Re: Emoji: Public Review December 2008

From: Asmus Freytag (
Date: Sun Dec 21 2008 - 02:25:18 CST

On 12/20/2008 10:23 PM, Doug Ewell wrote:
> David Starner <prosfilaes at gmail dot com> wrote:
>> I'm not sure how much more needs to be said. Plain text is what
>> people are using as plain text.
> The philosophical/ethics response to this would be that if the
> boundaries of plain text are that subjective and fungible, then there
> really are no boundaries, and we ought to either (a) encode all text
> as images or (b) accept William Overington's proposals to define
I'm all in favor of adding such things to CSS - if there's enough of a
use case for "red with yellow sparkles"...
>>> Well, thank goodness I never suggested any of these.
>> Then what are you suggesting?
> *Not encoding* the images that do not meet the established criteria
> for encoding. Letting the cell-phone vendors define private-use
> characters for them. (That is not the same as "endorsing permanent
> private use code assignments," which implies that Unicode itself is
> doing the endorsing.)
That's not altogether too helpful in solving the practical issue of
interoperability - which is nearly nil with PUA codes.
>>> How, for example, are we supposed to distinguish between CHICK and
>>> HATCHING CHICK unless our fonts and rendering engines (or printed
>>> pages) support animation?
>> It's the difference between a chick and a chick popping out of an egg.
> Have you seen the comparison chart? Two of the three vendors used the
> same image for both "characters." The third vendor defined an
> animation. If any of them had offered a static picture of a chick's
> head popping out of an egg, that might have been more convincing --
> though it remains to be proved that these images both have semantic
> value, and are not just pictures of what they are pictures of.
No, it does not have to be proven. By having enabled a large and diverse
group of end-users to create texts that contain these items a
text-elements, they have ipso facto become encodable. Whether the heart
in "I heart NYC" means "heart" or "love" or whatever, is actually
uninteresting to the character encoder. If the symbol has recognizable
shape and is used in regular (text) interchange between writer and
reader it it subject to encoding.

In other words, the symbol clearly has some sort of conventional meaning
(even if the character encoder doesn't know them or all of them at the
start) and it is possible to draw a rough line that designates what
acceptable ordinary (as well as fancy) glyph variations might be. Those
are the very practical questions that have to be resolved when actually
encoding something.

Those other kinds of questions are more interesting when a symbol is not
(yet) demonstrably used in text interchange. Much of the WG2 guidelines
were drawn up with an eye towards discouraging people to use Unicode
merely as a catalog for any random type of symbol, whether used (or even
usable) in text or not. Once symbols are demonstrably used in plain text
interchange, the whole equation changes.


This archive was generated by hypermail 2.1.5 : Fri Jan 02 2009 - 15:33:07 CST