Re: [cssom-view] colorDepth/pixelDepth, match implementations or theoretical purity?

On 6/20/13 12:07 AM, L. David Baron wrote:
> On Wednesday 2013-06-19 14:49 -0700, Tab Atkins Jr. wrote:
>> These features are basically useless.  Nearly every device that
>> matters has 24/8 color, and even for those that have 30/2 or something
>> more exotic, you need the entire browser and rendering pipeline to
>> support it before it makes sense to return the value to authors.
>>
>> I support speccing colorDepth/pixelDepth as a static 24 for now, and
>> separately, checking usage to see if we can just drop it.  I support
>> *not* implementing alphaDepth at all, for the reasons stated above.
I made them return 24 and didn't add alphaDepth. Feel free to 
investigate whether dropping them would break anything. :-)
> I'm skeptical about hard-coding it to 24.  I think a bunch of mobile
> devices these days are still using 16-bit color (though I'm not
> sure).  We also have media queries for roughly the same information.
Are there any pages that check for 16 do something useful with it?
> I tend to think if it's something that authors never care about we
> wouldn't have added it to two different properties on the screen
> object and to media queries.  Though maybe I'm wrong.
See https://www.w3.org/Bugs/Public/show_bug.cgi?id=17522#c7
> Otherwise, though, I agree; I don't see the rationale for adding
> any new features here.
OK.

On 6/20/13 1:44 AM, Robert O'Callahan wrote:
> The Screen object properties are very old. I think it's very plausible 
> they were added for no good reason. We did things like that back then 
> :-). MQ less so, but still pretty old. 
Yeah. It's not clear to me that there is a strong use case for exposing 
this information.

I've filed https://www.w3.org/Bugs/Public/show_bug.cgi?id=22415

-- 
Simon Pieters
Opera Software

Received on Thursday, 20 June 2013 09:05:10 UTC