Section: New Results

Improving Reusability and Automation in Software Process Lines

Software processes orchestrate manual or automatic tasks to create new software products that meet the requirements of specific projects. While most of the tasks are about inventiveness, modern developments also require recurrent, boring and time-consuming tasks (e.g., the IDE configuration, or the continuous integration setup). Such tasks struggle to be automated due to their various execution contexts according to the requirements of specific projects. We propose a methodology that benefits from an explicit modeling of a family of processes to identify the possible reuse of automated tasks in software processes [60] . Then, we propose a tool-supported approach that integrates both reuse and automation [61] . It consists of reusing processes from an SPL according to projects' requirements. The processes are bound to components that automate their execution. When the variability of a process to execute is not fully resolved, our approach consists of resolving this variability during the execution of this process. We illustrate our approach on industrial projects in a software company, as well as on a family of processes for designing and implementing modeling languages. Our approach promoted the identification of possible automated tasks for configuring IDEs and continuous integration, their reuse in various projects of the company, and the automation of their execution, while enabling to resolve process variability during the execution.