JPH09265393A - Program quality evaluation method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide an automatic program quality evaluation system capable of standardizing a program quality evaluation method, substantially shortening operation time required for evaluating the quality of a program and the operation time required for testing the program and efficiently automatically managing the quality of the program by setting quality elements and a quality evaluation procedure to be the decision index of quality evaluation. SOLUTION: The quality elements 15 are defined so as to normally analyze and evaluate the quality of the program, a reference value 16 is set for the quality elements, the quality of the program is evaluated 14 by automatically summing up and judging whether or not it matches with prescribed conditions and the quality element 15 (residual defective internal point) to be a problem in terms of program quality is presented. Thus, by the residual defective internal point, a defective element present inside the program and not extracted yet is specified, and by narrowing points where defects are latent, the efficiency of the test of the program is improved and the quality is improved.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a quality evaluation technique for a program used in a computer, and in particular, defines quality factors to standardize the quality evaluation method of the program.
The present invention relates to a program quality evaluation / improvement technique which is favorable for improving the efficiency of work related to program quality evaluation.

[0002]

2. Description of the Related Art As a conventional technique, a software quality control method (Japanese Patent Laid-Open No. 62-28845) is used.
It predicts the number of remaining defects from the number of defects actually extracted in the test process of program development and the predicted number of defects, and there is no one that analyzes and predicts the contents of defects inherent in the program to evaluate the quality of the program. It was

[0003]

In the prior art, the number of defects actually extracted in the test process of program development and the number of remaining defects are predicted from the predicted number of defects. There is no one that analyzes and predicts and automatically evaluates the quality of programs.

An object of the present invention is to standardize a program quality evaluation method by setting a quality element and a quality evaluation procedure, which are determination indexes for quality evaluation. Also,
By applying the automatic quality evaluation method, it is possible to significantly reduce the work time required for the program quality evaluation and the work time required for the program test.

[0005]

[Means for Solving the Problems] To analyze and evaluate the quality of a program, quality factors such as defect importance, defect code, defect creation process factor, defect processing form, defect factor, defect oversight process, etc. are defined. Then, a standard value is set for the quality element, and it is automatically judged whether or not it meets a predetermined condition, and the quality of the program is evaluated. By doing so, the program quality can be evaluated as standard.

[0006] Further, the residual defect inherent point can identify a defective element that is inherent in the program and has not been extracted yet. By narrowing down potential defective points, the efficiency and quality of the program test can be improved. Can be planned.

[0007]

BEST MODE FOR CARRYING OUT THE INVENTION The definition contents of quality elements are as follows.

[0008] 1. Defect importance It is composed of three types of factors: important, normal, and minor.
The meaning of each is important, file, DB destruction, ABE
This is a problem in system operation such as ND (abnormal termination), program loop, and fraudulent amount of money. Minor items are those that do not interfere with the operation of the system such as typographical errors on screens and forms, omissions, and column shifts. It is usually something else.

[0009] 2. Bad code It consists of four factors: new, degraded, insufficiently corrected, and latent. The meaning of each is a new defect that has newly occurred for the program creation. Degradation is a defect that occurs when a program is modified and mistakes are made to the unnecessary areas. Insufficient correction is a defect that occurred because the correction was insufficient when the program was modified. Latent is a latent defect inherent in the remodeling mother when the program was created by remodeling.

3. Factors of defective manufacturing process BS (system planning), FS (system design), DS
It is composed of four factors: (program function design) and C (program description). The meaning of each is BS
Indicates that there is a bad factor in the system plan. The FS has a defect factor in the system design. The DS has a defect factor in the program function design. C has a defect factor in the program description.

4. Defective processing form It consists of three types of factors: normal processing, abnormal processing, and boundary / limit processing. The normal processing is a defect in the normal processing portion of the program. The abnormal processing is the existence of a defect in the abnormal processing part of the program logic. The boundary / limit processing is a defect in the check processing portion necessary for the system specifications.

