Re: bits per color component.

On Wednesday, May 31, 2006, 10:45:32 PM, Rotan wrote:

RH> Yes, of course. That's what I get for trying a quick copy/paste. I take
RH> it that the revised definition fits with your understanding of color
RH> properties 

Yes.

RH> and that it resolves the anomoly. Whether or not the function
RH> itself is useful is a separate issue. 

I could imagine it being useful, to send images of differing fidelity depending on the capabilities of the device.

RH> The starter functions are merely
RH> to ensure that DI-Select can build upon existing Media Queries with
RH> minimal effort. Obviously people will want to extend DI-Select with
RH> their own functions, and that's when the fun begins.

RH> Regards,
RH> ---Rotan. 

RH> -----Original Message-----
RH> From: Chris Lilley [mailto:chris@w3.org] 
RH> Sent: 31 May 2006 21:35
RH> To: Rotan Hanrahan
RH> Cc: public-diselect-editors@w3.org
RH> Subject: Re: bits per color component.

RH> On Wednesday, May 31, 2006, 10:17:15 PM, Rotan wrote:

RH>> Yes, I think we've noticed that anomoly. The starter functions in 
RH>> DISelect are merely wrappers around the Media Queries functions, so 
RH>> the interpretation should be exactly the same. I'm sure the wording 
RH>> can be clarified. With this in mind, and observing the MQ definition

RH>> [1] I think the definition should be:

RH>> 'In a device with indexed colors, the di-cssmq-color function will 
RH>> return the minimum number of bits per color component in the lookup 
RH>> table.'

RH> a 5:6:5 system is not indexed - it has no palette. Its a truecolor
RH> (well, hicolor) system.

RH> So just

RH> "the di-cssmq-color function will return the minimum number of bits per
RH> color component"


RH>> Thus, for the common case of 5 bits red, 6 bits green, 5 bits blue 
RH>> (16 bits total), "di-cssmq-color(8) > 5" would be false. (Because

RH>> "5 > 5" is false.)

RH> Yes.


RH>> -----Original Message-----
RH>> From: public-diselect-editors-request@w3.org
RH>> [mailto:public-diselect-editors-request@w3.org] On Behalf Of Chris 
RH>> Lilley
RH>> Sent: 31 May 2006 20:56
RH>> To: public-diselect-editors@w3.org
RH>> Subject: bits per color component.


RH>> Hello public-diselect-editors,

RH>> under  9.9 The di-cssmq-color function 
RH>> http://www.w3.org/TR/2005/WD-cselection-20050502/#sec-dcn-cssmq-colo
RH>> r

RH>> I read:

RH>>  "number of bits per color"

RH>> Please clarify this by changing to "number of bit per color
RH> component"
RH>> which is, I think, the assumed meaning. For example, a truecolor 
RH>> display with 8 bits of red, 8 bits of green and 8 bits of blue would

RH>> be described here as having 8 bits, not as 24 bits, so

RH>> di-cssmq-color(8) < 9

RH>> would be true.

RH>> The wording needs to be clarified because it is common to talk of 
RH>> both the number of bits in total and also the number of bits per 
RH>> component. A typical computer display might be described as a "24bit

RH>> display" and a higher quality, medical imaging display might be
RH> described as a "10 bit"
RH>> display (10 per component; 30 in total).

RH>> Also, for the common case of 5 bits red, 6 bits green, 5 bits blue 
RH>> (16 bits total) is di-cssmq-color(8) > 5 true or false?









-- 
 Chris Lilley                    mailto:chris@w3.org
 Interaction Domain Leader
 Co-Chair, W3C SVG Working Group
 W3C Graphics Activity Lead
 Co-Chair, W3C Hypertext CG

Received on Thursday, 1 June 2006 08:57:21 UTC