<html>
<head>
<meta http-equiv="Content-Type" content="text/html; charset=UTF-8">
</head>
<body>
<div class="moz-cite-prefix">On 10/14/22 18:28, Adib Behjat via
Unicode wrote:<br>
</div>
<blockquote type="cite"
cite="mid:D9AB4CCF-FA1F-4F8F-8A76-F5B928B34C27@apple.com">
<meta http-equiv="Content-Type" content="text/html; charset=UTF-8">
I second Mark’s sentiment and really like the suggestion. I also
do think it would be wiser if this process was handled and managed
by OSs/Apps versus Unicode.
<div class=""><br class="">
</div>
<div class="">After reading Mark’s suggestion, I was reminded of a
game where you combine basic/generic elements to create more
complex elements (e.g. <a href="https://littlealchemy.com/"
class="moz-txt-link-freetext" moz-do-not-send="true">https://littlealchemy.com/</a>).<br
class="">
<div class=""><br class="">
</div>
<div class="">For example, if someone wants to generate “Physics
Teacher”, a user can type in their device:</div>
<div class="">🧑🏫⚛️</div>
</div>
</blockquote>
<p>Yes, this is essentially the "emoji kitchen" approach. It has
definite appeal and could work... and also distinct downsides:
lack of control over what you actually mean, different people
thinking of different formulæ, etc. No scheme is perfect. Those
alchemy games are indeed a good example of this kind of thinking,
but it has its ups and downs.<br>
</p>
<p>~mark<br>
</p>
<blockquote type="cite"
cite="mid:D9AB4CCF-FA1F-4F8F-8A76-F5B928B34C27@apple.com">
<div class="">
<div class=""><br class="">
</div>
<div class="">And based on this combination, the OS/App can give
the user the option to generate a custom avatar to represent
Physics Teacher. With regards to rendering, tools like DALL-E
(or other similar diffusion models) can enable this
capability. In addition, this process will help encourage the
introduction of more generic emoji characters to help expand
the foundational building blocks for these tools.</div>
<div class=""><br class="">
</div>
<div class=""><br class="">
</div>
<div class="">
<div class="">
<div class="">
<div class="">
<div class="">
<div class="">
<div class="">
<div>
<blockquote type="cite" class="">
<div class="">On Oct 14, 2022, at 2:43 PM,
Mark E. Shoulson via Unicode <<a
href="mailto:unicode@corp.unicode.org"
class="moz-txt-link-freetext"
moz-do-not-send="true">unicode@corp.unicode.org</a>>
wrote:</div>
<br class="Apple-interchange-newline">
<div class="">
<div class="">There are really two distinct
animi(?) behind the push for
ever-more-detailed emoji, or rather, two
animi for using emoji, and they pull in
different directions, almost opposite.<br
class="">
<br class="">
On one hand, people want an emoji that
looks JUST like they want it to look.
Maybe not even only for people! But of
course we see it most with people. People
want an emoji that looks like _they_ look
(do they really use it? or do they just
feel left out if it isn't available? I'm
not a heavy emoji-user, so I'm no judge,
but I bet the second motivation is
non-trivial as well.) So the really do
want "tall ectomorph female physics
teacher with dark hair in Princess Leia
buns, an eyebrow piercing (right), and a
birthmark on the left side of the chin."
This motivates the various suggestions of
somehow encoding a teeny image and
pretending it's somehow "plain text", or
encoding a link to it or something. I had
what I thought were helpful ideas about
the recent notions being floated at the
last Unicode meeting I attended, and maybe
they're even being thought about... But
although I've thought that maybe sending
images like this was the best of the
suggested solutions, it's still awful.<br
class="">
<br class="">
The other problem is the other animus
involved. Because it isn't just a picture
that people want when they use emoji. It
isn't enough that there's a little picture
of a person wearing a mortarboard hat or
something, there's actual semantic
information embedded in the encoded text
as well. It isn't just a picture, it's a
codepoint(-sequence) that means something,
a bit-sequence that _means_ "TEACHER."
Just like U+0065 means something more than
the ink used to draw it in whatever font.
People want some snippet of "text" that
not only looks like them, but also *means*
them. Hence in my example above,
expressing some of the various physical
traits are one thing, and you can
represent TEACHER with some cultural
convention like a mortarboard hat (or
THIEF with a mask), but how do you get
across "PHYSICS teacher"? Or a dozen
other subjects, arbitrarily finely
divided?<br class="">
<br class="">
When I think about it, I don't know that
people would really be satisfied with
image-sticker emoji. After all, not
everyone has the skill to draw them (which
is why we rely on emoji-font artists in
the first place), or make them Just So,
and I really do think that people would
feel the lack of semantic meaning. A lot
of messaging services already let you
include little graphics images, but I
don't think the people using them feel
this desire for new emoji any less. How
many homedrawn emoji do you really think
will be made? How many used more than
twice? How many used by more than one
person? How many will even be understood
by the recipient?<br class="">
<br class="">
Asmus' point about comparing it to
swapping out lego bits is well-taken.
There have long been these kind of "avatar
engines" that let you swap around features
to get something kind of like you (I
remember one on the Wii way back when.)
And maybe there is some reasonable limit
to how much customization can be provided
(though I'd bet anything we'd take forever
agreeing on it.) And even within specific
limits like hair-style and hats, there'll
always be one more that we're lacking, one
more Mr Potato Head piece people will push
for.<br class="">
<br class="">
(Actually, thinking about Asmus' line
about "faces drawn on," how's this for an
idea, combining raster with standard?
