checked and reported. For example, if there are 12 places corresponding to a certain check item. , Even if the person in charge checks all the relevant parts, or even if only 10 of them are checked and the remaining 2 parts are missed, the problem that only the same checked report was found is improved. ..

That is, in a preferable example of the invention, after the self-check by each person in charge, each person in charge is made to report the number of confirmation points 202 for each check item 201 (hereinafter referred to as self-check viewpoint) in the self-checklist 200. We propose to use it as an index for estimating the software quality of the software.

In the self-check viewpoint 201 adopted in the preferred example of the invention, the number of confirmation points is not one or a fixed number regardless of the scale (number of steps, etc.) of the target program, but is confirmed substantially linearly as the scale of the target program increases. The number of locations tends to increase. A self-checklist that summarizes a plurality of self-check viewpoints having such a tendency is created, provided to the software development department, and the number of confirmation points is counted and reported for each self-check viewpoint at the time of self-check.

Hereinafter, examples will be described with reference to the drawings.

FIG. 1 is an example of a configuration diagram of a quality control device 100 that outputs an index for estimating the quality of the program by inputting the number of self-check confirmation points of the program reported by the software development department.

The quality control device 100 can be configured on a computer (server), and its hardware configuration includes a calculation unit 110 composed of a CPU (Central Processing Unit) and the like, an HDD (Hard Disk Drive), a flash memory, and the like. It is provided with a storage unit 120 composed of an SSD (Solid State Drive) or the like using the above, an output unit 130 composed of various output devices and the like, and a communication unit 140 composed of a NIC (Network Interface Card) or the like.
The communication unit 140 is connected to a plurality of quality control unit terminals 160 and a plurality of development unit terminals 170 via a network 150.

The calculation unit 110 realizes the following functional units by loading the quality control program (not shown) stored in the storage unit 120 and executing it on the CPU.
The number of confirmation points registration unit 111 receives the number of confirmation points for each self-check viewpoint for each program input from the quality control unit terminal 160 by the quality control unit member, and records the number of confirmation points in the storage unit 120 in the storage area 121. Alternatively, the number of confirmation points reported by the person in charge of the software development department from the terminal 170 of the development department after the self-check may be received and recorded in the number of confirmation points storage area 121.

The number of defects The registration unit 112 is the number of cases found to be defective by the quality control staff based on the result of the inspection (test) of the program whose quality is suspected from the result of the point addition analysis (described later) or the result of the acceptance inspection of the outsourced program. Is received and recorded in the defective number storage area 122 of the storage unit 120.

The score calculation unit 113 reads out the data recorded in the number of confirmed points storage area 121 and the number of defective cases storage area 122, multiplies the score coefficient (described later) determined from the importance of each self-check viewpoint, and adds points (described later). The score (described later), the deduction (described later), and the total score (described later) are calculated and recorded in the score information storage area 123 of the storage unit 120.

The point addition analysis unit 114 reads out the additional points recorded in the score information storage area 123, calculates the additional points for the scale division, creates the additional points analysis result ranked according to the magnitude of the scale division additional points, and creates the quality control unit. It is displayed on the terminal 160 and recorded in the point addition analysis result storage area 126 of the storage unit 120. In addition, the point addition analysis result is output to the form from the output unit 130.

The deduction analysis unit 115 reads out the total score recorded in the score information storage area 123, totals the total score for each person in charge, calculates the total score of the scale division, and is in charge of the size of the total score of the scale division. The deduction analysis result in which the persons are ranked is created, displayed on the quality control unit terminal 160, and recorded in the deduction analysis result storage area 127 of the storage unit 120. In addition, the output unit 130 outputs the deduction analysis result to the form.

The viewpoint importance and importance score coefficient setting unit 116 receives a setting from the quality control unit terminal 160 in advance for inputting the importance for each self-check viewpoint and the score coefficient for each importance from the quality control unit terminal 160, and the storage unit 116. 120 viewpoints Record in the importance storage area 124 and the importance score coefficient storage area 125.

The storage unit 120 has a storage area for storing each information of the number of confirmation points 121, the number of defects 122, the score information 123, the viewpoint importance 124, the importance score coefficient 125, the point addition analysis result 126, and the point deduction analysis result 127.

