Software Público Brasileiro

\section{Related work}
\label{sec:relatedwork}

Discussions on how to introduce new management methods into an organization are
present in several works.  Nerur et al. recognized critical issues concerning
the migration from traditional to agile software development by comparing
practices of both methodologies \cite{nerur2015challenges}. The authors point
out managerial, organizational, people, process, and technological issues to be
rethought and reconfigured in an organization for a successful migration.
Strode et al. investigated the relationship between the adoption of agile
methodologies and organizational culture \cite{impactOfOrganizationalCulture}
by evaluating nine projects. They identified a set of six factors directly
linked to agile methods and concluded that the presence of these aspects in an
organization is proportional to the value of agile methodologies usage for
their projects. As Nerur et al., Strode et al. also said that the adoption of
agile development techniques does indeed produce changes in an organization's
culture.

Some works also discuss how academia can collaborate with the industry in the
management of software projects. Chookittikul et al. evaluated the increasing
use of the agile techniques in software development companies in Thailand.  The
authors suggested that universities should create curricula that develop in
their undergraduate students practical skills required by industry (mainly
agile practices) to promote growth in local software businesses
\cite{cho2011gap}. Sandberg et al. report the use of Scrum in an
industry-academia research consortium (involving ten industry partners and five
universities in Sweden) \cite{sandberg2017iacollaboration}. Through a case
study, they demonstrate that being able to bring together the meaningful
activities of the stakeholders is essential to the success of collaborative
research between industry and academia.

Complex and large-scale organizations, such as the public administration, have
to deal with multiple project variables. Alleman et al. describe a production
deployment for the US government, focusing on the methodology applied to
address long-term planning and value estimation \cite{alleman2003making}. In
the Brazilian context, Melo et al. \cite{melo2013agileBr} investigates the
growing adoption of agile methodologies in this country's IT industry. The
results of their survey highlight some mismatch that companies faces when
developing software for public administration.

Several works tried to highlight the FLOSS practices, while others attempted to
determine the relationship between FLOSS practices and agile methods. Capiluppi
et al. examined about 400 projects to find FLOSS project properties
\cite{capiluppi}. In their work, they extracted generic characterization
(project size, age, license, and programming language), analyzed the average
number of people involved in the project, the community of users, and
documentation characteristics. Warsta et al. found differences and similarities
between agile development and FLOSS practices \cite{warsta}. The authors argued
that FLOSS development may differ from agile in their philosophical and
economic perspectives, on the other hand, both share the definition of work.
Finally, Eric Raymond describes many of his experiences and decisions in his
work with FLOSS communities \cite{raymond}, and his report in 1999 has many
intersections with the agile manifesto in 2001.

This paper differs itself from others by studying the government-academia
collaboration for developing a production-level solution. From questionnaires,
interviews, and development activities data, we extracted best practices that
helped to harmonize the interactions between two different development process
and satisfied the management process of both sides. We analyzed the decisions
made from the FLOSS and agile perspectives.