Instead of drawing some random picture
yourself for an emoji, you have an image
of, say, facial features that's to be
projected onto a blank emoji-face, which
can be any "standard" emoji or whatever.
Could have other distinct ways of
specifying headgear images, etc. The
renderer would be smart enough to scale or
transform the image appropriately for
different kinds of emoji with faces in
different places and orientations etc.
Probably a beast to implement, but I'm
just floating ideas. This one actually
has signs of MAYBE bridging the gap
between the two drives for emoji.)<br
class="">
<br class="">
Perhaps the best "generalized emoji"
implementation is something along the
lines of the Emoji Kitchen, where you can
combine arbitrary emoji in arbitrary
numbers and orders and the system does its
level best to figure out SOME way for the
resulting image to make sense. This gives
you some genericness, and you can express
all kinds of shades of meaning by
combining enough emoji, but retains the
semantic meaning of them as well. Of
course, it puts you at the mercy of how
the system chooses to combine things, how
good the designers are, etc etc. You have
no control over what really emerges at the
end. (I suppose if this ever became
something widespread there would develop
conventions for combining with a little
more control (like Egyptian hieroglyph
combiner marks?? Probably not, but with
some semantic similarities.))<br class="">
<br class="">
Anyway. Wanted to rant a bit on this "two
desires" notion that I was thinking about
since the last meeting. I think it's
important to remember the second one,
which gets missed out on when people focus
on controlling the picture just so (though
it is what's behind the idea of using
Wikidata codes.) And the "images of
features" notion occurred to me while
typing this, and I think it's interesting.<br
class="">
<br class="">
I'm not really trying to suggest answers
here (though I did remark on some things
favorably); this is more asking the
questions. There's always going to be
these two conflicting needs, and there'll
always be people who want ever-finer
distinctions in emoji, and there may
simply not be any really good answers.
Emoji probably never should have been part
of Unicode (not "plain text"), but that
ship sailed long ago, and even there it's
not cut and dried (webdings? map symbols?)<br
class="">
<br class="">
Thoughts?<br class="">
<br class="">
~mark<br class="">
<br class="">
<br class="">
On 10/14/22 00:54, Asmus Freytag via
Unicode wrote:<br class="">
<blockquote type="cite" class="">People
that grew up on games are used to
character editors that allow any avatar
to be assembled from building blocks.
Short of a common "avatar engine" shared
across all platforms, a limited set of
emoji-legos isn't that unreasonable.<br
class="">
<br class="">
We have skin tones, male/female, some
limited use of color (black + cat).<br
class="">
<br class="">
Because of their small size, emoji faces
would support more customization; it's
hard to create a full character emoji on
the level of detail of a game character.
So you'd be limited to less detail than
you can implement with real lego blocks.
(And yes, the ones for the heads of the
little figure have removable hair (and
head gear). Plus a variety of of faces
(pirate) painted on.<br class="">
<br class="">
If that can be done in the physical
world, there's no reason a subset of
that couldn't be supported in emoji
rendering.<br class="">
<br class="">
People will intuitively sense that that
should be possible and thus the pressure
to innovate in that direction won't
stop.<br class="">
<br class="">
Just my $1/50.<br class="">
<br class="">
A./<br class="">
<br class="">
On 10/13/2022 4:38 PM, Mark E. Shoulson
via Unicode wrote:<br class="">
<blockquote type="cite" class="">Again,
this way lieth madness. People aren't
satisfied with an emoji for "female
teacher with dark hair"; they want
"TALL, THIN, female PHYSICS teacher
with dark hair IN PRINCESS-LEIA BUNS
AND A PIERCED EYEBROW (GOLD RING)."
And if you give in on "welllllll,
okay, we'll give in on the
tall/short...," you're only
encouraging them to beg for the rest.
("How about only a _little_ tall? How
about broad-shouldered?
small-breasted?")<br class="">
<br class="">
(Though my opinion isn't actually
quite what that sounds like: even I
admit that there probably *are* things
that are appropriate to give in on,
and I know we all can argue all the
day long about them.)<br class="">
<br class="">
~mark<br class="">
<br class="">
On 10/13/22 09:22, William_J_G
Overington via Unicode wrote:<br
class="">
<blockquote type="cite" class="">Thank
you for posting about this.<br
class="">
<br class="">
Could one use variation selectors
with this too, so as to have a
default style of glasses and various
styles of glasses available?<br
class="">
<br class="">
Or would one need to have separate
styles of glasses each encoded
separately?<br class="">
<br class="">
If both approaches are possible,
which one would be better?<br
class="">
<br class="">
If it is to be encoded, and I hope
it will be, it would be good to go
for the lot all at once. Lots of
styles as glasses are in lots of
styles.<br class="">
<br class="">
In my opinion it is no use just
doing one and leaving the rest for
some future time as that is often a
recipe for the rest never getting
done.<br class="">
<br class="">
If the lot is done as one grand
forward leap then that is the way to
keep Unicode thriving.<br class="">
<br class="">
William Overington<br class="">
<br class="">
Thursday 13 October 2022<br class="">
</blockquote>
</blockquote>
<br class="">
<br class="">
</blockquote>
</div>
</div>
</blockquote>
</div>
<br class="">
</div>
</div>
</div>
</div>
</div>
</div>
</div>
</div>
</blockquote>
</body>
</html>