Firefox/Projects/Multitouch: Difference between revisions
(→Plan) |
(→Plan) |
||
| Line 31: | Line 31: | ||
=== Plan === | === Plan === | ||
* Work on our events implementation for desktop and mobile to get them close to a possible spec (following smaug's proposal?) | |||
* Get rid of the manual document.multitouchData switch and work on switching modes when the page requests any touch events | |||
** This will deactivate page scrolling and win7 gesture recognition | |||
* '''[LAND 1]''' Deliver the raw events input first | |||
*Check on using windows gesture processors or implement our own gesture engine | |||
*[LAND 2] Deliver the gesture events | |||
*Re-implement [kinetic] panning on our engine for the pages using the touch events | |||
*[LAND 3] Deliver panning on our engine | |||
* Check on using windows gesture processors or implement our own gesture engine | |||
* [LAND 2] Deliver the gesture events | <br> Note 1: for pages that will not make use of the touch apis, nothing will change, and we will continue to use window's own gesture and scrolling processors. If, after LAND 3 or later we think our own stack is good enough to replace windows's one we can make the full switch | ||
* Re-implement [kinetic] panning on our engine for the pages using the touch events | |||
* [LAND 3] Deliver panning on our engine | |||
Note 2: Most demos out there related to touch (e.g. [http://scripty2.com/demos/touch/]) never combine it with a page with scrolling, so disabling it as a first step doesn't seem to be a big problem. | Note 2: Most demos out there related to touch (e.g. [http://scripty2.com/demos/touch/]) never combine it with a page with scrolling, so disabling it as a first step doesn't seem to be a big problem. | ||
Revision as of 15:17, 2 March 2010
Overview
This project aims to improve Firefox support for touchscreen devices. These improvements include making the browser experience better for users of touch screens (on desktop and mobile) and making it possible for webpages to make use of this new form of user input.
Main bugs:
- bug 548100 Tracking bug to everything related to touch
- bug 508906 Bug to implement a multitouch API via DOM Events
UX front
Some bugs:
- bug 547997 – Perform hit target detection/correction on tap clicks
- bug 547996 – Be able to tell when a click was generated by touch
- bug 548005 – Browser zoom should stay centered when coordinates are available
- bug 548012 – Browser zoom should be cancelable to be handled by webpage
(More on the tracking bug)
Multitouch API front
- bug 508906 [Win7] bug 544614 Fennec
- At the moment, experimental patch for Windows 7 present on Bug 508906.
- Stuck in two issues:
- Waiting for a real (draft) spec to be published
- Thinking on how to deal with Windows 7 limitation that we cannot get gestures and raw data at the same time
Plan
- Work on our events implementation for desktop and mobile to get them close to a possible spec (following smaug's proposal?)
- Get rid of the manual document.multitouchData switch and work on switching modes when the page requests any touch events
- This will deactivate page scrolling and win7 gesture recognition
- [LAND 1] Deliver the raw events input first
- Check on using windows gesture processors or implement our own gesture engine
- [LAND 2] Deliver the gesture events
- Re-implement [kinetic] panning on our engine for the pages using the touch events
- [LAND 3] Deliver panning on our engine
Note 1: for pages that will not make use of the touch apis, nothing will change, and we will continue to use window's own gesture and scrolling processors. If, after LAND 3 or later we think our own stack is good enough to replace windows's one we can make the full switch
Note 2: Most demos out there related to touch (e.g. [1]) never combine it with a page with scrolling, so disabling it as a first step doesn't seem to be a big problem.