JPH04100161A - Method and device for estimating software reliability - Google Patents

Abstract

PURPOSE:To obtain high compatibility and reliability for wide-range software by graphing and outputting the cumulative curve of an obtained data group and optionally and visually selecting whether a regressive analysis is taken as a single curve or composite curve. CONSTITUTION:The cumulative curve G(X) of obtained bug data is divided into a phase G1(X) which starts at X0 and ends at X1 and a phase G2(X) which starts at X1 and ends at Xn and characteristic regression models F1(X) and F2(X) adapted to the respective phases to decrease an error and increase persuasive power. When the cumulative curve G(X) of the obtained bug data group is a monoton increasing function, the single curve is sufficient for the regression model. Thus, the cumulative curve of the obtained data group is graphed and outputted and whether the regressive analysis is taken as the single curve or composite curve can be selected visually and optionally, so the models can be adapted widely to cumulative curves of obtained data groups in various shapes.

Description

DETAILED DESCRIPTION OF THE INVENTION [Industrial Application Field] The present invention relates to a method and apparatus for estimating reliability in software quality control.

[Prior Art] In recent years, reliability estimation methods using bug growth curves have been widely used in quality control of software, typified by computer programs. This technique is useful because it allows us to scientifically understand the quality of software by performing regression analysis on bug data obtained during the software development and testing process. First, a conventional example will be explained using FIGS. 1 and 4. In Figure 1, (1) is an information processing device, (2) is a computer (with FDD) (
3) is a display, (4) is a printer, (5) is a keyboard, and (6) is a floppy disk.

A floppy disk containing program routines (
6) into the computer (with FDD) (2), run the operation software, and start the information processing device (1). A program routine is loaded from the floppy disk (6) and enters an input waiting state.

FIG. 4 is a flowchart for explaining the conventional example,
Its operation is shown below. Step (50) is a key manual input of the bug acquisition data from the keyboard (5), and is a step in which, for example, the number of bugs acquired every day is input continuously.

Step (53) is a processing stage in which a regression analysis is performed on these data groups and a hest fit coefficient of a regression model (such as a Gompertz curve or a logistic curve) is calculated. Step (56) is the step of displaying the cumulative curve of the data group and the regression analysis results on the same screen, and step (57) is the step of printing out the plotted display contents and the required items obtained by data processing. be. Display (3)
is the hardware in step (56) above, and printer (4) corresponds to the hardware in step (57) above. As mentioned above, in the conventional estimation method and equipment of Software Trust Co., Ltd., regression analysis is performed by limiting one regression model to one curve, and the cumulative curve of the acquired data group is not necessarily superior. There was no indication of suitability.

[Problem to be solved by the invention]

As explained above, although the conventional methods always provide regression analysis results, the drawback is that they lack compatibility with cumulative curves of data groups consisting of multiple curve groups.

[Means for Solving the Problems] The method and apparatus for estimating software reliability according to the present invention first plots and outputs a cumulative curve of an acquired data group, and visually performs regression analysis as a single curve, or It is possible to arbitrarily select whether to perform regression analysis as a composite curve.

[Effect] Being able to select regression analysis as a compound curve means that
1 for the cumulative curve of the acquired data group showing various shapes.
It becomes iJ Noh if it is widely adapted.

〔Example〕

FIG. 1, FIG. 2, and FIG. 3 are diagrams for explaining an embodiment of the software reliability estimation method and device of the present invention. In FIG. 1, (1) is an information processing device; ) is a computer (F I) with D).

(3) is a display, (4) is a printer, (5) is a keyboard, and (6) is a floppy disk.

The hardware configuration of this invention is the same as that of the conventional example. A floppy disk (b) on which programs and routines are stored is inserted into a computer (with FDD) (2), and the operation software is driven to operate the information processing device (1). ). The program routine is loaded and
It will be in a state of waiting for input.

By manually pressing a key on the keyboard (5), the program routine will operate, displaying a diagram of the processing results on the display (3), and outputting a printout of the processing results on the printer (4). Here, I will briefly explain what is called regression analysis. FIG. 3 is an explanatory diagram of the analysis results. As is known, depacking work is carried out in the software development and testing process, but this is a bug removal work to improve the quality of the software. The graph that calculates the number of bugs that occur is called a bug growth curve. Every day, when I look at the cumulative curve G (X) of the bug deso that I have acquired, I start from Xo and once it shows a convergence state around X1, but after 1 point P, a bug appears again and I get X. We often encounter situations where the state of convergence occurs again. In this field, such a shape has (1) a test phase change at point P;

(2) A change was made to the specifications at point P.

