W3C home > Mailing lists > Public > public-media-capture@w3.org > April 2013

Re: Overconstrained

From: Harald Alvestrand <harald@alvestrand.no>
Date: Mon, 01 Apr 2013 11:04:34 +0200
Message-ID: <51594DA2.5020708@alvestrand.no>
To: public-media-capture@w3.org
On 03/28/2013 07:37 AM, Adam Bergkvist wrote:
> I pretty much agree with everything you're saying here. Together with 
> the media types, "camera facing" is the single most important 
> constraint. I think you bring up an important factor with the OS 
> default selection as well.
>
> On the topic of solving the few remaining cases, I just listed a few 
> ways to get there. One that simply could be implemented in JS (ask 
> again), another that would require some minor implementation (Harald's 
> UI proposal), and a last one that would be us inventing something that 
> requires a bit more implementation work. I think our current approach 
> with constraints belongs to the third category.
>
> So what do we really need? I would say the options to select "media 
> type" and "camera facing". Let's have Harald's grant all cameras 
> approach as well, and we can always implement the "ask again" approach 
> in JavaScript. Should "media type" and "camera facing" be part of the 
> constraints/settings concept? Perhaps not. It's not likely that you 
> would like to change either media type or camera facing with the 
> constraints/settings API later. We could just have a simple structure 
> to represent these.

Note that I did not want to change at all the mechanics of camera 
*selection*.
I wanted to extend the semantics of camera *permission*.

The two concepts are different, they just happen to be both mediated by 
GetUserMedia at present.

(Whether that is a design mistake is a different question.)
Received on Monday, 1 April 2013 09:05:05 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 21:24:40 UTC