Re: Chromatic font research

From: Sampo Syreeni (
Date: Thu Jun 27 2002 - 06:05:43 EDT

On Wed, 26 Jun 2002, John Cowan wrote:

>I looked at the image (less than ideal) at
> and fed
>it through the Gimp to strip out color information (specifically,
>Image/Colors/Desaturate followed by Image/Colors/Threshold, taking the
>127 default, which leaves stark black and white).

The point was, there are a lot more. You can start with . And that is only for
the international flags. There are lots of national and proprietary ones,
too, if I'm not entirely mistaken. Unify the whole set and you'll end up
with a number of distinct symbols with no difference besides color.

>In any event, this is plainly a letter-by-letter cipher for the basic
>Latin alphabet (A-Z), and the fact that single letters or combinations
>may be used as codes for cross-linguistic concepts is of no more
>interest to a character encoding standard than that "ABALC" can mean
>"Abandon all claims" in any of various natural languages.

To a degree, yes. But the line is hardly clear-cut. By the same token, we
might as well deprecate many of the alphabets in Unicode as they can be
considered a cipher of IPA. The point is, a widely used cipher does
constitute a new alphabet. (We might get back into the Klingon debate,
here. I'd rather refrain.)

Again, I'm not saying signalling flags should be coded. Neither am I
disputing the fact that such signalling systems are subordinate to
conventional alphabets. I'm not even saying a single set of such flags
couldn't be validly coded in monochrome. (Any single set of signalling
flags is usually designed to be read in adverse lighting conditions where
color is poorly perceived. It is not an accident you found little trouble
distinguishing the flags in monochrome.)

What I'm suggesting, instead, is that in certain contexts where
printing/wrinting technology hasn't presented an obstacle, meaning can
indeed have been assigned to colors in what is basically running text.
This makes me think that the monochromatic nature of characters is more a
technological necessity than an inherent part of the definition of a
character. If this is the case, it wouldn't be a bad idea to prepare for
the inclusion of colored glyphs in Unicode, should the need arise, and to
be careful not to dismiss coloring as a potential primary feature of new
character when considering coding proposals in the future.

Sampo Syreeni, aka decoy -, tel:+358-50-5756111
student/math+cs/helsinki university,
openpgp: 050985C2/025E D175 ABE5 027C 9494 EEB0 E090 8BA9 0509 85C2

This archive was generated by hypermail 2.1.2 : Thu Jun 27 2002 - 04:30:18 EDT