- From: Mounir Lamouri <mounir@lamouri.fr>
- Date: Fri, 06 Jun 2014 19:46:17 +1000
- To: frederick.hirsch@nokia.com, public-device-apis@w3.org
On Fri, 6 Jun 2014, at 0:10, frederick.hirsch@nokia.com wrote: > We’ve had some discussion on the list regarding the Vibration API and > alignment of implementations. > > The only item of concern at this point appears to be step #5 [1]: > > “5. If the length of pattern is even and is not zero, then remove the > last entry in pattern." > > Anssi provided an example to highlight the issue, as well as a pointer to > chromium code [2] > > To reiterate, if the pattern is even, then the last item is a pause, > which should have no effect since there is no vibrate time following. In > addition, as Mounir points out, a subsequent API invocation will cancel > the previous one so the pause should have no effect. it is generally > agreed that this step is not testable. > > Thus I propose the following change: > > [[ > > Remove step 5. > > In its place add a Note “Note: if the length of the pattern is even and > not zero then the last entry in the pattern will have no effect so an > implementation can remove it from the pattern.” > > ]] > > I suggest we stay away from RFC normative language for step 5 (e.g. MAY) > given the testability concern. > > Given we are changing normative language, I suggest we have another Last > Call (currently Vibration is in LC), and publish this along with the > other LC specifications (after a CfC) > > Am I missing any other implementation alignment concerns? > > Please comment on list, otherwise can the editors go ahead and update the > editors draft so we can see this in place (along with an update to the > status section)? The change sounds good to me. Hopefully, Firefox Android and Chrome Android will both pass the test suite now and we will be able to move forward. Exciting! :) -- Mounir
Received on Friday, 6 June 2014 09:46:49 UTC