面向对象软件柔性的度量

为促进柔性软件的开发和研究,在总结柔性研究的基础上,通过分析柔性软件的特征提出了新的软件柔性的定义,并结合面向对象的技术给出了面向对象的软件柔性度量模型,该模型包括有效适应性、易用性、响应速率和成本比率4个侧面。为了使度量客观、具有操作性,对每一个侧面又进一步细化为若个要素,并具体定义了每一个要素的度量方法。     

混源软件质量模型与度量方法研究

由于混源软件包含自主代码、开源代码等不同来源代码,从而具有更高的多样性和复杂性,对其质量的度量评估与传统软件存在极大区别。为了度量混源软件质量,建立混源软件质量度量模型和方法是非常必要的。通过分析混源软件质量特性,提出混源软件质量模型。然后利用层次分析法、幂性法及线性法构建度量方法体系。最后对UbuntuKylin操作系统进行了实验性的度量评估,验证了模型与方法的可行性和有效性。     

基于LDA的软件演化确认建模

演化是软件生命周期中一个重要的部分。现在有大量软件已经演化了数个版本,而如何确认演化后的软件与演化目的相符合,成为了一个需要解决的问题。由于目前还没有一个系统的方法来处理此类问题,提出了采用LDA主题模型的方法对演化确认进行建模分析。用LDA方法对软件源代码中的某些特征进行建模,通过模型能够分析出源代码内潜在的主题。将提取分析出来的主题与软件演化发布的相关报告做对比,找出它们之间的区别,以此确认演化后的软件是否符合演化目的。     

软件柔性的评价

为了促进柔性软件的研究和开发,通过对柔性软件特征的分析,从软件的适应范围、适应的代价和反应速率等方面重新定义软件柔性;以此为基础提出了软件柔性的层次度量模型.该模型包括软件的有效适应性、可使用性及成本比率3个侧面,对侧面又进一步细化,分为适应范围、变性范围、易操作性、响应时间、实际成本和基线成本等6个要素;为使度量更客观准确,对每一个要素进行了定义,并提出每个要素的度量的方法,从而保证该模型具有可操作性,易于推广.     

基于多维属性的构件化软件演化相似性度量方法研究

对不同软件进行演化相似性度量和比较能帮助软件维护人员理解软件演化及预测其演化趋势。然而,传统的研究大多度量单一软件演化属性的变化,虽然有些也涉及到多维演化属性,但并没有与软件的演化相似性相联系;同时亦缺乏在较高抽象层次度量软件演化相似性的有效途径。为此,以构件为基本单位,提出了一种基于多维演化属性的构件化软件演化相似性度量模型。即通过在原子构件层和系统(复合构件)层定义演化属性,进而度量原子构件之间以及系统(复合构件)之间的演化相似性。初步的实验表明,在原型工具的支持下该方法能辅助软件维护人员进行软件演化相似性的判断。     

构件化软件演化信息及演化相似性度量技术研究

对软件演化信息的度量能发现软件演化过程中的问题,预测其变化趋势。传统的软件演化信息度量技术主要是以文件、目录或者项目作为度量的基本单位,缺乏在较高抽象层次对软件系统进行度量的能力。在基于构件的软件配置管理模型的基础上,以构件作为软件演化度量的基本单位,提出了一组基于属性变化、适用于构件以及软件系统演化度量的公式（包括较近、较早演化度量,演化相似性度量等）,并开发了原型支撑系统。     

考虑随机不确定性的常态仿真模型确认试验设计

模型确认试验是一种新的试验,其目的在于度量仿真模型的可信度.为了得到低成本、高可信度的模型确认试验方案,提出一种随机不确定性模型确认试验设计方法.首先,基于面积确认度量指标提出一种新的无量纲的模型确认度量指标(面积确认度量指标因子),并且在其基础上发展了基于专家系统的仿真模型准确性定性评判准则;然后,建立随机不确定性模型确认试验优化设计模型,提出该优化模型的求解方法;最后,通过两个数值算例对提出的模型确认试验设计方法进行验证.结果表明,小样本情况下,试验方案的随机性会影响模型评判结果的可信度;面积度量指标因子随试验样本数量的增加而收敛;随机不确定性模型确认试验设计方法能够避免试验方案对模型确认结果的影响.     

面向CMMI的软件度量分析过程研究与应用

为了辅助软件组织有效实施符合CMMI模型的软件度量过程,提出了一种GREDR过程框架模型,以指导软件度量的目标策划与过程实施。在实际软件项目过程中的应用结果表明：GREDR模型能够提高软件度量过程的系统性与效用。     

基于源代码的面向对象软件可维护性度量模型

针对当前软件可维护性的度量难度大的突出问题,在详细阐述软件可维护性的概念和内涵的基础上,以ISO/IEC9126为指导,建立基于源代码的面向对象软件的可维护性度量模型。该模型主要包括可维护性的4个子特性模型和一个对子特性的综合模型,能够客观定量地计算出软件的可维护性水平。最后通过一个实例,验证该模型具有一定的科学性和合理性,在实际应用中有一定的指导意义。     

面向对象软件度量工具的设计实现

利用软件度量工具对程序源代码的各类质量属性度量,对于提高程序质量具有重要意义。提出了一个适用于多种面向对象程序设计语言的软件度量工具的灵活软件结构,该结构的核心是以数据库为中间件。给出了该工具的结构,总结了该结构的特点和实用性,给出了以图形形式显示的分析结果和度量指标与质量特征之间关系的实验结果。     

Definition and collection of metrics for comprehensive software measurement

The approach to defining software metrics in terms of software models such as control graphs or call graphs has two limitations. First, since a model of the software will not, by its very nature, have all the properties of the software itself, the model-based approach is not well suited as a basis for comprehensive programmes of software measurement. Secondly, although software metrics are usually well defined in terms of the software models to which they relate, the models themselves are often not well defined in terms of the source code, or other software artefact, which they purport to measure. The paper describes approaches to metric definition designed to overcome these limitations.     

