Re: Handling device permissions

On 4/17/12 8:19 AM, Timothy B. Terriberry wrote:
> Josh Soref wrote:
>>> There has been no discussion that I can tell that describes how and
>>> if permissions should be handled or stored by user agents.
>>
>> See above.
>
> I think the question is, should there be an API to allow the app to
> know, before calling getUserMedia(), if such a call will require a
> prompt for user consent? (N.B. that this is different from knowing such
> a call _won't_ require a prompt.) Applications may want to have
> different behavior depending on the answer to this question.
>
> The only simple example I can think of is to stop asking the user if
> they've said no once, and that's easy enough for the UA to enforce by
> itself (with a "Never allow this site to access the camera" option), or
> by the site simply remembering the result of its first attempt. Are
> there any other good examples that would require such an API?

There might be, depending on whether or not we add an API to reserve 
resources, or at the least allow the webapp to express the intent that 
it will call getUserMedia at a later point.

If we roll both of these into getUserMedia, then an app might choose to 
ask for resources without user interaction on startup, but if a prompt 
is required, defer it until is actually needs the camera.

-Anant

Received on Friday, 27 April 2012 21:25:22 UTC