对日嵌入式软件外包工作量估算方式比较

【作者】 张凯；

【导师】 王安宇； 沈华红；

【作者基本信息】 复旦大学 ， 项目管理， 2009， 硕士

【摘要】 随着全球经济和社会发展需求的变化,信息产业的全球化,越来越多的日本企业开始往中国外包他们的业务,对中国企业来说这是一个很好的机会。然而项目中估算的不准确往往成为限制中国企业发展的桎梏。汽车导航软件是丰田公司的一套产品,其主要功能是通过GPS实现车辆定位,可以给用户提供诸如地图查询,路径规划,自动导航等功能。作者所从事的开发内容主要是该软件用户界面(UI,User Interface)层的功能,在对项目进行工作量估算的时候,因为准确性不够,与日本客户之间常常产生矛盾,因此我们希望能够找到一种估算方式,提高准确性,减少客户矛盾,扩大业务量。作者结合问卷调查的结果,选取功能点分析法和影响因素分析法这两种估算方法着重进行论述和分析,以期找到更加合适汽车导航软件项目的工作量估算方法。功能点分析方法是一个传统的估算方法,它是站在客户的角度去估算工作量的,通过分析软件系统所要实现的功能来计算软件项目功能点数,从而达到估算软件项目规模的目的,这是估算中最常用的方法。而影响因素分析方法是一种新的适合于对日嵌入式软件外包的估算方法,它是站在开发者的角度去估算一个软件项目规模。它关心的是用户界面中一个控件自身的变化以及对其他控件的影响。作者以历史数据为参考,建立了两种方法的估算模型。分别将这两种模型运用于汽车导航软件开发项目中,通过十个样本的分析比较,结果发现,功能性估算的工作量和实际工作量之间的偏差还是较大的,15%之内的偏差仅占40%。而影响因素分析方法估算出来的偏差相对较小,15%之内的偏差占76.6%。再进一步对利用功能点估算的结果偏差大的原因进行剖析,发现项目自身存在的局限性、软件开发人员的理解不一致等都会导致该方法得出的偏差大。因此,得出了影响因素分析方法更加适合汽车导航软件开发估算的结论。以汽车导航为切入点,是因为作者比较熟悉该项目,但是影响因素分析法并不局限于此,它还能被引入到其他的对日嵌入式软件开发中去,如数字家电,通讯工具等。通过本文的理论和实例分析,使我们对影响因素分析法有了更为全面的了解,也认识到了此方法的所存在的一些问题,这也是我们这些专业从事软件外包业务的人员在今后的工作中需要细致而深入地进行思考的关键。更多还原

【Abstract】 With the change of global economy and social development, also the globalization of information industry, more and more Japanese company transfer some of their business to Chinese company, which is an excellent opportunity for Chinese company. But the inaccurate estimation usually becomes the obstacle of Chinese company development.Car Navigation Softwar comes from TOYOTA Company. Its main function is to realize the car orientation through GPS, which can offer consumer some service, such as looking up map, designing the way to destination and navigating automatically. The main task for this project was to develop the function of user interface (UI). Due to inaccurate calculation of each workload, some contradiction often happened between two sides. In order to decrease the contradiction and expand the volume of business, a good method of estimation should be found.Combining with the result of questionnaire, the author emphasized to introduce and compare of two estimation methods, Function Point Analysis (FPA) and Influencing Factor Analysis (IFA), to find the most suitable method of estimating workload in the project of car navigation.FPA was a traditional estimation based on the consumer. It estimated the function point in the software by analyzing the target functions, to reach the objective of estimating the scale of this project, which is the most common way. IFA is one kind of new method suitable for embedded software in the Japanese engineering offshore, it is to estimate the project scale based on the software designer. It mainly focuses on the change of one component on the user interface and its influence for other components.The author took the history data as reference, then estabilish the estimation model of these two methods. The author used the two models into the car navigation software development, then compared ten samples and showed the result that the deviation of theoretical and real workload from FPA is larger, the deviation value within 15% only contained 40% among all samples, but the deviation of theoretical and real workload from IFA is smaller, the deviation value within 15% contained 76.6%. Then the author explained the reason why the big deviation came from the 1st method. Some aspects, such as the limitation of project itself and different understanding of software designers, caused the big deviation. It indicated that IFA is more suitable for workload calculation in car navigation software development. Using car navigation as the breakthrough point is because the author is more familiar with the project. However, IFA can be used in other fields, not only into embedded software in Japanese Software Outsourcing. It can be used into other Japanese Software Outsourcing, such as digital appliances or mobilephone etc. The analysis on theory and examples in this article made us have more comprehensive understanding of IFA, meanwhile also realized some problems in this method, which will make us do the detailed and in-depth work. The embedded software engineering offshore will be improve and perfected in the future.更多还原