Re: color names in SVG-1.0 conflict with /usr/lib/X11/rgb.txt

On Tuesday, April 23, 2002, 7:29:28 PM, John wrote:

JE> I'm not sure who is "right" but there are a few conflicts
JE> between the color values in SVG-1.0 and in X11 as represented
JE> in /usr/lib/X11/rgb.txt from XFree86.

JE> SVG colors are descibed in <http://www.w3.org/TR/SVG/types.html>

JE> Specifically I found the following conflicts:

JE>                   SVG-1.0           X11

JE>         gray   128 128 128       190 190 190
JE>         grey   128 128 128       190 190 190
JE>         green    0 128   0         0 255   0
JE>         maroon 128   0   0       176  48  96
JE>         purple 128   0 128       160  32 240


JE> The most serious of these conflicts, it seems to me, is
JE> the meaning of "green" since green is a primary color
JE> in the RGB color space.

The color name 'green' has been defined as a half-intensity green
primary since the very early days of HTML 3.2, then HTML 4, CSS1,
CSS2, and finally XSL and SVG. The color name for the full intensity
green primary is 'lime'. Similarly for maroon, the half-intensity red
primary.

These values have been implemented for very many years in existing
browsers before being standardized. You will find these values in
Netscape 1, MSIE 2, and Opera 1.

The values are also colors, as opposed to digital values. By that I
mean that they are tied to a particular color space of monitor
phosphors, transfer curve, and viewing conditions. They are colors in
that they are objectively measurable using international standards
for color which go back to 1931.

Thus the green primary, in terms of W3C colors, has a specific
measurable value. The green primary in X11.txt means whatever the
current hardware says that it means.

JE> Can you advise on how these conflicts should be resolved?

The problem with X servers is that there was a divergence of X11
colors. Each manufacturer added their own ones. SGI changed all the
values because of the different gamma correction on SGI hardware. HP
changed some values because they used different monitor phosphors. And
so forth.

JE> Perhaps W3C and XFree86 could establish a joint standard?

I believe that since the W3C color set is so tightly specified, the
best thing should convergence be deemed desirable would be for X
servers that implement color management to use the W3C set.

However, I also believe that there is no actual conflict. The set of
named colors, while clearly influenced by one particular set of X11
colors (amongst other influences), is not specified to mean 'display
what the X11 server calls this color name' but 'display the sRGB color
specified, taking into account any different monitor gamma, phosphors,
and ambient light conditions" for a fully accurate result. Thus, if an
SVG or CSS or XSL implementation is running on a system that uses X11,
it still gives conformant results regardless of the differences you
cite above.

-- 
 Chris                            mailto:chris@w3.org

Received on Tuesday, 23 April 2002 21:37:55 UTC