\section{Introduction}

E-government projects differ from others due to their complexity and
extension\cite{anthopoulos2016egovernment}. They are complex because they combine construction, innovation, information \& communications technologies, politics and social impact. Their extension, on the other hand, is related to their scope, target audience, organizational size, time and the corresponding resistance
to change. Government-academia researches can be considered a way to create novelty for e-government projects and to meet the needs of society. However, this collaborative work also has challenges, such as to organize the project, to align goals, to synchronize the pace of both sides\cite{sandberg2017iacollaboration}, and to overcome the failure
trend of e-government projects \cite{goldfinch2007pessimism}.

Poor project management is one of the main reasons why
e-government projects fail\cite{anthopoulos2016egovernment}. In Brazil, while
industry and academia prefer agile approaches to manage their projects, government organizations generally use traditional
methods to discipline its software development. When government and academia decide to join forces to develop an e-government solution, these differences in project management become an issue. Changing the software development process in a large-size institution represents a complex organizational change that has impacts on structure, culture, and management practices \cite{nerur2015challenges}, which will limit its feasibility in projects with tight deadlines and short budgets.

This paper presents practices based on open source ecosystems and agile
methodologies, and adopted to harmonize differences between government and academia project management. We map
the management practices of the referred project by examining a 30-month government-academia collaboration case. Then we show benefits of this empirical model, using collected data from repository management tools and from project participants surveyed: analysts from the Brazilian Ministry of Planning (MPOG) and developers from the University of Brasília and the University of São Paulo. At the end, we compare the results of this current work with lessons learned in a previous paper\cite{meirelles2017spb}. 

Section \ref{sec:relatedwork} describes related work. Section 
\ref{sec:researchdesign} explains our research questions and research
methodology with a brief description of the case study. Section \ref{sec:results}
presents results derived from our quantitative and qualitative analyses.
Finally, we discuss our findings and future work in section \ref{sec:discussion}.