SoD: Supporting Cased-Based Design for Packaged Software Implementations

本项目侧重于基于案例的设计（CBD）。CBD解决了被认为是打包软件的软件类缺乏有组织的设计理论的问题。CBD通过重用已有的设计知识来创造新的设计知识，从而依靠一个不断循环的知识创造。通过创建基于案例的设计的理论框架（作为组织知识创建过程）以及在打包软件实现中支持CBD的创新工具，CBD从这个项目中获益。该工具允许用户协同创建、编辑和分类设计案例。用户能够沿着多个用户进化的方面探索复杂的设计空间。该工具通过利用用户分类和推荐工作，系统地推荐现有设计，以便在新情况下重用。我们的基本信念是，通过利用一大群人的集体力量，可以以更低的成本和更低的风险实现更好的设计。这个项目的动机是由于信息系统使用大型软件包（如ERP（企业资源规划）软件）的日益增长的趋势所带来的问题。特别成问题的是大型系统实现的复杂性和持续的维护。这些软件系统很少打包使用。相反，它们是定制的、重新设计的、增强的、扩展的，并集成到其他现有的和正在发展的系统中。打包软件实现中的设计（PSI）是通过定制和集成现成软件包来解决业务问题的过程。PSI中的设计侧重于附加软件组件或对开箱即用软件包的修改。有效和高效的设计可以带来可观的回报和更低的实现成本，并可能最终导致更广泛地采用打包软件。这不仅有利于软件供应商，也有利于社会，因为它减少了重新发明轮子的定制软件开发工作。然而，PSI缺乏通用的设计方法和CASE（计算机辅助软件工程）工具。PSI的设计很大程度上依赖于以前的设计经验。事实上，PSI专家在面对新的问题情境时经常采用基于案例的设计（CBD）。他们探索过去的设计案例，找到类似的案例，并在新的问题情况下重用该案例的设计。由于同一个软件包被部署在许多实现中，CBD可以显著提高质量，降低成本和降低设计风险。然而，CBD只能作为设计案例的知识库。使用CBD的挑战之一是缺乏支持设计知识库增长的CASE工具。因此，该项目开发并评估了一个开源工具，以支持打包软件实现（PSI）中的基于案例的设计（CBD）。如果与大型设计知识库一起有效使用，基于案例的设计可以显著降低成本、降低风险并提高设计质量。

This project focuses on Case-Based Design (CBD). CBD addresses the lack of an organized design theory for the class of software identified as packaged software. CBD reuses existing design knowledge to create new design knowledge, thus relying on a continuous cycle of knowledge creation. CBD profits from this project through the creation of a theoretical framework of case-based design as an organizational knowledge creation process as well as from an innovative tool to support CBD in packaged software implementations. The tool allows users to collaboratively create, edit and classify design cases. Users are able to explore complex design spaces along multiple user-evolved facets. The tool systematically recommends existing designs for reuse in new situations by utilizing user classifications and recommendation efforts. The fundamental belief is that by utilizing the collective power of a large group of people, better designs can be achieved at lower costs with lower risks. The motivations for this project are the problems imposed by the increasing trend in information systems to use large software packages such as ERP (Enterprise Resource Planning) software. Particularly problematical are large system implementation complexities and on-going maintenance. These software systems are rarely used as packaged. Rather, they are customized, redesigned, enhanced, expanded, and integrated into other existing and evolving systems. Design in packaged software implementation (PSI) is the process of solving business problems by customizing and integrating the off-the-shelf software package. Design in PSI focuses on the add-on software components or modifications to the out-of-box software package. Effective and efficient designs can bring substantial returns and lower implementation costs and may eventually lead to wider adoption of packaged software. This benefits not only the software vendors but also society by lowering custom software development efforts of reinventing the wheel. However, there has been a lack of common design methodologies and CASE (Computer Aided Software Engineering) tools for PSI. Designs in PSI rely heavily on previous design experience. In fact, PSI experts frequently practice case-based design (CBD) when facing a new problem situation. They explore the past design cases, find a similar case, and reuse the design for that case in the new problem situation. As the same software package is being deployed in many implementations, CBD can significantly improve the quality, reduce the cost and lower the risks of designs. However, CBD can only be as good as the knowledge base of design cases. One of the challenges to utilizing CBD is the lack of CASE tools that support growing the design knowledge base. Consequently, this project develops and evaluates an open-source tool to support case-based design (CBD) in packaged software implementations (PSI). If used effectively with a large design knowledge base, case-based design can significantly reduce the cost, lower the risks and improve the quality of designs.