KR20030056232A - Software Process Appraising Method - Google Patents

Abstract

PURPOSE: A method for evaluating a software process is provided to improve the software process by suggesting the software process necessary for a level after evaluating a software development process through an objective evaluation and understanding a present software development level. CONSTITUTION: All processes for developing a software are evaluated(S11). Respective performance data of a software project evaluating the process is collected(S12). The process to be preceded at an infra side is extracted by using an evaluation result(S13). The probable improvement of the software performance by improving the process is analyzed by performing a regression analysis(S14). The process needing the improvement is defined(S15). The software development process of the evaluated processes is subdivided and evaluated(S16). The development process needing the improvement is cleared according to the collected performance data(S17). The method to improve the cleared process is suggested(S18).

Description

Software Process Appraising Method

The present invention relates to a software process evaluation method, and more particularly, to a software process evaluation method that proposes a development process for determining software quality, that is, a method for improving software to be developed by evaluating a process.

The method of evaluating the process of developing conventional software is by a diagnostic program (evaluation based on the compliance of international standards of software processes). That is, it is configured to evaluate key software activities according to international standard criteria to add or improve processes for items rated "no".

1 is a diagram illustrating the conventional software process evaluation method.

First, the first step is to check the compliance of the software process according to the international standard Key Process Area (KPA). (S1)

In the second stage, the software process to be evaluated recognizes the problem of not complying with the international standard and analyzes the problems of the current software process. (S2)

In the third step, the level of the current software process is presented according to the result of the check in the first step (S3), and in the fourth step, a main item that needs to be improved at the current level is proposed (S4). do.

In the fifth step, we standardize it and apply it to the software development process. (S5)

However, the conventional software process evaluation method as described above does not improve the process by evaluating the process based on the productivity, quality, and schedule compliance performance of the software project, so that the improvement after the evaluation may not achieve the intended purpose. In other words, after assessing the current process, the items that are to be observed at the current level are blindly applied so that they can be applied regardless of the end purpose.

In addition, there may be a difference between the individual evaluating subjective evaluation in evaluating the evaluation, which may cause the current level to be different for each evaluator and thus may not focus on the areas requiring improvement. This can result in wasting resources on unnecessary parts.

In addition, in evaluating processes, it may be objective and not subjective to selecting areas for improvement. In other words, objective verification of major areas of improvement is not made, so it is difficult to predict the results after improvement before improvement.

Finally, applying all KPAs proposed by national standards takes considerable time and can result in higher loads than currently available, which can lead to poor performance of software projects. In other words, the prior art is not an improvement on a few important processes, but at least the KPA for each software development level has to be adhered to, so there is a problem that it may cause considerable time and developer's work efficiency to execute all of them.

The present invention has been made to solve the above problems of the prior art, the purpose of which is to evaluate the process of developing the software to be developed through an objective evaluation and grasp the development level of the current software and then It is to provide a method for evaluating a software process to improve the software process.

1 is a flow chart showing a conventional software process evaluation method,

2 is a flow chart illustrating a software process evaluation method in accordance with the present invention.

According to a feature of the software process evaluation method according to the present invention for solving the above problems, the first step of evaluating the overall process of the software development, and the performance data of each software project evaluated in the first step A second step of collecting, a third step of deriving a processor to be preceded in terms of infrastructure by using the evaluation result of the first step, and the evaluation result of the first step as independent variables collected in the second step A fourth step of analyzing the extent to which the software performance can be improved by performing the regression analysis using the performance data as a dependent variable, and a fifth step of identifying a process requiring improvement according to the analysis result of the fourth step; The software development process of the process evaluated in the first step is subdivided A sixth step of applying, a seventh step of identifying a development process requiring improvement as in the fourth step according to the performance data collected in the second step, and a process identified in the fifth and seventh steps The eighth step is to propose a possible way.

Hereinafter, exemplary embodiments of the present invention will be described in detail with reference to the accompanying drawings.

The software process evaluation method according to the present invention compares the evaluation results so that subjective evaluation can be described as objectified data and various matrices (defects, productivity, schedule compliance rate, hereinafter project performance) of the target software development project. Through the statistical process of analysis, it is through the process of analyzing what items (software detailed process, KPA, etc.) objectively affect the software project performance. It is designed to improve software project performance by further subdividing and evaluating bays.

Looking at the operation of the present invention configured as described above with reference to FIG.

First, in the eleventh step, the overall process of the software process (management process, organizational process, development process) is evaluated. (S11) The organizational process is "is there an education and training organization", "is there an SQA organization?" Is assessed through such a questionnaire. The management process is "requirement management?", "Is there a plan for the software project?" Is assessed through such a questionnaire. Is the development process "required analysis" or "is there a step-by-step review?" Is assessed through such a questionnaire. In particular, the questionnaire of each process is evaluated by dividing the process into four stages of “Yes / No / Not / Not Know” for each process. If the answer is “Yes,” the questionnaire is always kept well. 5 points for ", 3 points for" Keep as needed "and 1 points for" Keep for one-time only "to evaluate the project team leader and senior manager. .

