Firefox/Projects/Perception of Performance: Difference between revisions
Jump to navigation
Jump to search
No edit summary |
|||
| Line 13: | Line 13: | ||
* Actually improve real performance :) | * Actually improve real performance :) | ||
= Current Progress= | = Current Progress = | ||
* Margaret has an initial implementation of an acceleration-based scroll model working on Windows | *Currently the biggest concern is users who encounter double acceleration ([https://bugzilla.mozilla.org/show_bug.cgi?id=509651 bug 509651]), we need to also buy some mice that include their own driver with an acceleration model so that we can get a sense of what double acceleration feels like<br> | ||
* The two thresholds we are now interested in experimenting with and tweaking are the: | *Margaret has an initial implementation of an acceleration-based scroll model working on Windows | ||
** rate of acceleration | *The two thresholds we are now interested in experimenting with and tweaking are the: | ||
** time chunk of events to analyze before processing them | **rate of acceleration | ||
* Still looking for research and documentation in this general area of UI toolkit design | **time chunk of events to analyze before processing them | ||
* Two patches are up for review that add the functionality described below | *Still looking for research and documentation in this general area of UI toolkit design | ||
*Two patches are up for review that add the functionality described below | |||
= Design: Acceleration Model for Scrolling = | = Design: Acceleration Model for Scrolling = | ||
Revision as of 04:50, 11 August 2009
Overview
Sprint lead: faaborg
Sprinters: faaborg, margaret
Goals / Use Cases
- Identify top 5 ways to improve perceived performance
- Full list of brainstormed ideas I need to organize this list, grouping items into categories, and identifying the top 5 items.
- Resolve one on trunk (we selected acceleration model for scrolling)
Non Goals
- Actually improve real performance :)
Current Progress
- Currently the biggest concern is users who encounter double acceleration (bug 509651), we need to also buy some mice that include their own driver with an acceleration model so that we can get a sense of what double acceleration feels like
- Margaret has an initial implementation of an acceleration-based scroll model working on Windows
- The two thresholds we are now interested in experimenting with and tweaking are the:
- rate of acceleration
- time chunk of events to analyze before processing them
- Still looking for research and documentation in this general area of UI toolkit design
- Two patches are up for review that add the functionality described below
Design: Acceleration Model for Scrolling
We're using a timeout function to group mousewheel clicks into separate series, where each series represents a unique scroll gesture. Within each series we increment a counter that keeps track of the number of clicks of the scroll wheel within the series, and we can use this value to determine when to start acceleration and how much acceleration to create.
We're using three integer preferences to modify scroll wheel behavior:
- mousewheel.withnokey.numlines (default=6)
- The number of lines the page moves with one click of the scroll wheel.
- This pref is used to determine a scroll delta before acceleration computations are applied.
- mousewheel.acceleration.start (default=3)
- The scroll wheel click number at which acceleration begins to take effect.
- mousewheel.acceleration.factor (default=10)
- The multiplicative factor used to scale constant acceleration.
Bugs
- bug 462809 - Interpretation of scroll events on Windows and OS X
- bug 508747 - mousewheel.withnokey.numlines overrides the OS setting
- bug 508785 - Extremely high speed of mouse-wheel scrolling
- bug 509189 - The application level mouse wheel scrolling acceleration should be disabled in default settings
- bug 509651 - Disable Firefox's acceleration based scrolling model if the mouse driver already provides it