Re: Screen Orientation Status

On Tue, 8 Apr 2014, at 9:08, Bruno Racineux wrote:
> On https://www.w3.org/Bugs/Public/show_bug.cgi?id=25088 :
> 
> You cannot fire on window without also having the window.orientation
> property (with its own issues*). That would break existing code relying
> on
> the webkit api (expecting to read a value on window.orientation).

I understand your concerns. Blink is implementing the legacy Webkit
window.orientation API so we will definitely send feedback here, and the
specification will be modified accordingly if we can't implement this
wrt backward compatibility.

> I am not sure, I was able to get my point across previously, for lack of
> feedback on it, but I strongly believe that mixing 'device' orientation
> and 'screen' orientation is a mistake going forward and important point.
> 
> In the not so distant future, we will likely use screens 'attached'
> to mobile devices. In that, the screen will be entirely independent from
> the mobile table/phone/watch/etc.
> 
> Sample use case:
> Let's imagine an ipad like 'iScreen' plugged to an iPad with dual screen
> capability with two different browser viewport on each screen.  When the
> iPad rotates, it shall only fire for the iPad screen, not the external
> iScreen if the later doesn't rotate, and vice-versa.
> 
> How will you differentiate those two from an implementation standpoint?
> I have no idea how that goes, but I'd like to raise that question.

I don't think the specification should handle multi-screen scenarios
explicitly given that window.screen doesn't and the Web usually doesn't
too. However, we should make sure that we actually match window.screen
so I will change the specification to make sure the screen is the
"output device". That way, we should get consistency between the values
of window.screen and window.screen.orientation.

Note that if you are interested in the scenarios you are mentioning, you
might want to have a look at the webscreens CG.

-- Mounir

Received on Thursday, 10 April 2014 16:54:50 UTC