I often judge that. That is, performing regression analysis using a single regression model F (X) increases errors and is inconvenient. Therefore, the cumulative curve G (
Phase G + (X) starting from Xo and ending at Xl
If it is divided into phases G2(X) starting from Xl and ending at growing. In addition, if the cumulative curve G (X) of the acquired bug data group is a monotonically increasing function, then 1
It is sufficient if the regression model is also a single curve.

FIG. 2 is a flowchart for explaining the present invention in detail, and its operation will be explained step by step. Step (
50) is the key manual stage of bug acquisition data, step (51) is the initial display stage of the cumulative curve of the data group, and step (52) is the control character J. Should it be zero? Or J. A judgment stage where 1 is set, step (
54) is the control character J. If is 1, the step of inputting the data number of the branch part according to the composite number,
Steps (53a), (53b), and (53c) are processing steps for calculating the best fit coefficients of the regression model;
Step (56) is a step of displaying the cumulative curve of the pig group and the regression analysis results on the same screen, and step (
Step 57) is a step of printing out the plotted display contents and the required items obtained by data processing.

In the description, an example has been described in which the cumulative curve of the data group is divided into two phases, but the present invention is not limited to this and can be applied to a plurality of divisions.

〔Effect of the invention〕

As explained above, according to the method and apparatus for estimating software reliability of the present invention, first, the cumulative curve of the acquired data group is plotted and output, and regression analysis is performed visually as a single curve, or Because it is configured so that regression analysis can be selected as a composite curve, it provides higher compatibility and reliability with a wide range of software, including small-scale device embedded software and large-scale, complex information processing software. can do.

[Brief explanation of the drawing]

FIG. 1 is a system configuration diagram according to an embodiment of the present invention, and FIG.
The figure is a flowchart for explaining the invention in detail,
FIG. 3 is an explanatory diagram for explaining the invention in detail, and FIG.
The figure is a flowchart showing a conventional example. In the figure, (1) is an information processing device, (2) is a computer (
(with FDD), (3) is display, (4) is printer, (5) is key fob, (6) is floppy disk, (50) is key manual stage, (51) is initial display stage, (52) is judgment Stage, (53a), (53b
), (53c) is a data processing stage, (54) is a branch point instruction stage when a compound curve is selected, (56) is a display stage, and (57) is a printing stage. In FIG. 1, the same or corresponding parts are designated by the same reference numerals.

Claims (2)

[Claims] (1) In a software reliability estimation method that performs regression analysis on bugs that occur in the software testing process and uses the analysis results to understand the quality status, You can arbitrarily choose whether to treat the cumulative curve as a single curve or as a compound curve, and if you choose to treat the cumulative curve as a single curve, you can use all of the acquired bug data. If you choose to treat the latter as a compound curve, divide the obtained bug data group into multiple units, perform regression analysis on each unit independently, and then perform a regression analysis on each unit again. A software reliability estimation method characterized by performing polymerization. (2) It is equipped with a display unit, a keyboard, a printer, a memory in which the following program routines are stored, and a central processing unit that executes according to the contents of this memory, and is capable of processing input bug data groups. An apparatus for estimating software reliability, characterized in that the software reliability estimation apparatus is configured to perform regression analysis, visually observe the processing result on the display unit, and obtain it as a printout of the printer. (a) A first program routine for graphically displaying a bug accumulation curve using the input bug data group. (b) Visually check the cumulative bug curve, determine whether to process it as a single curve or a compound curve, and if you select processing as a compound curve, enter the data number of the branching point, etc. Second program routine. (c) A third program routine that performs regression analysis using a predetermined number of bug data (all units for a single curve, multiple units for a compound curve) and determines the best fit coefficient for each regression model. (d) the cumulative curve of the input bug data group;
A fourth program routine for graphically displaying the single or multiple superimposed regression models obtained by regression analysis on the same screen. (E) A fifth program routine that prints out the content graphically displayed on the same screen and the required items obtained by data processing.