5. Defect factors Lack of understanding of specifications, lack of understanding of standards / standards, lack of consideration of operation,
It consists of five factors: interface and simple mistake. Insufficient understanding of specifications is a defect caused by misunderstanding of business specifications, system specifications, and program specifications. Lack of understanding of standards / standards is a defect caused by a lack of understanding of standards / standards set in the project and misunderstanding. The lack of operational consideration is a defect that occurred because the system operation was not considered. The interface is a defect caused by the interface mismatch between programs. A simple mistake is a defect caused by a simple mistake.

6. Failure overlook process UD (unit test process), CD (interlocking test process), S
It consists of three factors, D (Comprehensive test process). Each meaning means that UD is a defect that should be extracted in the unit test process. A CD is a defect that should be extracted in the interlocking test process. SD is a defect that should be extracted in the comprehensive test process.

Embodiments of the present invention will be described in detail below with reference to the drawings. FIG. 1 is a system configuration diagram for realizing an embodiment of the present invention, FIG. 2 is a record layout of data stored in a data file 18, and FIG. 3 is a record layout of data stored in a summary file 17. 4 is an example of the reference value table 15,
FIG. 5 is a comprehensive quality evaluation list output to the printer 13, and FIG. 6 is a flow chart of processing showing details of the processing unit in FIG.

According to the screen displayed on the display device 12, the data input person has a work name 21, a program name 22, a defect importance level 23, a defect code 24, and a defect making process factor 2.
5, the defect processing form 26, the defect factor 27, and the quality factor of the defect oversight process 28 are input 601 from the input device 11. The end determination 602 of the data to be input is performed, and the input data is stored 603 in the data file 18 until the data input is completed.

After the data input is completed, the data stored in the data file 18 is read 604, the end judgment 605 of the data file is performed, and if the data exists, the factor of each quality factor is counted up 606 for each job,
The data is stored in the tabulation file 607.

After the totalization of the factors of the quality factors for each work is completed, the totalization file 17 is read 608 and the end determination 609 of the totalization file is performed, and if data exists, the contents of the totalization file are used for each business and each quality factor. Then, the aggregated value of each factor is checked against the judgment standard 41 in the standard value table A16, and a check 610 is performed as to whether or not the judgment standard 41 is met. A flag (weight 42 for each quality element) indicating the remaining defective inherent point is set 611, and when the data of the totalization file 17 is completed, the process is terminated 609.

Further, the flags of the quality elements indicating the remaining defective inherent points are aggregated and compared with the quality points 43 in the reference value table B16 to perform the quality evaluation determination 612 (good / common / bad determination processing). As a result, it is possible to evaluate the quality of the program for each job without the judgment of an expert. Finally, the comprehensive quality evaluation list (FIG. 5) is displayed on the display device 12 and the printer 13.

The reference value tables A and B16 shown in FIG.
Judgment criteria 41, weights 42, judgment expressions registered in quality points 43, and values are determined by the combination of factors, the number of judgment expressions, the average value / variable specific value, etc. depending on the project difficulty of the program to be evaluated. , The structure of the expression, the value shall be changed.

FIG. 7 shows the transition of the residual defect status when the automatic program quality evaluation / improvement tool is applied.

In the first test, the remaining defective intrinsic points (defective importance-important and defective code-degrade,
(Insufficient repair) is focused on and defects are identified. Similarly, in the second test, the residual defect inherent points (defective manufacturing process factor-FS, DS, and defective processing form-abnormal processing, boundary / limit processing) are intensively tested and defects are extracted. In addition, in the third test, the remaining defect inherent points (defect factor-interface and defect oversight process-UD, CD) are intensively tested, and all programs included in the job name C remain. The number of defects is 0.

On the other hand, FIG. 8 shows the residual defect status for each residual defect intrinsic point when the program automatic quality evaluation / improvement tool is not applied. When the program automatic quality evaluation / improvement tool is applied, FIG. In contrast to this, since it is not possible to perform tests intensively according to the remaining defect inherent points, it is not possible to extract defects efficiently, and the number of tests for extracting defects of all programs is 5 times. From this, it can be seen that an efficient test can be performed when the program automatic quality evaluation / improvement tool is applied.

