[css-color] raise issues from San Francico discussions

Author: svgeesus

css-color/Overview.bs

@@ -1591,8 +1591,9 @@ Device-independent Colors: Lab and LCH</h2>
 	</pre>
 
 	The first argument specifies the CIE Lightness, the second argument is C and the third is H.
-	L is constrained to the range [0, 100]. C is an unsigned number, while H is constrained 
-	to the range [0, 360).
+	L is constrained to the range [0, 100]. C is an unsigned number, theoretically unbounded 
+	(but in practice does not exceed 230).
+	 H is constrained to the range [0, 360).
 
 		<p class="issue">Need to define handling for out of range numbers. 
 			Clip L, wrap H? Have also seen L up to 400, specifically 
@@ -1681,10 +1682,10 @@ gamut space capable of representing almost all visible real-world colors. Both a
 	seven-ink printer, for ProPhoto RGB, and for the P3 standard RGB space. 
 
 	<pre>
-		color(swopc, 0, 206, 190, 77);
-		color(indigo, 24, 160, 86, 42, 0, 18, 31);
-		color(prophoto, 233, 150, 122);
-		color(P3, 97, 253, 36);
+		color(swopc, 0, 206 190 77);
+		color(indigo, 24 160 86 42 0 18 31);
+		color(prophoto, 233 150 122);
+		color(P3, 97 253 36);
 	</pre>
 
 	All but the predefined colorpace example also need a matching ''profile'' at-rule somewhere 
@@ -1704,23 +1705,32 @@ gamut space capable of representing almost all visible real-world colors. Both a
 
 The color function takes an indentifier as the first parameter, followed by as many 
 numerical parameters as are needed (three for an RGB colorspace, four for CMYK, 
-one for greyscale, and so on). The identifier is either one of the predefined spaces 
-(<i>P3</i> or <i>Rec2020</i>) or the name of a profile.
+one for greyscale, and so on). An optional, comma-separated alpha value is also supported.
+The identifier is either one of the predefined spaces 
+(<i>p3</i> or <i>rec2020</i>) or the name of a profile.
 
+<pre class="issue">Decided at San Francisco to add a larger set of common predefined spaces 
+	like AdobeRGB, ProPhoto RGB, and so on. Also coated and uncoated swop, etc, etc. </pre>
 
 	<pre class='prod'>
-		<dfn>color()</dfn> = color( <<ident>> , <<number>>+ )
+		<dfn>color()</dfn> = color( <<ident>> , <<number>>+ [, <<number>>])
 	</pre>
 
 	<p class="issue">How to handle cases where there are too many, or too few, numbers supplied?
 		Seems like ignoring extra numbers, and treating unspecified numbers ar zero, would be 
 		the most robust?</p>
+		
+	<p class="issue">Need a fallback color, as decided at San Francisco. Add after the alpha value.
+		Can be an sRGB fallback or indeed another color() function.</p>
+		
+	<p class="issue">Need to add choice of quoted string or number+, to allow f
+		or named color profiles.</p>
 
 <h3 id="predefined">Predefined colorspaces: DCI P3 and Rec.2020.</h3>
 
-These two colorspaces are indicated by using the predefined identifiers <i>P3</i> or 
-<i>Rec.2020</i> in the color function. No ''<@color-profile>'' at-rule is needed; if one is 
-supplied, it will be ignored.
+These two colorspaces are indicated by using the predefined identifiers <i>p3</i> or 
+<i>rec2020</i> in the color function. No ''<@color-profile>'' at-rule is needed for these
+predefined names; if one is supplied, it will be ignored.
 
 The DCI P3 [[!DCI-P3]] colorspace has the following characteristics:
 
@@ -1764,14 +1774,14 @@ although in practice all but the first step are linear calculations and can be c
 <!-- make these steps links to the appropriate section in a calculations appendix -->
 <ol><li>Convert from gamma-corrected RGB to linear-light RGB (undo gamma correction)
 	<li>Convert from linear RGB to CIE XYZ
-	<li>Convert from a D65 whitepoint (used by both P3 and Rec.2020) to the D50 whitepoint used 
+	<li>Convert from a D65 whitepoint (used by both p3 and rec2020) to the D50 whitepoint used 
 		in Lab, with the Bradford transform
 	<li>Convert D50-adapted XYZ to Lab
 </ol>
 
 <h4 id="lab-to-predefined">Converting Lab to predefined colorspaces</h4>
 
-Conversion from Lab to P3 or Rec.2020 also requires multiple steps, and again in practice all 
+Conversion from Lab to p3 or rec2020 also requires multiple steps, and again in practice all 
 but the last step are linear calculations and can be combined.
 
 <ol>
@@ -1799,10 +1809,18 @@ data (<dfn data-dfn-type="descriptor" for="@color-profile/src">src</dfn>, just l
 ''src'' in font-face.) The ''src''
 descriptor takes a url() as it's value.
 
+<p class="issue">Same-origin and CORS for src.</p>
+
+<p class="issue">local() to use locally installed profiles. Profile stack 
+	like font-face rather than a single url. Avoid flash of uncalibrated color.</p>
+
 <!-- should we add a format field to this descriptor, like with font-face src? it is always ICC format
 in practice, but that could change in the future. or we could add format later, once there is
 a need, and make the default ICC if not specified -->
 
+<!-- we decided in San Francisco to not allow renering intent to be specified in the CSS; 
+just use the rendering intent provided by the profile.
+
 A third, optional descriptor defines the <dfn data-dfn-type="descriptor" for="@color-profile">rendering intent</dfn> to use when mapping colors 
 from a larger to a smaller gamut.
 
@@ -1878,13 +1896,21 @@ There are four values for rendering intent
 	pixels. Out of gamut colors should be converted to colors that have the same saturation 
 	but fall just inside the gamut.</p> </dd>
 </dl>
-	
-
+	-->
 
 <pre class="prod">
 	<dfn at-rule >FIX ME: MISSING</dfn>
 </pre>
 
+<h2 id="working-color-space">Working Color Space</h2>
+
+<p class="issue">Resolved at San Francisco to add a working-color-space at-rule, which 
+	affects the entire document. Compositing, interpolation, blending use this. 
+	Initial value is sRGB. linear-sRGB, p3, rec2020, and lab were also discussed as values.
+	Chris to read the canvas spec to see what it does there, particularly for the 
+	'optimal' value.</p>
+	
+
 <h2 id='cmyk-colors'>
 Device-dependent CMYK Colors: the ''device-cmyk()'' function</h2>