扫码阅读
手机扫码阅读

快速学习COSMIC方法之四:早期快速估算功能规模的方法

30 2024-10-02

我们非常重视原创文章,为尊重知识产权并避免潜在的版权问题,我们在此提供文章的摘要供您初步了解。如果您想要查阅更为详尽的内容,访问作者的公众号页面获取完整文章。

查看原文:快速学习COSMIC方法之四:早期快速估算功能规模的方法
文章来源:
麦哲思科技任甲林
扫码关注公众号
Summary of Approximate Estimation Methods for Software Sizing

Summary of Approximate Estimation Methods for Software Sizing

Before diving into the detailed COSMIC method, it is important to understand how to estimate the size of software in the early stages of a project when requirements are not detailed enough for testing. Many companies use approximate estimation methods to determine project budgets and reduce the workload of measurement.

The principle of quick estimation is to analyze the correlation between historical coarse-grained requirements and actual size to find a conversion relationship. The scale of coarse-grained requirements can be based on various factors such as the number of functional processes, number of use cases, number of document pages, or number of developers, as long as there is a strong correlation with functional point size evidenced by historical data.

In the early stages and quick guides of COSMIC, four approximate estimation methods are introduced:

  1. Average Size per Functional Process: Companies can use historical data to find a correlation between the number of functional processes and functional point size. By dividing the number of functional points by the number of functional processes, an average value 'a' is obtained, which is then used for future project size estimations.
  2. Fixed-Size Classification Approximation: By counting functional points for each functional process and classifying them into size categories (e.g., large, medium, small), companies can estimate the average functional point size for each category. Significant differences between categories are necessary, and criteria can be established to help in classification.
  3. Equal-Size Band Method: This more complex method involves sorting functional processes by size and dividing them into bands with equal cumulative functional point totals. Average sizes for each band are then calculated, and these categories can be used for future estimations with the help of objective criteria.
  4. Average Use Case Method: Similar to the first method, it also calculates the ratio between functional processes and functional points but includes measuring the number of use cases. This method may offer lower accuracy and requires a standardized template for writing use cases.

It is vital to understand that approximate estimation relies on historical data. A company must first measure some typical projects using the standard COSMIC method to establish a baseline for these approximate estimation methods. These relationships need to be regularly calibrated with actual data to ensure that estimations are as close to reality as possible.

While industry benchmark data might seem appealing due to its convenience and lack of effort in data accumulation, it comes with inherent inaccuracies that may not reflect individual company nuances. Therefore, companies should not rely solely on industry data but should invest time in accumulating their own data to benefit from the "slow first, fast later" approach of rapid estimation methods.

想要了解更多内容?

查看原文:快速学习COSMIC方法之四:早期快速估算功能规模的方法
文章来源:
麦哲思科技任甲林
扫码关注公众号

麦哲思科技(北京)有限公司总经理 敏捷性能合弄模型评估师 认证的Scrum Master 认证的大规模敏捷顾问SPC CMMI高成熟度主任评估师 COSMIC MPC,IAC 成员,中国分部主席

420 篇文章
浏览 63.3K
加入社区微信群
与行业大咖零距离交流学习
软件研发质量管理体系建设 白皮书上线