W3C home > Mailing lists > Public > www-style@w3.org > July 2008

Re: [css3-color] New last call for comments on CSS Color

From: Chris Murphy <lists@colorremedies.com>
Date: Tue, 22 Jul 2008 00:31:43 -0400
Message-Id: <DF7CA94B-E0D7-4462-878D-0BC68742DD63@colorremedies.com>
To: W3C style mailing list <www-style@w3.org>, "L. David Baron" <dbaron@dbaron.org>

I'm not sure what you mean by "handling out-of range or invalid  
values." Do you mean a JPEG or text that somehow has an encoded value  
for red greater than 255? I don't see how that happens in an 8bpc  
world. There simply isn't a place holder for a value greater than 100%.

But OK, if that can happen (?), then yes instead of the browser not  
displaying the image or content, or crashing or otherwise dramatic,  
fine just say anything higher than 100% is only 100% and proceed  
normally. But I would not call this gamut mapping or clipping, this is  
error correction in reading the page document or image data.

I see in the example RGB 300,0,0. Umm, 300 levels don't exist in 8bpc.  
110%? Doesn't exist. To me this implies floating point where values  
can be greater than 100% and less than 0%.

That sRGB 255,0,0 may be out of gamut for my display, but how is this  
even knowable? You send the color to a CMS and ask it "is this out of  
gamut or not?" It can be done, but why bother? Just send the value to  
the CMS and say "display this" and it does the best job it can.  
Whether it's in or out of gamut doesn't, in my view, need to enter  
into the equation. The gamut check is a feature in Photoshop, but it's  
not particularly useful in my view, so I guess I'm just confounded why  
in and out of gamut colors is even a relevant issue for the developer  
of a web browser. Tag the data, let the CMS do the rest.

Chris Murphy



On Jul 21, 2008, at 11:59 PM, L. David Baron wrote:

> On Monday 2008-07-21 23:47 -0400, Chris Murphy wrote:
>> That whole paragraph.
>
> So the normal behavior in CSS for handling out-of-range or invalid
> values is to treat them as parse errors.  Since device gamuts vary,
> we don't want to do that here.  So I think the spec needs to say
> something so that it's clear that the colors specified should
> somehow be mapped into the device gamut rather than dropped.
>
> Is there a more general term than "mapped" that you would prefer?
>
> -David
>
> -- 
> L. David Baron                                 http://dbaron.org/
> Mozilla Corporation                       http://www.mozilla.com/
Received on Tuesday, 22 July 2008 04:41:05 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 27 April 2009 13:55:10 GMT