Re: [mediaqueries] status and moving forward

> On Feb 29, 2016, at 00:42, fantasai <fantasai.lists@inkedblade.net> wrote:
> 
> On 02/26/2016 02:04 PM, Tab Atkins Jr. wrote:
>> On Wed, Feb 24, 2016 at 8:46 PM, fantasai <fantasai.lists@inkedblade.net> wrote:
>>> On 02/09/2016 08:17 PM, Tab Atkins Jr. wrote:
>>>> What problem do you have with light-level?  Aside from the "should we
>>>> also map a11y concerns to this", I think the feature itself is 100%
>>>> stable and well-designed.
>>> 
>>> That's exactly my concern. :)
>>>   https://lists.w3.org/Archives/Public/www-style/2016Feb/0295.html
>> 
>> That's not a problem with light-level in any way.  At most, it's a
>> question of "should we do this future MQ, or just recommend using
>> 'light-level' for that purpose".  a11y concerns won't change our
>> design of 'light-level' at all - it does exactly what it's supposed
>> to.
> 
> I agree that the design of 'light-level' shouldn't change. I don't
> expect it to. I do expect that whether it should be masquerading
> as an a11y MQ is likely to change based on further thought on those.
> 
> Imho, light-level and contrast and bg/fg preferences are three
> different things. light-level handling will often be ONE OF
>  a) contrast changes
>  b) bg/fg inversion
>  c) both a) and b)
> and which one of these three options the author takes is IMHO
> not something we or the OS should be assuming. E.g. I think it's
> perfectly reasonable to have a light-on-dark high contrast design
> for dim lighting. Telling UAs that triggering 'light-level: dim'
> for low-contrast users is not going to serve them well on my website.

It would serve users with dyslexia somewhat. Lower contrast would likely be better, but light-on-dark is likely to be an improvement.

While I believe that in most cases this will be an improvement, and only extremely rarely will it actually make things worse, whether or not this is true does vary on a site per site basis. Would you be more comfortable with the a11y paragraph if we called that out, and encouraged UAs that wish to expose this as a preference for a11y reasons to have the ability to switch on a per site basis?

> I can live with light-level shipping as-is if we remove the a11y-related
> suggestions, if you feel this is critical.

I understand that having specialized MQs for a11y, and having authors use them, would lead to better results than using a vague proxy. But until we have these a11y MQs — and even after we do since most authors cannot be expected to tailor their stylesheets for audiences they are barely aware of — I fail to see the harm in calling out the fact that light-level is a half-way decent approximation of these a11y concerns.

> However, I would *prefer* if we delayed it to L5 so that we can release
> 'light-level', 'preferred-contrast', and 'preferred-fgbg' *together*.
> This lets us set up authoring best practices (through spec recommendations
> and examples, and hopefully any tutorials based on them) to encourage
> anyone handling light-level to also opt in for the relevant a11y queries.

We can (and should) still write these best practices when we do L5, but the fact that light-level works out (even if only approximatively) without any best practice guidelines is a big part of the appeal.

 - Florian

Received on Monday, 29 February 2016 08:20:01 UTC