FIG. 3 shows the data structure of the number of confirmation points table 121. The control number 121a, which is the data item column of the number of confirmation points table 121, is the control number of the data record assigned to each program to be recorded, the person in charge 121b is the name of the person in charge of developing the program, and the person in charge 121c is in charge. The leader of the person, the development frequency 121d is the development unit representing the development lot in the project, the program ID 121e is the identification name unique to the program, the development scale 121f is the number of steps of the development program, etc., and the number of confirmation points in charge. 1 (121g) is the number of confirmation points checked by the person in charge regarding the self-check viewpoint 1, and the number of confirmation points 1 (121h) is the number of confirmation points 1 from the number of confirmation points checked by the person in charge regarding the self-check viewpoint 1. The number of confirmation points n (121k) in charge is the number of confirmation points checked by the person in charge regarding the self-check viewpoint n, and the number of confirmation points n (121l) in charge is the number of confirmation points n (121l) in charge by the person in charge regarding the self-check viewpoint n. This is the difference obtained by subtracting the number of confirmed points n in charge from the number of confirmed points checked.

FIG. 4 shows the data structure of the defective number table 122. The defect number table 122 corresponds to the self-check viewpoint as a result of the control number 122a, the person in charge name 122b, the person in charge name 122c, the development frequency 122d, the program ID 122e, the development scale 122f, and the program inspection (test) by the quality control department. Viewpoint to store the number of newly discovered defects It has a data item column such as 122h.

FIG. 5 shows the data structure of the score information table 123. The score information table 123 reads data from the control number 123a, the person in charge name 123b, the person in charge name 123c, the development frequency 123d, the program ID 123e, the development scale 123f, and the number of confirmation points table 121, and the person in charge (and the person in charge) for the program. 123g of additional points to store the evaluation points of the self-check result, 123h of deductions to store the negative evaluation points due to the defects found as a result of the program inspection (test) by the quality control department, and the total points calculated from the additional points and deductions. It has a data item column such as 123i.

FIG. 6 shows the data structure of the viewpoint importance table 124. The viewpoint importance table 124 has a data item column of a viewpoint number 124a representing an identification number of a self-check viewpoint and a data item column of a importance 124b representing a rank of importance assigned to the self-check regarding the corresponding self-check viewpoint.

FIG. 7 shows the data structure of the importance score coefficient table 125. The importance score coefficient table 125 includes importance 125a representing the rank of importance assigned to each self-check viewpoint in the viewpoint importance table 124, and weighting when calculating points to be added or deducted according to the rank of each importance. It has a data item column of a score 125b for storing a score coefficient representing.

FIG. 8 is an example of the menu screen 300 of the tool interface displayed on the quality control unit terminal 160. Select the tool menu (301-305) to initiate the process performed by the quality control personnel.

FIG. 9 is a flowchart of the number registration flow. The quality control member selects "number information registration" 301 from the tool interface screen 300 (S10) to start the number registration process.

In step S11, the quality control staff registers the self-check result of the program reported by the software development department (or, when the outsourced software is delivered, the subcontractor reports the self-check result of the program. In the case of receiving and registering), the confirmation location registration unit 111 is activated and accepts the input of the module information (program ID, scale) and the self-check result.
When the quality control member registers the result of the program inspection (test), the defective number registration unit 112 is activated and accepts the input of the module information (program ID, scale) and the defective number.

In step S12, the confirmation location number registration unit 111 registers the module information and the self-check result in the confirmation location number table 121.
In step S13, the defective number registration unit 112 registers the module information and the defective number in the defective number table 122.
In step S14, FIG. 13 for notifying that the data has been registered on the quality control unit terminal 160 shows the viewpoint importance, the importance score coefficient setting unit 116 of the viewpoint importance by the quality control unit, and the importance score coefficient. It is a flowchart of the process of accepting a setting and registering it in a viewpoint importance table 124 and an importance score coefficient table 125. The quality control member selects "viewpoint importance / importance score coefficient setting" 305 (S50) from the tool interface screen 300 to start the score calculation process.

In step S51, the quality control member sets a score coefficient 125b for each importance 125a. Prioritize the self-check viewpoints that are especially important and increase the score coefficient. The importance X sets the score coefficient to be multiplied by the number of defects corresponding to the viewpoint, and the importance Y sets the score coefficient to be multiplied by the number of defects outside the viewpoint.

