Supporting software engineering with knowledge acquisition, formalization, and application

软件开发是一个设计过程，在这个过程中，每个人都必须做出许多决策，每个决策都有几个可能的选择。这些决策涉及不同的方面，例如，设计、实现、测试、维护等。软件从业者需要支持来决定哪种替代方案更适合他们的特定环境。提高开发“好的”软件的能力意味着分析以前几个项目的结果以确定它们之间的相似和不同之处。这些分析的结果可以表述为知识。这些知识的例子有模式、启发、最佳实践、评估模型等。我们认为，如果在软件开发和项目期间生成的知识被捕获、结构化和形式化，则软件工程性能可以得到改善。然而，许多知识是隐含的，驻留在软件从业者或工程师的大脑中。在这种情况下，我们认为：1)不同的知识来源和类型可以通过不同的方法来处理，并可以在软件生命周期的不同阶段为软件从业者提供支持；例如，识别相关的模式和模式组合，促进可重用性，做出设计决策，修复测试过程中发现的软件缺陷，分析维护过程中的更改影响等。2)我们需要得到适当的技术和工具的支持。人工智能技术正被用于知识发现、获取和形式化；例如，用于定义/使用各种本体、用于知识表示和利用、机器学习、形式概念分析等。

Software development is a design process where every involved person has to make a number of decisions, each of them with several possible choices. These decisions concern different aspects e.g., designing, implementing, testing, maintaining, etc. Software practitioners need support for deciding which alternative is better for their specific context. Improving the ability to develop "good" software implies activities that analyze results from several previous projects in order to identify similarities and differences between them. The results of these analyses can be formulated as knowledge. Examples of such knowledge are patterns, heuristics, best practices, assessment models, etc.We consider that software engineering performances could be improved if knowledge generated during software developments and projects is captured, structured, and formalized. However, a lot of knowledge is implicit, residing in the brains of the software practitioners or engineers. This fact makes knowledge elicitation, acquisition, and formalization a challenge.In this context, we believe that:1) Different sources and types of knowledge can be handled by various approaches, and can support software practitioners at different stages of the software life cycle; for instance, for identifying relevant patterns and pattern combinations, promoting reusability, making design decisions, fixing software defects found during testing, analyzing change impact during maintenance, etc.2) We need to be supported by appropriate technology and tools. Artificial intelligence techniques are being used for knowledge discovery, acquisition, and formalization; for instance for defining/using various ontologies, for knowledge representation and utilization, machine-learning, formal concept analysis, etc.