We are starting a new series, “Project of the day”, where we will weekly introduce a research project underway at the Pervasive Computing lab. First up is an Academy of Finland project GloFuMe, which is just ending.
GloFuMe stands for “Optimizing Global Software Projects with Fuzzy Meta-Heuristics”. As the title suggests, the project deals with a software projects (in the context of global software development), and meta-heuristics – with the added flavoursof utilizing fuzzy logic to help with non-numerical input that would be given to the heuristic algorithms. The main idea was to study global software development projects particularly from an architecture-centric viewpoint – how could the software architecture and the project organization (tasks, teams, people, skills, etc.) be aligned to optimize the projects in terms of effort, cost and time – among other things.
Thus, the project originally had the following objectives:
• Providing a framework for optimizing cost, duration, effort, skills utilization, communication and work allocation in GSD projects
• Linking software architecture solutions to GSD project management issues
• Providing a tool realizing the GSD project optimization framework, augmented with architecture solution proposals
The ultimate goal was thus a tool, where software architecture design would co-evolve with the software project setup, and the two would be continually aligned and optimized to best fit the organization and its goals.
The project started out by mapping software architecting challenges and practics. Together with researchers from Lero – the Irish Software Research Centre, we performed a systematic literature review to answer the questions of what kind of challanges exist in architecting in GSD projects, and what kind of practices there are. We then continued the work by interviewing architects from several companies engaged in distributed software development. We asked about the challenges they had encountered in distributed software development, what kind of practices they had and particularly what kind of practices they had in software architecting.
Throughout the project there was a need to gain access to actual project data from a company involved in distributed software development. Due to a number of hiccups in securing this data, work that would lead to the practical implementation of the tool and algorithms (all requiring the project data) has only started in the past six months. We plan to do case studies with the data and then begin constructive research by actually building the algorithms and the tool and testing it within the company setup – at the same time reflecting on what we have learned from the background studies.
The project started in September 2014, and had funding for three years from the Academy of Finland. As a post-doc project, there is only funding for the main researcher, the undersigned, instead of having an actual research team. The project has been carried out solely by the undersigned, Dr. Sievi-Korte, who has a background in software architecture and utilizing genetic algorithms. There has been very close collaboration with people from Lero, and now with a Finnish company.
The project timeframe was extended for one year as the main researcher took time away to have a baby, and so the project is now ending in mid-September, 2018. The work will still continue, though, as we now have working collaboration with a company. Thus, the project goals are well underway, although in a slower timeframe than expected. We already have as results checklists for architecting in GSD, practice-challenge frameworks and guidances on how to aid the social aspects of GSD. More results to be expected in the coming years.