In step S52, the quality control member sets the importance level 124b for each self-check viewpoint 124a. While the development time (self-check time) is limited, a weight difference is provided in the score coefficient so that the most important ones can be checked. Increase the score for items that are difficult to confirm by self-check and take time and effort.

In step S53, the data input from the quality control unit terminal 160 is received and registered in the viewpoint importance table 124 and the importance score coefficient table 125.

FIG. 10 is a flowchart of a process in which the score calculation unit 113 reads out the data recorded in the confirmation point number table 121 and the defective number table 122, calculates points to be added, deducted, and total points, and registers them in the score information table 123. be. The quality control member selects "score calculation execution" 302 from the tool interface screen 300 (S20) to start the score calculation process.

In step S21, the number of confirmed points in charge and the number of confirmed points of the person in charge for each self-check viewpoint are extracted from the number of confirmed points table 121.
In step S22, the importance and the score coefficient for each self-check viewpoint are extracted from the viewpoint importance table 124 and the importance score coefficient table 125. The total value) is multiplied by the score coefficient for each viewpoint and added up to calculate the additional points (Ap) by the following equation (Equation 1).

Here, Ap: points added, i: self-check viewpoint number (= 1 to n), Npi: number of confirmation points checked by the person in charge regarding self-check viewpoint i, Nri: confirmation points checked by the person in charge regarding self-check viewpoint i. Number (this value stores the difference value that the person in charge later corrected the number of confirmation points checked by the person in charge), Sfi: Score coefficient assigned to the self-check viewpoint i.
In step S24, the calculated additional points (Ap) are registered in the score information table 123.
In step S25, the number of defects corresponding to the viewpoint and the number of defects outside the viewpoint are extracted from the defect number table 122.
In step S26, the score coefficient of importance X is extracted from the importance score coefficient table 125 for the number of defects corresponding to the viewpoint, and the score coefficient of importance Y is extracted for the number of defects outside the viewpoint.

In step S27, the number of defects and the score coefficient are multiplied and added up to calculate the deduction (Dp) by the following equation (Equation 2).
(Number 2) Dp ＝ Ndv ・ Sfx ＋ Ndo ・ Sfy
Here, Dp: deduction, Ndv: number of defects corresponding to the viewpoint, Ndo: number of defects outside the viewpoint, Sfx: score coefficient of importance X, Sfy: score coefficient of importance Y.
In step S28, the calculated deduction (Dp) is registered in the score information table 123.

In step S29, the points added and deducted are added up and registered in the score information table 123 as a total score.
This process can be executed even if the number of defects table is not registered. Since the number of defects is 0, the "point deduction" becomes 0, and practically only the point addition calculation is performed.

In addition, if the quality control staff registers the number of defects as a result of inspection (test) of a program whose quality is suspected from the result of point addition analysis (described later) and executes this process, the point addition calculation has already been completed. Since it is registered in the score information table 123, only the deduction calculation and the total score calculation are executed in that case.

FIG. 11 shows a flowchart of the point addition analysis process executed by the point addition analysis unit 114. The quality control member selects "point addition analysis execution" 303 from the tool interface screen 300 (S30) to start the point addition analysis process.

In step S31, the point addition analysis unit 114 extracts module information (program ID, scale) and points added from the score information table 123.

In step S32, in order to evaluate the points added for each program extracted in S31 equally, the scale addition points obtained by dividing the points added by the scale of the corresponding program and the median value of the scale addition points of all the extracted programs are calculated. ..

In step S33, the point addition analysis result as shown in FIG. 14 is displayed on the quality control unit terminal 160. For example, the additional point data calculated from the self-check results of all the programs related to one project is divided by the scale of the program, and the calculated additional point 401 is added, and the data for each program is arranged.

