From examples to knowledge to software engineering support and automation

在过去的二十年中，已经做出了很多努力来生产支持软件开发和维护的环境和工具，例如，建模语言、库等。然而，使用这些环境和工具并不容易，只有一些专家具有正确有效地使用它们所需的专业知识和知识。我们的研究计划的目标是利用这有限数量的专家的知识和专长，以一种能够使更多的开发人员受益的方式。为此，我们建议使用这些专家执行的用法示例来学习必要的知识，这些知识可以自动化操作，或者支持开发人员使用这些环境和工具。得出这样的例子可以采取不同的形式。例如，当专家开发一个正确使用库的程序时，可以收集执行示例并从中学习如何正确使用库的模式。然后，将这些模式转换为推荐系统，以指导新开发人员使用此类库。另一个说明性的情况是操作根据特定于领域的语言编写的模型。通过观察操作这些模型的专家，即在操作任务之前和之后的模型，可以得出示例。然后，可以从这些示例中学习操作任务。

During the last two decades, much effort has been made to produce environments and tools to support software development and maintenance, e.g., modeling languages, libraries, etc. However, using these environments and tools is not easy, and only some experts have the necessary expertise and knowledge to use them correctly and efficiently. The goal of our research program is to leverage the knowledge and expertise of this limited number of experts in a way that can benefit to a larger number of developers. To this end, we propose to use examples of usages performed by these experts to learn the necessary knowledge that is operationalized to automate, or to support developers in, the use of these environments and tools. Deriving such examples can take different forms. For instance, when an expert develops a program that correctly uses a library, one can collect execution examples and learn from thempatterns on how to correctly use the library. Such patterns are then, transformed into recommending systems to guide new developers in using such a library. Another illustrative situation is one of manipulating models written according to a domain-specific language. Examples can be derived by observing an expert that manipulates these models, i.e., models before and after a manipulation task. Then, the manipulation tasks can be learned from these examples.