Applications of information theory to software engineering measurement

In the 1990s, there is an emphasis on finding ways to lower software cost and improve quality. Thus, it is very important to quantify and measure factors, such as software complexity, which have been shown to affect cost and quality. Researchers have defined many software complexity measures, and have exploited them to identify fault-prone program modules, to predict the number of faults that testing and operation will reveal, or to assess maintainability. Information theory based software measures are attractive because they quantify, with a standard unit, the amount of information in an abstraction of a program. The unit of measure is a bit. The purpose of this paper is to survey the state of the art of applications of information theory to software measurement, beginning in 1972. Information theory based measures have been applied to most phases of the development lifecycle. However, there have been relatively few papers. Most measures have not been empirically validated. One can conclude that the field is in its infancy. Relevant concepts of information theory are briefly described, and tables summarize the references from various perspectives. Since the amount of research, thus far, has been very limited, researchers will find numerous opportunities to validate, refine, and improve the measures presented here. Such research should give future practitioners useful software measures for each phase of the lifecycle.     

Properties of software measures

During recent years much attention has been directed towards the measurement of the properties and the complexity of software. The major goal of using software measures is to get reliable software and an objective representation of the properties of software and the software development process by numbers. Many software measures have been developed in order to determine the static complexity of single programs (intra-modular complexity) and of entire software systems (inter-modular complexity) during the phases of the software lifecycle. As a consequence of these developments many authors have discussed the properties of software measures. In this paper a measurement theory is introduced which gives conditions for the properties of measures. The properties of software measures and conditions of the use of software measures as an ordinal and ratio scale are given. As an example, these are applied to the measures of McCabe.Because composition and decomposition operations are major strategies in software development, theorems which describe the properties of software measures related to this type of operations are also presented. Properties of software measures, as required by masses in the literature, are discussed and explained with statements of measurement theory. The results show that it is possible to explain most of the required properties of software measures in the literature with conditions of measurement theory. This makes the properties of software measures during the software lifecycle and their application in practice more visible.     

基于度量的嵌入式软件缺陷风险分析研究

通过对COSMIC-FFP模型的扩展优化提出了嵌入式软件系统度量的方法,从而解决了COSMIC-FFP模型不支持对含有复杂数学算法的嵌入式实时系统度量的问题,基于软件规模度量提出了软件缺陷度量的方法。通过对软件规模的准确度量和对软件缺陷风险的分析,发现软件项目过程风险管理的不足,达到降低软件项目过程风险的目的。     

软件发布机制体系结构研究

针对现有软件发布方式主要以开发者的角度进行软件开发后生命周期管理所存在的问题,提出了更为合理的以用户的角度进行软件开发后生命周期管理的思想,建立了一种新的软件发布机制体系结构.该发布机制试图解决如下问题:让用户透明地使用软件,即用户不需要显式地进行软件的安装、更新和卸载;支持软件的即点即用和使用态发布,使用态发布指安装了某软件的计算机可以作为该软件的发布源;使得系统重装后能够以较小的代价恢复到指定的状态点.最后,分析了软件动态更新和分布式软件系统发布所应解决的问题.     

软件防反汇编技术研究*

为了保护软件所有权,根据一般的反汇编算法的特征,提出代码重叠、跳转地址重定向和控制流混淆等几种代码混淆技术。这些技术能使反汇编结果出现混淆,误导攻击者对程序理解,从而提高软件防反汇编的能力,有效地阻止对软件的逆向分析,保护了软件的知识产权。     

Anchoring the software process

Software organizations need common milestones to serve as a basis for their software development processes. The author proposes three such milestones, gives an example of their use, and discusses why they are success-critical for software projects. To avoid the problems of the previous model milestones-stakeholder mismatches, gold plating, inflexible point solutions, high risk downstream capabilities, and uncontrolled developments-software projects need a mix of flexibility and discipline. The risk driven content of the LCO, LCA, and IOC milestones let you tailor them to specific software situations and yet they remain general enough to apply to most software projects. And, because they emphasize stakeholder commitment to shared system objectives, they can provide your organization with a collaborative framework for successfully realizing software's most powerful capability: its ability to help people and organizations cope with change     

基于源代码分析的逆向建模

逆向建模通过对源代码进行分析,提取出代码中的对象信息、结构信息、流程信息等,生成对象间的关系描述、结构描述、系统流程描述等设计模型描述。逆向建模过程中对源代码的分析处理与编译过程的前端处理相似,只是处理的复杂程度与产生的目标结果不同,因此可以采用编译技术对源代码进行处理。通过逆向建模可以弥补软件设计中缺少或缺失的模型设计文档,帮助代码阅读者更好地理解程序,帮助软件的测试和优化。介绍了对C/C++源代码进行逆向建模的实现。     

基于度量的软件过程改进研究

在软件开发的整个过程中,需要对各个阶段进行评估、分析、改进,它们是一个不断循环的过程.即软件过程改进贯穿于软件开发的始终.如何对软件过程进行度量、度量元是什么,已经成为当前国内外研究的热点.对当前国内许多机构开展的质量管理体系认证模型进行研究,提出了一种基于CMMI的6σ软件度量框架,在此基础上分析了软件过程改进和评估中的关键问题.通过对软件过程的改进,从而提高软件产品质量和提升软件开发效率.