- From: Lars Erik Bolstad <lbolstad@opera.com>
- Date: Thu, 08 Mar 2012 14:54:36 +0100
- To: James Graham <jgraham@opera.com>
- CC: public-geolocation@w3.org
James, Sorry for the rather late response here. We discussed this at our f2f meeting earlier this week. We're about to request transition to CR for the DeviceOrientation Events spec and from a purely process-related view we're concerned about introducing a dependency to DOM4 at this point. In the interest of wrapping up the work we're chartered to do as a working group now and not in 9-12 months, we decided to keep the reference to DOM2 Events. Lars Erik On 02.12.2011 14:09, James Graham wrote: > There are a number of problems around the events in the device > orientation spec: > > The spec refers to DOM 2 events which is hideously out of date. It > should refer to DOM 4 [1] which actually has an accurate description of > how events work. > > The spec introduces a new init*Event method. There has been general > agreement that these should be phased out of the platform in favour of > constructors for event objects. > > The description of event dispatch should use the correct terminology to > integrate with the event loop in HTML5. See e.g. this example [2] from > websockets to get an idea of how it should look (the relevant sentences > are those following the pattern "queue a task named X to fire a simple > event named Y at the Z object". > > [1] http://dvcs.w3.org/hg/domcore/raw-file/tip/Overview.html#events > [2] http://dev.w3.org/html5/websockets/#feedback-from-the-protocol
Received on Thursday, 8 March 2012 13:55:06 UTC