"Our data source is constituted by e-mails sent to KDE mailing-lists and archived by MARC"
"Two problems quickly arise: neither the e-mails addresses nor the names can be considered unique. Consequently, we used an in-depth search algorithm to put together “name-email” couples corresponding to a same contributor. Indeed, the algorithm suggests possible merges."
"There is a specific mailing list in our data set, kde-commit, which gathers automatic notifications from the revision control system (RCS)....We measure “commit” by the number of messages sent to the “kde-commit” mailing list. However, we did not count “silent” commits, nor usual messages sent to this mailing list."
"We measured activities done in BTS in two ways: “bug opener” and “non bug opener”. First, we counted the number of modifications done by the contributor who opened the concerned bug report. "