In the twelfth step, the historical data on the project (evaluation in the eleventh step) (system test results, software related claims, pre-set schedule and actual completion schedule, number of lines of code of the developed software, and total number of inputs) To calculate the number of defects per line of code, claims per line of code, schedule compliance rate (actual completion time / planned time), and productivity.

In the thirteenth step, a process to be preceded in terms of infrastructure is derived using the result evaluated in the eleventh step. That is, the process corresponding to the infrastructure (management, organizational aspect) among the items commonly evaluated as "no" in the eleventh step is identified and a standard process thereof is presented.

In the fourteenth step, the relationship between the evaluation result and the performance of the eleventh step is analyzed by statistical analysis. (S14) That is, the total of the points given for the "yes" evaluated in the eleventh step as an independent variable, the performance collected in the twelve steps as a dependent variable is performed on the subject evaluated by performing a regression analysis It verifies whether it is related to performance, that is, improving the process can improve software performance. This, in turn, proves that performance can be improved if improvements are made to items with a rating of "no" and a process with a low score.

In the fifteenth step, the results of the fourteenth step are used to identify an organizational standard process that needs to be improved or added. (S15) Here, if the regression formula is established as described above as a result of the fourteenth step, improvement is made to the process (organizational and administrative aspects) with the priority among the items which are evaluated as "no" and the items with low grant points. If the regression equation does not hold, all the processes of the twelfth step are improved and the process proceeds to the sixteenth step.

In the 16th step, only the development process aspects are evaluated in detail to define the process according to the characteristics of each project. In other words, the evaluation is performed by the method evaluated in the eleventh step with respect to the process further refined with respect to the development process. The fragmented process presented here is a breakdown of the development process evaluation item of the eleventh step. For example, has the DFD been used for the needs analysis for the "requirement analysis" item of the eleventh step? Has been derived. "

In the seventeenth step, using the performance derived in the twelfth step, a development process that requires improvement for each project is identified through the same analysis as in the fourteenth step. That is, the main improvement items are derived in the same manner as in the fifteenth step by analyzing the analysis method in the fourteenth step. If the regression equation does not hold, the item with the highest priority among the items that are answered “no” is derived as the improvement item.

In the eighteenth step, an improvement method is proposed for the processes derived in the 15th and 17th steps. That is, the methodology that can be executed for the processes derived in the above steps 15 and 17 is presented. For example, if a project plan process is identified, the methodology suggests a PERT chart, schedule, and resource forecasting method. For the requirements analysis process, we propose the requirements analysis methodology such as state chart, ER_D, DFD.

The software evaluation method of the present invention configured as described above is a conventional software to reduce the software defects by improving the process in the development of the software into the electrical and electronic products, and to prevent the software development period delay with accurate resource allocation and accurate scheduling. By omitting the process of improving the process that was not used for development and objective analysis, only the areas that need improvement are identified so that the software performance can be closely linked with the process improvement and prevent the performance deterioration due to the improvement of the whole process. It can be effective.

Claims (8)

A first step of evaluating the overall process of software development; A second step of collecting performance data of each software project whose process is evaluated in the first step; A third step of deriving a processor to be preceded in terms of infrastructure by using the evaluation result of the first step; A fourth step of analyzing the degree to which the software performance can be improved by improving the process by performing a regression analysis using the performance data collected in the second step as the dependent variable as the independent variable as the independent variable; A fifth step of identifying a process requiring improvement according to the fourth step analysis result; A sixth step of subdividing and evaluating the software development process among the processes evaluated in the first step; A seventh step of identifying a development process requiring improvement as in the fourth step according to the performance data collected in the second step; And an eighth step of proposing a method of improving the processes identified in the fifth and seventh steps. The method of claim 1, The first step is to evaluate whether the process corresponding to each project for the organization process, the management process and the development process in the overall software development process divided into four stages of "Yes / No / Not applicable / Unknown" Software process evaluation method. The method of claim 2, If the evaluation is "Yes", the software process evaluation method characterized in that 5 points if the corresponding process is always made, 3 points if it is made as needed, 1 point if only one-time. The method of claim 2, The second step is a software process evaluation method, characterized in that to collect the historical data on the project in the first step evaluation, the number of defects and claims per line of code, schedule compliance rate and productivity. The method of claim 2, In the third step, the software process evaluation method, characterized in that it grasps items corresponding to management and organizational aspects among the items evaluated as "no" as the result evaluated in the first step, and presents standard processes thereof. The method of claim 2, In the fifth step, if the regression equation is established in the fourth step, improvement of the item in terms of organization and management in which the priority is higher among the items that are evaluated as "no" in the first step and the items with low scores are given. And if the regression equation does not hold, improve all items of the second step. The method of claim 2, In the sixth step, the software process evaluation method characterized in that the evaluation for the granular process of the development process evaluated in the first step. The method of claim 2, In the seventh step, the method for evaluating the software process as described in the fifth step through the analysis method of the fourth step.