Best Practices Tailored for Research Teams Handling Diverse Projects in Parallel

Kurzfassung

Established technology management (TM) methodologies need to be tailored for small research teams with multiple projects in parallel. This paper aims to demonstrate some customized agile practices that our team adapted for our diverse research goals and software development in the last four years. Henceforth, the practices, tools and methods are categorized into automation, collaboration, visualization and management. For instance, Scrum is an agile method intended for a single project. But despite the challenges or obstructions in the earlier stages of adapting Scrum, it has shown success in our team of six people working on four to five different projects. Additionally, for one research software project a separate Scrum board was needed with traceability from the primary board. This helped create an agile work environment with improved efficiency and tackling of challenges that earlier went unnoticed until the later stages of the work. For project version control and database management Gitlab is used. In Gitlab - milestones, issue boards, iterations and performance charts - are useful for TM and performance measurement and management (PMM). Moreover, the Gitlab CI/CD pipelines help automate and improve the speed, efficiency and accuracy of the research software toolchain being developed. Moreover, Unified Modeling language diagrams, Architectural Decision Records (ADRs) and design decision meetings have helped across the different phases of the projects. It is also important to show which challenges were faced in the implementation and which trade-offs had to be made. Concluding, all key aspects, conventional or tailored, that helped us in our diverse team with limited time and human resources, will be presented in this paper. As a result, in all our projects some key facets of TM have improved. These include planning, milestones tracking, cross-functional collaboration and feedback, development and operations (DevOps), re-prioritization and PMM. It also helped us in knowledge management for the faster onboarding of new developers as well as recording the tacit knowledge of team members. This was achieved with sprint retrospective logs, mind maps, established workflows. In summary, despite hindrances we were able to customize conventional TM methods for the improved working of a diverse research team.