How to estimate test workload (3) cocomo Model

How to estimate the test workload

(3.) CocomoModel-Common Software Scale Estimation Methods

As a measure estimation method, the code line analysis method has been widely developed in the S and S. It has developed many parameter models for estimating workload and progress in the industry, the most famous one is the cocomo model. The latest version is the cocomo II model.

Cocomo is called constructive cost model in English, and Chinese is a constructive cost model. it is a precise and easy-to-use model-based cost estimation method, which was first proposed by Boehm in 1981. In essence, it is a parameterized project estimation method. parameter modeling uses some features of the target as parameters, build a digital model to predict the project cost (similar to the overall housing cost calculated by using the residential area as a parameter ).

In the cocomo model, the workload adjustment factor (EAF) represents the comprehensive effect of multiple parameters. These parameters enable the project to be characterized and normalized according to the project in the cocomo database. each parameter can be located very low, low, normal, high, very high. each parameter is used as a multiplier, and its value usually ranges from 0.5 to 1.5. The product of these parameters is used as a coefficient in the cost equation.

Cocomo uses three models of different levels to reflect different degrees of complexity.

L basic model is a static single variable model. It uses a function that uses the estimated number of source code lines (LOC) as the independent variable to calculate the workload of software development.

L intermediate model ). then, based on the workload of the function compute software development that uses LOC as the independent variable, the workload estimation is adjusted based on the factors that affect the attributes of products, hardware, personnel, and projects.

L The detailed model includes all the features of the intermediate cocomo model. However, when workload estimation is adjusted using the various influencing factors described above, the impact on the analysis, design, and other steps in the software engineering process should also be considered.

At the same time, according to different application fields of different application software, the cocomo model is divided into the following three software application development modes:

L Organization Mode ). the main feature of this application development model is to develop projects in a familiar and stable environment. There are many similarities between the projects and other recently developed projects, and the projects are relatively small, there is no need for many innovations.

L embedded application development mode (embedded mode ). in this application development model, projects are restricted by interface requirements. the interface is very demanding for the development of the entire application, and requires great project innovation, such as developing a brand new game.

L semidetached mode, which is the type between organization mode and embedded application development mode.

The cocomo model is accurate and easy to use. The basic quantity used in this model includes the following: (1) DSI (number of source command lines), defined as the number of code lines, including all the Code except the comment line. If a row contains two statements, a command is executed. (2) mm (measurement unit is month) indicates the development workload. (3) tdev (measurement unit: month) indicates the development progress, which is determined by the workload. (4) The cocomo model focuses on 15 factors that affect the software workload and defines multiplication factors to accurately and reasonably estimate the software workload.

However, cocomo also has some serious defects. For example, the input during analysis takes precedence and cannot handle unexpected environment transformations. The obtained data is often not used directly and must be calibrated, you can only get a Summary of the past and cannot calibrate future situations.

Reference: Some Excerpted from <function analysis methods and practices>