There is no good grasp of the performance of the visual, 2017, and 2010 editors relative to each other. The visual and 2017 editors have bad worst-case performance against the 2010 editor as evidenced by slow load times on exceptionally large articles, but the average-case performance is better against some performance metrics such as time to first meaningful paint and time taken to become interactive.
In order to understand and improve the performance of these editors, it is a Q2 goal of the Editing team to carry out rigorous performance analyses of the editors. Multiple factors must be considered, such as:
- best-, average-, and worst-case performance
- performance on fast and slow machines
- performance on fast, slow, and lossy connections
- evaluation against multiple different performance metrics
From this, benchmarks can be set, and actionable performance improvements can be identified.