FORFLOW

BAVARIAN RESEARCH ASSOCIATION FOR PROCESS- AND WORKFLOW SUPPORT FOR PLANNING AND CONTROLLING PROCEDURES IN THE PRODUCT DEVELOPMENT

3.1 - Context-based search for reusable components

Success in development processes is contingent on providing the develop- ment engineers with the right information at the right time. Attaining this goal is not feasible by merely digitalising information. The challenge lies in endeavoring to supply the engineer with information specific to his needs, as the costs of information search must be in due proportion to their antici- pated benefits. Hence, informing the development engineer of the where- abouts of significant documents is only a first step; in particular, an imple- mentation of easy access methods and a proactive notification regarding potentially relevant documents impact effectiveness and efficiency in de- velopmental processes. Relevant information comprises reusable artifacts as well as documents being helpful in decision situations or indicating best practises; those documents can contain results of previous projects as well as catalog items, or they can be found in publicly available sources. As supplying engineers with information specific to their needs is of para- mount importance to mastering developmental processes, numerous question need to be addressed. Was a similar problem solved in the past? To what result? What documents could be helpful in a current situation? How can manual methods, like keyword searching or directory tree browsing, be replaced by more sophisticated approaches? Partially automated methods, employed to retrieve relevant components and documents, should draw on a developer’s context, which is constituted by activities and contents typical to the process at hand. To provide a com- mon ground for inter-project searches, the notion of a task, characterized by a source and an effect document, is introduced. Within this framework, supporting an engineer working on a particular project implies looking for effect documents from other projects, whose source documents match the current project’s source document. Offering searches based on this approach demands modelling the context of developmental processes and offering a similarity search based on this context, thereby facilitating a concerted supply of meaningful information. It is this project’s objective to conceptually design and prototypically imple ment a system potentiating context-based similarity search for the product development domain. The project draws on preliminary works following an analogous procedure in the field of software engineering.