Archetypal Internet-Scale Source Code Searching - Finding Source Code on the Web for Remix and Reuse - page 49

3.6.1 Documentation
An initial assessment is done using web pages and documentation, as one user put
it, “...(the coredeveloper)had a good documentation of his code with lots of com-
ments too (by which I could also modify his code), hence I decided to use that
code.”
A piece of software with the required functionality may be eliminated if it can't
be easily determined whether it has the required features and documentation. The
basic functionality has to be in place, and supported by requisite documentation.
3.6.2 Peer Recommendations
Peer recommendations were the most trusted and valued - especially if the person
has used the software before. For instance, one respondent stated that “...friends
recommended Graph.pm; searched for that on search.cpan.org, and found it was
just what I needed.” In the absence or inaccessibility of peer advice programmers
then look for availability of help from people within their online social network, or
within the project context.
3.6.3 Help from One's Social Network
Help from a local expert, an electronic forum, a mailing list archive, or active
users who are doing similar tasks and are willing to answer questions is a major
consideration while choosing open source software. One respondent said he looked
for“...issues whicharethen found by people, solved and posted on mailing lists of
discussion forums.” A glance at the forums tells the users how friendly a project is
and how likely they are to obtain help when needed.
Availability of help is also determined from the project activity. Respondents
preferred large open source projects that were very active. For example, “The cri-
teria that I used were: (1) if the tool was in java (2) if the tool was web based (3)
the activity of the project.” Activity can be quantified as the number of contribu-
tors, frequency of builds and updates, traffic on newsgroups, and number of users.
Larger projects have more resources, are more responsive, and are more likely to
rank highly on these criteria.
Overall, social characteristics of the project, such as the level of activity, presence
of discussion forums, and recommendations by peers seem to have precedence over
characteristics of the source code, such as code quality (i.e. whether the code is peer
reviewed and tested), and reliability (Table 3.3 ).