Developing Software with Open Source Software Components - Finding Source Code on the Web for Remix and Reuse - page 176

9.4.2 RQ2: How Is OSS Related Knowledge Managed
in the Industrial Practice?
Although all integrators recognized the importance of exploiting experience and
knowledge sharing to reduce learning costs and minimize risks, only five projects
(i.e. projects a, b, k, l, q) out of 19 had established (either human or computer sup-
ported) mechanisms inside their company to support the selection of OSS com-
ponents. These mechanisms might use dedicated knowledge management systems
to help store and locate usable knowledge. They also might inform about “gurus”
for providing further information, as well as experience for a coupled to distribu-
tion lists and “wikis”. In addition, some of these respondents also stated that their
company had recently set up a dedicated department or person for gathering and
monitoring the OSS marketplace.
The other 14 respondents recognized that there were not established mechanisms
inside their company to support the reuse of knowledge and experience. Instead,
they proceeded informally by directly asking information from internal and/or ex-
ternal colleagues.
In both cases, navigating on the Internet was stated as an important mechanism
for knowledge extraction and management, especially because of the existence of
Internet resources such as collaborative communities (e.g., TheServerSide, Experts
Exchange, Java Users group) that offer forums and other collaborative mechanisms
to exchange ideas and solutions to common problems.
9.5 Discussion of Main Findings
The previous section aimed to present a general view of the results. This section
aims at discussing the most important findings and observations from the data.
9.5.1 Use of Informal Procedures for Selecting Components
We found that the component searching phase was informally performed and mainly
influenced by previous experience in all the companies. This result is in line with
the claim that component selection methods proposed in the literature mostly focus
on the component evaluation phase, setting aside the problem of identifying com-
ponents and related information [ 7 ]. Regarding component evaluation, it was very
interesting to see that in contrast to previous studies that stated that companies nei-
ther used nor knew of any formalized methods to select components [ 14 , 32 , 34 , 44 ];
our results suggest that there is an incipient interest or awareness of some compo-
nent selection methods , as some companies intend to apply evaluation methods and
tools for supporting evaluation. This was mainly motivated by the need to succeed