With the development of information technology to increase the size of the system and applications, both the construction side of the system or the construction side, have an urgent need to build the organization's own data measurement system, in order to strengthen project control process, improve student rate, reduce production costs and enhance market competitive advantage.
Metrics for organizations to build their own needs, we must first determine the tools and methods of measurement used, and secondly to identify elements of the measure, that is what you want to measure data, and then build credible metrics database, and finally analyze the data metrics, continually improve process performance. The following construction practice through an organization based on Function Point Analysis metrics library, outlined the basic process of building a measure of system is as follows:
1, using Function Point Analysis (FPA)
Function Point Analysis has 30 years of development history, by IBM engineers Allan Albrecht in 1984, the first public release function point analysis method for software functional size measurement. Since 1986, the International Function Point Users Group (IFPUG) was established, its growing software functional size measurement method of Albrecht, has formed an international standard function point metrics, namely, ISO / IEC 20926 "IFPUG Functional Size Measurement Method." This standard provides detailed function point measurement method, from a user's point identification data function (ILF internal logical files and the EIF external interface files) and transaction function (EI external input / EQ outer query / EO external output), by calculating the 14 combined with the complexity of adjustment factor derived estimated function point (i.e., the size of software data). NESMA (Netherlands Software Metrics Association) functional point measurement method is improved, formation of the international standard ISO / IEC 24570 "Function Point Analysis application definitions and counting guidelines." The Standard noted that in different stages of demand, using different estimation parameters, such as in the early stages of the product, and demand has not been entirely clear split, FPA ILF only count the number of data files and ELF software to obtain a preliminary scale. Calculation rules are as follows: Overall the UFP (unadjusted function point) = 35xILF + 15xELF; gradually clear after the system requirements, the calculation method using the estimation function point function points. Calculation rules are as follows: Overall the UFP (unadjusted function point) = 10xILF + 7xELF + 4xEI + 5xEO + 4xEQ. Beijing software cost evaluation techniques Innovation Alliance (hereinafter referred to as "Union") launched calculation rules, it is based on the estimated function points under the two scenarios above method, by appropriate adjustment factor to calculate the function points the application system adjusted.
Function Point Analysis (FPA) has become the international standard, and has been recognized by most countries, which is the consistency of the standard, meet the application and consistent guidelines acceptable. Since 2013 the Ministry of industry-standard "software development cost metric specifications" and Beijing local standards "information software development project cost is estimated norms" has been released, more and more domestic software construction side and construction side of positive organizational learning FPA method, and in the metric system construction organization, and implementation of the project has been widely used.
Construction party applications, can be used Function Point Analysis, combined with the industry benchmark data, estimated the cost estimates of the project, to declare the budget, provide the basis for project bidding or audit basis; in the process of building the application system, can be collected by process data, to evaluate and manage the construction side, to identify problems and resolve project risks. Application system construction side, the use of function point method, you only need to estimate the function points of the project, combined with the industry benchmark data or organize their baseline data, multiplied by the corresponding adjustment factor, you will be able to obtain the corresponding metrics.
The benefits of the use of Function Point Analysis is that the method is based on a business perspective, from the perspective of the user identification data files and basic processes, without concern for how the technical realization, which will help users understand and apply. For the original source data is organized metric, this method has a high degree of credibility. In the metric system during the construction of the organization, function point analysis is a core capability need to master.
2, measure elements
To organize the project effective management, improve project performance and organizational processes, it is necessary to improve the measurement system by establishing organization-level metrics database. Function Point Analysis as a method consistent understanding of the parties, in favor of metrics collection and calculation of data, more and more organizations are currently using this method to obtain the desired metrics.
According to software process improvement features, organizations elements of software metrics there are many, measure factors related to the function point analysis includes estimating function points, the final function point, function point deviation rate, scale change factor, defect density, productivity, etc. data. Through these valid quantitative data, organizations can improve their own software project metrics system, and gradually form a production efficiency standard value data organization, data quality standard values and other data. Measurement element functionally relevant points are shown in Table 1:
Table 1 metric elements
After the measurement elements Function Point Analysis is determined based on, organizations can improve or build your own library of metrics.
3, the construction of metrics database
Organizational metrics database is a collection of project metrics data. To build metrics database, organizations must be done through the appropriate statistical measure template or construction applications. In addition, the organization also through the system in the form of provisions when the project team to provide any data, provided by what means, and which department the collection, collation and verification. Metrics collection and management of data must be by the person responsible, leadership attention, can not be a mere formality, otherwise the data collected up is inaccurate, useless data, not the purpose of the library building measure.
The following detailed description made by measuring the sample library building an organization, as shown in Table 2. In addition to measure elements of the table shows, the organization also collected other metrics, such as earned value analysis data, customer satisfaction.
Note 1: The initial data in Table 2 is derived from the functional point function point estimation method, the default project requirements has made it clear, otherwise, according to the estimated function point method into the estimates.
Note 2: For the system after the application delivery system maintenance tasks, due to the many factors involved, within the scope of the relevant data need to discuss improved function point measurement method, not in this article.
Table 2 project metrics data
4, measurement data analysis
After obtaining a measure of organization of data for each item, up to a certain number of project size, to form a measure of project data repository. With a measure of organization library is a first step, the key is how to make full use of these data, organizations can be analyzed to measure different dimensions of data, found that the characteristics and laws of the project, Dialectic policy. For example, the above organization by analyzing nearly three years of data, found that the average productivity Financial Management application system is 11.2 people / function point, the average productivity core business class application system is 14.6 people / function point, office automation category the average productivity of the application is 9 / feature point, in doing so or similar items, the average productivity can be applied as a reference the class, binding function point after adjustment, and maintenance workload estimate applications. In addition, statistical analysis, the average size of coefficient of variation application system of the organization to undertake three years is 1.2, then the estimating function point adjustment factor to change the scale set at 1.2. Obtain these data, effectively guide the smooth development of the organization's projects, the project implementation process clear, accurate and timely data collection and statistics. At the same time, the use of metric storehouse of data analysis, but also to evaluate the performance of the project's case, or by comparing process data, to identify problems and reduce project risk. On the organizational level, the results can be through data analysis, in order to continually improve organizational performance assessment project, update the organization to provide strong support measure aspects of the reference data.
Therefore, based on the construction of metric system Function Point Analysis (FPA), and for enhancing process control and the ability to organize and manage projects it is very effective. I believe that with the development of information technology, national standards published, as well as continuous improvement of function point analysis method, more and more organizations bound by the understanding, learning and use of Function Point Analysis, to build the organization's own metrics database, build efficient the metric system. (This article is copyrighted software cost Beijing Innovation Alliance evaluate all technology, Reprinted with attribution)