Chromium Accessibility Performance
- Mentors
- Pallavi, Aaron Leventhal
- Organization
- Chromium
- Technologies
- javascript, c++, html, ARIA
- Topics
- web, performance, accessibility, memory, stability, code health, threading
The accessibility code in Chromium is used in Chrome and Edge to enable users of assistive technology to access the web. However, enabling it can significantly impact the performance of web applications, memory usage, battery life, and crash frequency. Many users get this code enabled without realizing it through various utilities that request platform accessibility APIs. This project aims to improve Chromium's performance by optimizing its accessibility code, which can benefit the web browsing experience for a significant number of users.
With the goal of improving performance and code health, the project will complete three or more sub-projects, including bounding box computation improvement, avoiding unnecessary computations using AXMode flags, optimizing expensive property computations, and moving Accessibility deserialization off the main thread.
The project will also potentially investigate performance problems that occur when large subtrees have display: none toggled on them and identify the slowest properties computed for the AX tree to speed up or defer processing.