Assuming that the person in charge of self-checking each program and the person in charge have faithfully fulfilled their responsibilities, the number of confirmation points according to the scale of each program is reported, and points are added calculated from the number of confirmation points. Is also a value according to the scale of each program. In order to evaluate those points equally, if the points are expressed by the scale addition points divided by the scale of the corresponding program, the scale addition points of each program show values that are close to each other within a range of some errors and variations. I know from my experience. If the scale addition point value is extremely small compared to the scale addition point value of other programs, the person in charge has a low level of understanding of the self-check, and the self-check is performed properly. It is possible that the number of confirmed points is small and reported.
Also, if the scale addition point value shows an extremely large value compared to the scale addition point value of other programs, the person in charge misunderstands the self-check and increases the number of incorrect confirmation points. It is possible that the report was made or the number of confirmed points was falsely reported. In either case, the quality of the program is expected to be low. Therefore, in the point addition analysis result, the median value 402 of the scale addition points of all programs is calculated and displayed. A program with a scale addition point that deviates from the median scale addition point above or below a predetermined value is predicted to have low quality and is subject to inspection (test) by the Quality Control Department. Further, the quality prediction may be calculated and displayed by the quality control device 100 from the scale addition points 401 and the median value 402 thereof.

In this example, the analysis is performed for each program, but point-added analysis may be performed for each person in charge or responsible person (group of persons in charge). The appropriate number of confirmation points may depend on the program, and the scale addition point 401 will vary to some extent. On the other hand, it is considered that the scale addition points are close to each other among the programs included in one development project, and the quality is considered to depend on the person in charge and the group. Therefore, it is considered that the quality of the person in charge and each group is easily reflected in the scale addition points.

Returning to step S33 of FIG. 11, the point addition analysis unit 114 records the point addition analysis result in the point addition analysis result storage area 126 of the storage unit 120.
In step S34, the point addition analysis unit 114 outputs the point addition analysis result to the form from the output unit 130.

FIG. 12 shows a flowchart of the deduction analysis process executed by the deduction analysis unit 115. The quality control member selects "execute deduction analysis" 304 from the tool interface screen 300 (S40) to start the deduction analysis process.

In step S41, the deduction analysis unit 115 extracts the module information (program ID, scale) and the total score from the score information table 123.

In step S42, in order to evaluate the total score of each person in charge extracted in S41 equally, the total score is divided by the scale of the corresponding program, and the total score of each person in charge is divided by the scale of the corresponding program. Calculate the value 404.

In step S43, the deduction analysis result as shown in FIG. 15 is displayed on the quality control unit terminal 160. For example, the points added based on the results of self-checks by all developers involved in one project and the points deducted calculated from the number of defects found from the results of inspections (tests) conducted by the quality control department were totaled. The total score is totaled for each person in charge, and the total score for each person in charge is divided by the scale of the corresponding program to calculate the total score for scale division 403. It is conceivable to do. Instead of each person in charge, it may be aggregated by person in charge (by group of person in charge).

The person in charge with a large total score of 403 can evaluate that the quality of the developed program is high. This is because in the point addition analysis, the person in charge with high quality has a large number of self-check confirmation points (appropriate) and few defects. This is because if a false declaration is made for the number of confirmed points in the self-check, the quality is suspected in the point addition analysis, the quality control department inspects it, and the defect is found, and the points are deducted. In addition, even if there are few defects in the program, if the number of self-checked points is so large that the intention of fraud is felt, a penalty such as deduction may be imposed.

In step S43 of FIG. 12, the deduction analysis unit 115 records the deduction analysis result in the deduction analysis result storage area 127 of the storage unit 120.
In step S44, the deduction analysis unit 115 outputs the deduction analysis result to the form from the output unit 130.

In system development, it is premised in the present invention that the number of confirmation points (defined by the person in charge and the person in charge in the checklist) generated per program scale does not fluctuate significantly. Therefore, the number of confirmation points per program scale is evaluated, and if the value deviates significantly, it is considered that the person in charge has a low level of understanding and self-check is inappropriate, and therefore the program quality is also low. is expected.

Depending on the program, the quality is high, but due to the nature of the program, it is considered that the value of the scale addition point may deviate from the median value 402 of the scale addition point in the point addition analysis result, but the quality of the program is suspected. When a quality manager actually performs a defect check in an inspection (test), it serves as a guide for determining a program to be checked for defects. High quality programs have fewer defects and fewer deductions.

