How to get away with refactoring

How do you refactor major, core functionality in a million-line codebase without disrupting the entire system? Maude Lemaire explains how Slack overhauled channels and shares the many obstacles the company overcame to boost both application performance and company-wide developer productivity (with only a few hiccups).

Slack is the leading global collaboration hub that makes people’s working lives simpler, more pleasant, and more productive. From global Fortune 100 companies to corner markets, businesses and teams of every kind use Slack to bring the right people together with all the right information. Just two years ago, the product only had 2 million daily active users (DAUs); today it boasts over 6 million. Unfortunately, with quick growth come tough growing pains, and Slack wasn’t (and still isn’t) immune to them. Around April of last year, the product’s largest customers were plagued with relentless performance-related outages. Instead of shipping scrappy, short-term fixes in the hopes that Slack could survive another Monday morning, the company opted to address the root problem and completely refactor problematic functionality. Maude Lemaire tells the tale of that undertaking, following one engineer (her) through many obstacles and hiccups she overcame to boost both application performance and company-wide developer productivity. Through Maude’s experience, you’ll learn how to properly assess the risks associated with a large refactoring effort, narrow the project scope to only its necessary components, convince your leadership that the intimidating, months-long rewrite is worth it, document the complex process for teammates and leadership alike, prepare for the unexpected, and stay optimistic.