Making sense of 20 million emails: Introducing the LKML5Ws Dataset

Track: Artificial intelligence & Debian

Type: Short talk (20 minutes)

Room: Aula Magna - FADU

Time: Jul 23 (Thu): 16:30

Duration: 0:20

The Linux kernel uses mailing lists as the primary medium for all development, bug reporting, and pivotal discussions regarding the project’s future, a practice also observed in other large Free Software projects. However, as a consequence of this decentralized development model, these emails are spread across hundreds of mailing lists, each with its own community and code-maintenance model.

This talk presents the LKML5Ws dataset (“the What, When, Who, Where, and Why in the Linux Kernel Mailing Lists”). With over 20 million emails from 345 different mailing lists, our massive relational dataset provides a comprehensive overview of the last 20 years of Linux kernel development. Beyond shedding light on the awe-inspiring number of patches, discussions, and contributors, our dataset serves as a basis for studying the intricate, knowledge-rich nature of free software development. We want to share and discuss this data, along with some visualizations, with the Debian community.

URLs

etherpad