In FIG. 15, in order to increase the total scale score, it is effective to 1) perform an appropriate self-check, enter an appropriate number of confirmation points, and 2) reduce defects. By describing in a ranking format and disclosing this to each person in charge of development, the awareness of each person in charge will be raised. The person in charge improves the quality by understanding the development work properly, reporting the self-check properly, and creating a program with few defects.

According to this embodiment, when accepting an inspection of a program outsourced to another company, the self-check result reported by the other company is used to determine which sample program should be accepted and verified before accepting the quality of the received product. You can give a guide.

According to this embodiment, the person in charge of system development can visualize not only the quality of the program but also the quality of the self-check itself, so that it is considered that the quality mindset of the program and the self-check is improved. As a result, it is possible to reduce defects before the combined test of each program and prevent delay in progress after the start of the test. In addition, since the progress including quality can be managed at the time of program creation of each person in charge, delay risk can be extracted and countermeasures can be taken. In addition, by visualizing the quality, the quality awareness of the person in charge can be improved.

In this embodiment, the method of visualizing and managing the quality of software and the quality of self-checking of software and its device are described, but general products including not only software but also hardware are manufactured. In doing so, it is considered that it can be applied to the method of managing the quality of the work procedure, whether or not it was done according to the work procedure.
In that case, the effectiveness of this embodiment can be achieved by adopting a self-checking viewpoint that checks universal common functions as much as possible, rather than checking a specific function of the product. Considered a condition.

In the first embodiment, in the point addition calculation, the calculation is performed using both the number of confirmation points of the person in charge and the number of confirmation points of the person in charge. That is, the quality of the software is evaluated at the end of the review of the responsible person (the person in charge and the responsible person are evaluated. However, in many cases, the responsible person does not perform the review, and the responsible person If there is no number of confirmation points, only the person in charge will be evaluated). In the second embodiment, the point addition calculation is performed only by the number of confirmation points of the person in charge. This makes it possible to evaluate the person in charge. A competent person can add points to a reasonable value, and if he / she does not understand well or cheats, the points added will decrease or increase.

As a proposed change of the second embodiment, the point addition calculation may be performed using only the number of confirmation points of the person in charge. In this case, the number of confirmation points of the person in charge is the difference from the number of confirmation points of the person in charge, and the larger the absolute value, the larger the value corrected by the person in charge. The absolute value of the number of points confirmed by the person in charge is used for the point addition calculation. The larger the points added, the greater the role of the person in charge. However, it is necessary to consider that if the person in charge is competent, the correction of the person in charge will be reduced. For example, it is necessary to evaluate in consideration of the ability of the person in charge.

100 Quality control device 110 Calculation unit 111 Number of confirmation points Registration unit 112 Number of defects Registration unit 113 Score calculation unit 114 Point addition analysis unit 115 Point deduction analysis unit 116 Viewpoint importance, importance Score coefficient setting unit 120 Storage unit 121 Number of confirmation points Storage area (Table of number of confirmation points)
122 Number of defects storage area (table of number of defects)
123 Score information storage area (score information table)
124 Perspective importance storage area (perspective importance table)
125 Importance score coefficient storage area (importance score coefficient table)
126 Point addition analysis result storage area 127 Point deduction analysis result storage area 130 Output unit 140 Communication unit 150 Network 160 Quality control department terminal 170 Development department terminal 200 Self-checklist 201 Check items (self-check viewpoint)
202 Person in charge confirmation number 300 Tool interface menu screen 301 Tool menu "Number information registration"
302 Tool menu "Score calculation execution"
303 Tool menu "Additional point analysis execution"
304 Tool menu "Perform deduction analysis"
305 Tool menu "Viewpoint importance / importance score coefficient setting"
401 Scale Discount Points 402 Median Scale Discount Points 403 Total Scale Discount Scores 404 Median Scale Discount Total Scores

Claims (4)

