- From: Andrei Popescu <andreip@google.com>
- Date: Tue, 12 Jul 2011 19:37:46 +0100
- To: Bjoern Hoehrmann <derhoermi@gmx.net>
- Cc: "Wojciech Mas?owski" <wmaslowski@opera.com>, public-geolocation <public-geolocation@w3.org>
On Tue, Jul 12, 2011 at 4:34 PM, Bjoern Hoehrmann <derhoermi@gmx.net> wrote: > * wmaslowski@opera.com wrote: >>Theoretically yes. In practice I don't think this is a problem. Unless >>your application does something like counting how many times the >>orientation event fired this shouldn't really change the way your >>application works. Note that significant change is vague enough so that >>the applications should not make any assumptions about timing at which >>they receive event therefore it is reasonable to assume that receiving >>one additional event should not cause the application to behave any >>different. > > If someone makes a game where players have to click a cow while holding > their device steady, they might trigger game over when the listener is > invoked, test, find that does not work because it always gets invoked > shortly after registration, add a check for whether it's the first time, > test, find that works okay. Later they add new functionality not keeping > the odd semantics in mind, someone will sooner or later encounter issues > with that, and whether that is the developer when testing, or users who > complain, is anyone's best guess. > If all you need to know is whether the device is still or not, wouldn't it be easier to use the DeviceMotion Event? I think that would get around the need to special-case the first event in your usecase. Thanks, Andrei
Received on Tuesday, 12 July 2011 18:38:20 UTC