Re: HTML 3.2 PR color value syntax

David Perrell (davidp@earthlink.net)
Tue, 12 Nov 1996 11:24:44 -0800


Message-Id: <199611121925.LAA25414@hungary.it.earthlink.net>
From: "David Perrell" <davidp@earthlink.net>
To: <www-html@w3.org>, "Walter Ian Kaye" <walter@natural-innovations.com>
Subject: Re: HTML 3.2 PR color value syntax
Date: Tue, 12 Nov 1996 11:24:44 -0800

Walter Ian Kaye wrote:
> MacOS has used 48-bit color for many years, and gets "downgraded"
> for use with 24-bit HTML values.

There are visible color variations that 24-bit is too coarse to
resolve; shadow detail is lost and a discerning eye can see 'stepping'
in subtle gradients.

> 6%? 101/256 = 39.45% on my calculator. Did you use the Pentium
'fdiv'? ;-)

Nope. 101^3 = 1030301 colors. 256^3 = 1677216 colors. 1030301/1677216 =
approx. 0.06141.

> I, for one, cannot visually discern 0-100% from 0-255 in rendered
colors,

I can see a very distinct difference between some color pairs with only
1% variation in one color channel. Admittedly not enough to worry about
for WWW work, and it's a hell of lot better than NetSnake's 20%
increments. But I still think it's silly to limit resolution to avoid
thinking in non-decimal increments.

David Perrell