- From: Florian Rivoal <florian@rivoal.net>
- Date: Mon, 10 Sep 2012 23:07:28 +0200
- To: "Tab Atkins Jr." <jackalmage@gmail.com>, "Doug Turner" <doug.turner@gmail.com>, François REMY <fremycompany_pub@yahoo.fr>
- Cc: "Marcos Caceres" <w3c@marcosc.com>, "Doug Turner" <dougt@mozilla.com>, "CSS WG" <www-style@w3.org>, public-device-apis@w3.org
On Mon, 10 Sep 2012 22:43:56 +0200, François REMY <fremycompany_pub@yahoo.fr> wrote: >> It looks like they have to be, yes. For clarity I'd probably define, >> in a clearly-marked informative section, what lux values the ranges >> map to in an ideal sensor, but have a strong note that UAs should >> (RFC2119 SHOULD) adjust the ranges to respect the device's own >> sensitivity, so as to achieve the same results. > > The whole point is that the 'dim' and 'bright' modes depends of the > screen's minimal/maximal brightness and reflectance, too. > > While an approximative settings is important, ultimately, switching to a > high-contrast layout (either black on white or white on black) may > depends on more than just the value in lux returned by the ALS. As such, > I agree with Tab that the specification should advertise the fact that > the range in which a particular mode is triggered is completely to the > discretion of the user agent, which will do so in order to provide the > best experience possible to the user. This seems indeed to be our consensus position, and I agree with it. There remains a question though, which is to know if we can merge the functionality described above with the high contrast mode(s) that are desirable for accessibility reasons. I can add what we've agreed to so far to MQ4, and put a note or an issue about the possible overlap with accessibility. Is that fine with everyone, or do you want to discuss this aspect of the question further before anything gets specified? - Florian
Received on Monday, 10 September 2012 21:07:55 UTC