基于InCoME的软件成本估算方法及支撑工具

软件成本估算领域经过四十余年的发展，涌现出一大批估算理论与方法，但都没有在现实环境中的软件企业中得到广泛应用，在项目早期进行软件成本估算仍旧是一件非常难的任务。由于估算模型的复杂性，缺乏相关的自动化的支撑工具及在现实软件企业中进行成本估算的应用研究是造成这一局面的一个重要原因。 本文从解决现实估算问题出发，提出了基于支撑工具的软件成本估算应用的具体框架。使用专家知识初始化模型，收集历史数据进行模型校准，并使用jack-knife交叉验证对模型进行精度分析。在建立可接受的模型后，收集待估算项目规模和成本因子数据，基于集成成本建模与估算（InCoME）方法，提供COCOMO、类比、回归等多种估算方法的支持。估算结果结合不确定性分析和风险分析，给项目计划和决策提供参考。成本估算应用框架形成一套完整和规范的流程，是一个现实可行的软件成本估算的解决方案。本文的另一贡献在于定制开发此应用框架的支撑工具，即集成成本建模与估算工具。在分析调研主流的软件成本估算工具的基础上，为配合估算应用框架，使用eclipse RCP和关系数据库，开发出层次清晰、可扩展性强、可维护性高、易升级易部署、界面友好的支撑工具。我们将基于支撑工具的软件成本估算应用框架应用于现实中的软件企业进行经验研究，经验研究结果表明，企业的软件成本估算得到了明显改善，支撑工具也被很好的接受。

InCoME软件成本估算工具

针对现行的成本估算软件存在算法支持不够、算法不公开、对建模支持不够等问题,提出了一种基于最优加权算法的集成软件成本估算方法InCoME,并且在此基础上采用Eclipse RCP框架、Java开发语言、HSQLDB关系数据库开发出了In-CoME成本估算软件。该软件不仅实现了驱动因子管理、集成模型支持、模型校准、模型精度分析等模块,而且还实现了基于不确定性的成本估算和风险分析功能,解决了国内一些企业急需成本估算软件等问题。

职业倦怠与自我和谐、人际信任的关系研究：中学校长与一般教师的对比

Job Burnout has been a focus of the Occupational Stress Research. As a typical,helping occupation, teacher has attracted widely attention and researches in the areas of pedagogy and psychology. The special subgroup of teacher, headmasters who are the elites of the Basic Education, is ignored. The research about principals’ Job Burnout is nearly blank after analyzing related documents and information. With the development of the society, people pay more and more attention to the education and put more demands on the headmasters, especially middle-school principals. They are required not only to be good educators, who are equipped with all the inner qualities as a teacher, but also good managers. So the main purpose of this research was to compare the principal group with ordinary teacher group, and reveal underling factors, such as background variables and psychological protection variables. A representative sample of Wenzhou middle school principals sized 192 and a sample of middle school teacher sized 302 were sampled from various schools. The educational version of burnout inventory, self consistency scale, and interpersonal trust scale were administrated to the two samples, together with some demographic variables of interest. The applicability and equivalence of the three instruments used in this study were checked. Based on well-established reliability and cross-sample congruence of measures, the difference between principals and teachers was test. Then the contributing factors were analysis gradually. The five background variables were examined one by one in the two samples separately. A multiple covariance analysis was conducted to test whether there remained any difference between these two samples on the variables of interest. Regression analysis was used to further control the effect of self harmony and interpersonal trust to test the difference between two samples. Mediating analysis was conducted to build the relationship among the three constructs. The main results of the research were stated as following: 1. The internal consistency coefficients of all the scales were good, and no difference exited between the two groups. The measurement equivalence of three instruments was established well. The measures could be applied to and comparing the two samples. 2. The self-harmony, and interpersonal trust of principals were better than the ordinary middle-school teachers. Job Burnout of principals was significant lower than teachers. 3. Demographic variables like the gender, age groups, income levels, disricts, and the type of school, were important influencing factors. The difference patterns of the variables on these five variables in two samples had similarity and distinction. 4. After controlling the background variables, there remained significant difference between principals and teachers on the variables of interest. 5. Job Burnout negatively correlated with self-harmony and interpersonal trust. That is to say，the lower the degree of self-harmony and interpersonal are, the serious of the Job Burnout is, The correlation between the self-harmony and the interpersonal trust was positive. 6. After statistically controlling the background variables and psychological variables, there still exited significant difference between two groups of this study. Also, self harmony and interpersonal trust were significant protection predictors to different aspect of job burnout. 7. Mediating analysis was conducted to the residual score of the three constructs after controlling the five variables and group membership. Self harmony partially mediated the relationship between interpersonal trust and job burnout. That is, interpersonal trust had indirect effect to burnout mediated by self harmony, also had direct effect to burnout.