Original Post: Keep it simple, Scanner | Blog
Tom Shelton-Lefley discusses his recent work on tidying up the Burp Suite codebase, particularly focusing on the crawler component of Burp Scanner. Despite having a history of deleting more lines of code than adding, he has spent the last three weeks removing complex features, such as code for merging graph nodes, which went into hiding intricate complexities of the crawler’s operation.
Previously, Burp Scanner’s abstract view created merges based on unstable location features, resulting in complex collapses of directed graphs. Evaluating these merges was hard and added cognitive load for new team members and users. With new developments making crawler behavior visible to users, the old merge functionality was deemed too complex, insufficiently intuitive, and largely unnecessary. Data showed that 96% of recent crawls had zero merges.
Consequently, Tom removed merging to produce a more understandable and predictable crawler, reduce unnecessary dependencies, and simplify future development. The effort aligns with their principle: if it doesn’t benefit developers, users, or performance, it can be eliminated. He also hints at exciting new features enabled by these clean-ups.
Go here to read the Original Post