It is a method of controlling the quality of software using a computer.
When creating software, each person in charge of software development counts and reports as the number of confirmation points for each self-check viewpoint, for each statement or identifier of the target program corresponding to the check content, according to the self-check list. The process of acquiring and registering the self-check results
Based on the self-check result, the process of multiplying the number of confirmation points for each self-check viewpoint by the score coefficient indicating the importance of the self-check viewpoint and adding them up to calculate the additional points for each software.
The process of calculating the scale addition points by dividing the points added for each software by the scale of the software,
A point addition analysis process that displays a list of at least the person in charge, points added, and scale addition points for each target software, and calculates and displays the median scale addition points.
In the list of point addition analysis results displayed in the point addition analysis step, software having a scale addition point that deviates from the median scale addition point above and below a predetermined value is predicted to be of low quality and is subject to inspection (test). After the inspection (test) is completed, the process of getting the inspection result and the process of sending it to the quality control department
Among the above inspection results, the process of registering the number of defects of each software as the number of defects corresponding to the self-check viewpoint as the number of defects applicable to the viewpoint and the number of defects not corresponding to the self-check viewpoint as the number of defects outside the viewpoint.
The process of calculating and registering the deduction points by multiplying the number of defects corresponding to the viewpoint of each software and the number of defects outside the viewpoint by the corresponding score coefficients, and the process of registering.
The process of adding up the points added and deducted from each registered software and registering them as the total score,
The process of calculating the total score for each person in charge by dividing the total score for each person in charge of each software by the scale of the software, and
A deduction analysis process that displays a list of at least the total score for each person in charge and the total score of the scale discount, and calculates and displays the median total score of the scale discount.
A quality control method characterized by having. In claim 1,
In addition to each person in charge of software development counting and reporting the number of confirmation points for each self-check viewpoint, the person in charge follows the same self-check list and the number of confirmation points for each self-check viewpoint. It also has a process of counting, acquiring the reported self-check result, and registering the difference obtained by subtracting the number of confirmation points of the person in charge from the number of confirmation points checked by the person in charge as the number of confirmation points of the person in charge.
In the process of calculating the points added for each software, the total of the number of points confirmed by the person in charge and the number of points confirmed by the person in charge for each self-check viewpoint is multiplied by the score coefficient and added up to calculate the points added for each software. A quality control method characterized by. A device that uses a computer to control the quality of software.
When creating software, each person in charge of software development counts and reports as the number of confirmation points for each self-check viewpoint, for each statement or identifier of the target program corresponding to the check content, according to the self-check list. The number of confirmation points to be registered by acquiring the self-check results
Based on the self-check result, the score calculation unit that calculates and registers the points to be added for each software by multiplying the number of confirmation points for each self-check viewpoint by the score coefficient indicating the importance of the self-check viewpoint and adding them up. When,
The scale addition points are calculated by dividing the points added for each software by the scale of the software, a list of at least the person in charge, points added, and the scale addition points for each target software is displayed, and the median scale addition points are calculated. And the point addition analysis unit to display
In the list of point addition analysis results displayed by the point addition analysis unit, software with a scale addition point that deviates from the median scale addition point above and below a predetermined value is predicted to be of low quality and is subject to inspection (test). After the inspection (test) is completed, the inspection result is acquired, and among the inspection results, the number of defects of each software is checked, and the number of defects corresponding to the self-check viewpoint is the viewpoint of the number of corresponding defects and self-check. The number of defects registration department that registers the number of defects that do not correspond to the viewpoint as the number of defects outside the viewpoint,
The score calculation unit further multiplies the number of defects corresponding to the viewpoint of each software and the number of defects outside the viewpoint by the corresponding score coefficients, calculates the deduction, calculates the deduction, and registers and registers it. The points added and deducted for each software are added up and registered as the total score.
The total score for each person in charge of each software is divided by the scale of the software to calculate the total score for each person in charge, and at least the total score for each person in charge and the list of total score for scale division are displayed. A deduction analysis unit that calculates and displays the median total score of the scale discount,
A quality control device characterized by being equipped with. In claim 3 ,
In the confirmation point registration unit, each person in charge of software development counts and reports the number of confirmation points for each self-check viewpoint, and the person in charge follows the same self-check list. The number of confirmation points is counted for each self-check viewpoint, the reported self-check result is acquired, and the difference obtained by subtracting the number of confirmation points of the person in charge from the number of confirmation points checked by the person in charge is further registered as the number of confirmation points of the person in charge. death,
The quality calculation unit is characterized in that the total of the number of points confirmed by the person in charge and the number of points confirmed by the person in charge for each self-check viewpoint is multiplied by a score coefficient and added up to calculate the added points for each software. Management device.

Images