04-methods.tex 6.95 KB
\section{Research Design}
\label{sec:researchdesign}

The focus on this paper is investigating practical ways to reconcile cultural
differences in software development process between academia and government,
without modifying their internal processes. Our analysis was guided by the
following research questions:

\textbf{RQ1.}\textit{What practices based on open source development experiences would
help to combine teams with different management processes in a
government-academia collaboration project?}


\textbf{RQ2.}\textit{How do open source development practices benefit the process of
developing an e-government platform in a government-academia collaboration?}

To answer these questions, we use as a case study the evolution project of the
SPB portal \cite{meirelles2017spb}, a government-academia collaborative
development based on open source software integration. We designed two surveys
and an interview to the different roles performed by the ex-project
participants and collect public data from the project development environment
available on the developed platform itself. Our research approach is detailed
in the following subsections.

\subsection{The case study}

The project to evolve the Brazilian Public Software Portal
was a partnership between government and academia held
between 2014 and 2016\cite{meirelles2017spb}. To solve maintenance problems and fill
design-reality gaps in the portal, the Ministry of Planning (MPOG) joined the
University of Brasília (UnB) and the University of São Paulo (USP) to develop a
platform based on the integration and evolution of five existing open source
software.this environment was a novelty in the context of the Brazilian government, due to the technologies employed and its diverse features, including social networking (Noosfero), mailing lists (MailMan), version control system (GitLab), and source code quality monitoring (Mezuro), all integrated using a system-of-systems software (Colab).

The academic team carried out development activities in the Advanced Laboratory
of Production, Research and Innovation in Software Engineering (LAPPIS) at UnB.
The project management and development process in this laboratory is usually
executed adopting empirical practices from open source communities and agile methodologies. For this project,
a total of 42 undergraduate students and two coordinator-professors participated in the development team. Six IT professionals were also hired as senior developers due their experiences in open source projects and two designers specialized in User eXperience. 

The government team was composed of one director, one coordinator, and two IT
analysts from a department of MPOG. Although they were responsible for the
execution of this collaboration, their department generally does not
execute development of ministry's software, its responsibility is
contracting and homologating software development services, following
traditional management approaches, such as the RUP, CMMI, and PMBOK.

% Conteúdo OK melhorar construção
These two aforementioned teams periodically met in person for the purpose of
managing the project progress, discussing strategic/political and technical goals. Initially, these meetings took place at the ministry's headquarters and, usually, only directors and professors participated. The management of the development team was
concentrated in the academic side and was organized in
biweekly sprints and 4-month releases. However, with the progress of the project, this workflow proved to be inefficient. Conflicts between the internal management processes
and differences in pace and goals of each institution were compromising the
platform development.

\subsection{Survey and data collection}

We divided the UnB development team into two groups of target participants according to
their roles during the project: \textit{UnB Interns} and \textit{Senior Developers}. For each
group, we designed an online survey with topics related to project organization,
development process, communication and relationship between members, acquired
knowledge and experience with free software. We also interviewed two \textit{MPOG
analysts} who directly interacted with the development team and project
development process. The interview questions could be classified into four
parts: Professional profile; Organization, communication and development
methodologies in the context of government and project; Satisfaction with the
developed platform; Lessons learned.

We sent the link of the online survey through emails to 42 UnB interns (undergraduate students), who participated in any time of the project
as developer receiving scholarship. We received a total of 37 responses. Their
average age is 25 years old and 91.9\% of them are male. Currently, 35.1\%
continue at university as undergraduate or graduate students, 18.9\% work as
developer in a small company and 18.9\% in medium or large companies, 10.8\% are
entrepreneurs, 8.1\% are unemployed and the others work as teachers or civil
servants. 43.2\% said the SPB project was their first experience with free
software.

We also sent the link of the online survey through emails to eight senior developers (IT market professionals). All of them
answered the questionnaire. Their average age is 32 years old and 87.5\% are
male. They have an average of 11 years of experience in the IT market, and
currently 62.5\% of respondents are company employees, 37.5\% are freelance
developers, 25\% are master's degree students and 25\% entrepreneurs. They have
worked on average in 5 companies and participated in 4 to 80 projects. They
participated in this collaborative project between 7 to 24 months. 85.7\% of
them had some experience with free software before the SPB project.

Two MPOG IT analysts were interviewed separately.
Each interview took an average of 2 hours with 28 open questions. They are more
than 30 years old and have been government employees for more than 7 years.
Only one of them continues working in the same ministry. For both, this
collaborative project was their first experience of government-academia
development collaboration.

Finally, we quantitatively analyze data about the development of the project,
publicly available on the SPB platform. We collected from the repository manager tool of the platform all open issues and commits related to the main repository of the platform, that is, the development repositories of the integrated software were not considered.
For issues, we collected:
project name, author of the issue, opening date, issue title, and number of
comments. We also collected informations about: total open issues, total
commits, different authors of issues, total of different authors of issues,
total of comments, authors of comments, total of authors other than comments. 
During the period from April 2015 to June 2016, 879 issues was opened by 59 distinct authors with a total of 4658 comments and 64 distinct commentators. The development team made 3256 commits in the repository provided by SPB platform, the first one in July 2014 and the last one in August 2016.