改善软件开发工作强度，提高软件系统开发效率是软件工程研究多年以来的重要目的与目标。尽管软件体系结构的高层次抽象特性能够为软件系统的设计与开发提供高层次复用的手段与基础，软件体系结构技术的应用仍然受到从抽象到具体进行过渡的局限限制。本文提出面向软件体系结构和多Agent联邦技术的大型软件系统设计与合成方法可以有效地利用多Agent联邦技术可以为软件体系结构的设计结果到软件系统的实际开发提供一种从抽象到具体的平滑过渡手段。

当前，XML已经成为一种可广泛应用于Internet数据存储和交换的半结构化数据文档信息管理技术。但是，即使是在DTD的说明配合下，Internet上普遍存在的不同XML数据源之间的集成与连接操作对于广大程序员而言，依然是一项极其复杂和繁琐的工作。本文根据XML文档的DTD语义特点，提出并实现了一种基于软件Agent来辅助进行DTD文档分析的XML文档的集成系统框架。有关的研究与应用实践表明，通过采用软件Agent的来引导和辅助XML文档的DTD语义分析与集成，能够在增加处理工作的自动化程度的基础上，极大的提高XML文档与数据的集成和交换自动化程度，有效的减缓软件人员的工作强度。

When we look back to the history of Internet, we can see that the need for human being to communicate with computers through Internet has made HTML the most popular tool for application of first generation World Wide Web on Internet. This fact gives us a sense that the need for more and more computers to communicate each other through Internet will make XML the mostly used technology for computers on Internet to exchange and store distribute data in near future. But even with the help of DTD and schema to today's XML documents & data sources, the programming works for programmers to integrate XML documents through Internet is still very hard and complex. One of the important reasons to this situation is that the programmers who want to integrate the XML documents have not only to analyze the document's original DTD, but also to write a new DTD for the new integrated XML data. Because DTD is really a specification for one XML document, it is only suitable for software & programs that have basic knowledge of DTD to analyze and create them. Based on the character of XML documents and their DTD specifications, this paper proposed a new system framework to use software agent to help the work for XML document & data integration. While software agents have DTD knowledge for them to automatically analyze and memorize the character of XML's DTD, they are absolutely effective in works of DTD analysis and XML creation. This character makes the system with the framework of this paper efficiently relieve the programmers' work by the automatic processing in XML document’s integration.

While organizations have invested millions of dollars on their software systems and it is impossible to produce systems of any size that do not need to be changed, software evolution for requirement change is getting more and more important. So far, most of the efforts on software evolution have focused on system function change or system architecture transfer. The number of evolution for core data model adding is relative few. Software architecture is a good tool for software design and software analysis, but it has limitation on software evolution when there need some architecture inheriting for core data model adding. In this paper, we proposed method using an extended software architecture model to evolve exist software with additional data models. The method is applied to evolve a visualization software platform from plane data processing ability to a visualization platform with ability of processing both plane data and volume data.

可视化软件开发平台所支持的数据模型和系统框架是决定可视化应用的使用范围与处理能力的两个重要特性。传统的可视化开发平台在多数据模型处理和与分离场景数据包容等方面还不能做到有效的协和与统一。本文采用扩充的软件体系结构的系统分析与设计方法，对大型可视化软件开发平台进行了基于软件体系结构的进化分析与设计，在较好的保持传统系统的原有特性的基础之上，有效的将一个只支持面模型处理应用开发的可视化软件设计开发支持平台进化扩充成为一个能够同时支持面模型和体模型应用开发的综合模型支持系统。