Week 7 - Ownership

Various Team Members’ Work Summary

Work this past week has been based around continuing work with NLP (natural language processing), security adjacencies, and the concept of OWNERship. Additionally, the majority of our team has by now read one or more papers by Chris Bird. His work has been a great resource as far as investigating existing research on code reviews.

Code OWNERship

In the past week Richard and myself have finished the OWNERS script to scrape OWNERS hierachy data from the Chromium code base using their OWNERS file formatting rules. The script runs in a few minutes and builds a CSV (comma-separated values) file containing the release number, filepath, and owner email. We’ve made a variety of improvements on the script, including adding Trollop to parse commandline arguments such as where to place the CSV and where the codebase to scrape is.

With the scraper done, Richard started to move forward on creating a model and adding a table to our database to store the data, write a parser to go through the CSV and place the data into the table, and write some verifies to confirm the data is correct. I have started to branch out and look into another OWNERS-related task - a script to traverse the Chromium code base and collect the information on when a developer is first added as an OWNER in a filepath. We plan on using this initially to determine if people are being ‘promoted’ to an OWNER too quickly. Because giving explicit OWNERship allows the person through Rietveld code review tool to approve reviews, having inexperienced OWNERS could relate to CVEs and bugs.

Overall both progress on OWNERship research and other areas of the project are going smoothly.

This project is sponsored by the Collaborative Research Experiences for Undergraduates, funded by the National Science Foundation.