Re: [Vibration] Feedback on the Vibration API

On Wednesday, November 7, 2012 at 12:12 PM, Anssi Kostiainen wrote:

> [+to Justin]
>  
> Hi David, Marcos, Justin,
>  
> On 6.11.2012, at 15.22, ext Marcos Caceres wrote:
>  
> > On Tuesday, November 6, 2012 at 10:07 AM, David Bruant wrote:
> >  
> > > First, in the spec, it seems that only one app can vibrate at a time
>  
> As per the step 8, only one application can vibrate at a time. We landed on this design based on Mozilla's implementation experience, see:
>  
> https://bugzil.la/679966
>  
> > If this is so, this is a spec bug (or just not correctly worded). I'm fairly certain that notifications can cause the phone to vibrate independent of application (e.g., when you get an SMS).
>  
> I think you're right in that trusted (or privileged apps) should be able to vibrate the device even if not visible. Thanks for pointing this out!
>  
> To make this clearer, I propose we add something like the following as a note after the step 6 (feel free to propose a better wording):
>  
> [[
>  
> Note
>  
> A trusted (also known as privileged) application that integrates closely with the operating system's functionality may vibrate the device even if such an application is not visible at all, and thus may ignore the previous step.
>  
> ]]  
>  
> Alternatively, we could bake similar language in to the algorithm itself. As the algorithm is currently written, we're bailing out at step 6 if the privileged app invoking vibrate() is hidden.
>  
> Justin, David, Marcos - does that change make sense to you?
>  
> For reference, here's the current algorithm:
>  
> http://dev.w3.org/2009/dap/vibration/#dfn-processing-vibration-patterns
>  
I don't have a strong personal opinion… though I do worry about the spec becoming overly prescriptive. It's just a fact of life that other apps running in the background may cause the device to vibrate (e.g., a notification center). There is probably not much a browser can do about that (beyond controlling it's own tabs) so putting prose in there about interaction with the underlying OS might cause confusion (specially if it's too mixed in with the algorithm).    

--  
Marcos Caceres

Received on Wednesday, 7 November 2012 12:51:33 UTC