FIG. 9 is a continuous curve (Gompertz curve) of defect convergence generally in software product development.
A solid line indicates a defect extraction cumulative value curve when the program automatic quality evaluation / improvement tool is applied, and a broken line indicates a defect extraction cumulative value curve to which the program automatic quality evaluation / improvement tool is not applied. Compared to the case where the program automatic quality evaluation / improvement tool is not applied, when the program automatic quality evaluation / improvement tool is applied, the defects can be effectively isolated by conducting the test focusing on the residual defect inherent points. Therefore, a difference appears in the slope of the curve from the time of analysis,
This is the effect of the automatic program quality evaluation / improvement tool.

[0024]

According to the present invention, the quality of a program can be evaluated as a standard by setting quality factors for evaluating the quality of the program, and the quality of the program can be efficiently managed. Further, by conducting the program test with emphasis on the narrowed remaining defect inherent points, the program defects can be efficiently extracted.

[Brief description of drawings]

FIG. 1 shows a system configuration diagram of the present invention.

FIG. 2 shows a file layout of a data file.

FIG. 3 shows a file layout of a summary file.

FIG. 4 shows an example of reference value tables A and B.

FIG. 5 shows a file layout of a comprehensive quality evaluation list.

FIG. 6 shows a flow chart of processing for outputting a comprehensive quality evaluation list.

FIG. 7 shows the status of residual defects for each of the residual defect intrinsic points when the program automatic quality evaluation / improvement tool is applied.

FIG. 8 shows the status of residual defects for each of the inherent points of residual defects when the automatic program quality evaluation / improvement tool is not applied.

FIG. 9 shows a defect convergence curve (Gombertz curve) in general software product development.

[Explanation of symbols]

11 ... Input device, 12 ... Display device, 1
3 ... Printer, 14 ... Evaluation determination unit, 15 ... Quality element table, 16 ... Reference value tables A and B,
17 ... Totalization file, 18 ... Data file, 21 ... Business name, 22 ... Program name,
23 ... Defect importance degree, 24 ... Defect code, 25 ... Defect making process factor, 26 ... Defect processing form, 27 ... Defect factor, 28 ... Defect overlooking process, 41 ... Judgment standard, 42 ... Weight, 4
3 ... Quality points, 601 ... Data input processing,
602 ... Data input end determination processing, 603 ... Data file storage processing, 604 ... Data file reading processing, 605 ... Data file reading end determination processing, 6
06 ... Count-up processing of factors of each quality factor,
607 ... Aggregate file storage processing, 608 ... Aggregate file reading processing, 609 ... Aggregate file reading end determination processing, 610 ... Predetermined condition matching check processing, 611
Residual defect inherent point set processing, 612 ... Quality evaluation determination processing, 613 ... Overall quality evaluation list.

 ─────────────────────────────────────────────────── ─── Continuation of front page (72) Inventor Yukitoshi Chiba 1-6-27 Shinsuna, Koto-ku, Tokyo Hitachi Public Information Division (72) Inventor Takahisa Yagi 1-6-Shinsuna, Koto-ku, Tokyo No. 27 Hitachi, Ltd. Public Information Business Department (72) Inventor Hitoshi Tashiro 1-6-6, Shinsuna, Koto-ku, Tokyo Hitachi Ltd. Public Information Business Department (72) Inventor Shizuko Matsuo 1-chome Shinsago, Koto-ku, Tokyo 6-27 Hitachi Ltd. Public Information Division (72) Inventor Toshihiro Muto 1-627 Shinsuna, Koto-ku, Tokyo Hitachi Public System Engineering Co., Ltd.

Claims (1)

[Claims] 1. A quality evaluation method by defining a quality factor for a defect extracted in each test process (single unit, interlocking, and total) of program development, and performing a program quality evaluation from this quality factor. Standardize. Further, the program quality evaluation method is characterized in that a quality factor and a reference value are compared with each other, and a condition that meets a predetermined condition is automatically extracted as a residual defect inherent point.