Re: Support emoji fonts

[Guo Yunhe <guoyunhebrave@xxxxxxxxx>]

I do not think Unicode organization has a standard for any subsets, because their goal is to define the total scope. It also means, we can create a must-have subset from practice. My suggestion is to compare all open source Emoji fonts available on Linux, choose the common symbols they contain. Here is the list of open source Emoji fonts I found:

1. EmojiOne
   
2. Twemoji
3. Noto Emoji / Google Emoji / Android Emoji (Same thing on different products)
   

The most popular Emoji fonts could be:

1. Apple Emoji
2. Google Emoji / Android Emoji / Noto Emoji
3. Twitter Emoji
4. EmojiOne
5. Facebook Emoji
6. Samsung Emoji

Others, like DoCoMo, KDDI and SoftBank emojis, might never be used on Linux desktop systems.

Solution 1: Unicode Emoji 3.0

From my test result, these fonts contains all characters of Unicode Emoji 3.0 ( http://unicode.org/emoji/charts/full-emoji-list.html ) . So I think Unicode Emoji 3.0 could be such a subset of Emoji and it fits all existing Emoji fonts that Linux users might use.

Solution 2: Intersection of Apple, Google, Twitter, EmojiOne, Facebook, Samsung Emoji fonts

This is a much smaller subset of current Emojis. It should covered all emojis that people frequently use. That is why they are included by all important communication services.

What do you think?

在 2016年09月08日 03:44, suzuki toshiya 写道:

Oops,

cellarphone. However, I don't think recent emoji
pushers do not care about this subset.

I mean: however, I don't think recent emoji
pushers care about this subset.

Regards,
mpsuzuki


suzuki toshiya wrote::

Dear Tagoh-san,

One of the recognizable subset would be the set to
interchange original "emoji" used by legacy Japanese
cellarphone. However, I don't think recent emoji
pushers do not care about this subset.

To consider other new emojis, should we ask for the
comments from Unicode (or ISO/IEC JTC1/SC2/WG2) experts,
to define the subset to judge whether the font is
sufficient to use to display emojis.

Also I'm interested in that fontconfig is expected
to pick the font supporting color glyphs, and/or,
supporting VS to display existing symbols with emoji-
style.

Regards,
mpsuzuki

Akira TAGOH wrote::

Well, you may misunderstood my question. let me rephrase. the question 
is, is a font required to contain all of them to say "our fonts support 
emoji" or to indicate that in fontconfig? and how many emoji fonts has 
supported all of them at this moment? in other words, if a font is more 
or less missing them, it won't be recognized as emoji-aware.
I don't see any mention about it there at least. .orth files in 
fontconfig doesn't contain all of Unicode code points which is used in 
those languages because some of them isn't often used and may not be 
implemented for priority etc.

On Wed, Sep 7, 2016 at 8:30 PM, Guo Yunhe <guoyunhebrave@xxxxxxxxx 
<mailto:guoyunhebrave@xxxxxxxxx>> wrote:

    Here is the official define of emoji characters. (Opening this page
    may hang your browser for a while!!!)
    http://unicode.org/emoji/charts/full-emoji-list.html
    <http://unicode.org/emoji/charts/full-emoji-list.html>

    Hope it would be helpful.

    在 2016年09月07日 14:21, Akira TAGOH 写道:

    The problem on that idea is how to figure out what the minimal
    coverage in emoji block. at this point, the minimal glyph coverage
    for langs are defined in fc-lang/*.orth and cache files contains
    lang property only which fonts satisfies the coverage for. if
    there are any specs defining a must or an optional to have, that
    may be helpful otherwise we may need to think about another idea
    for that.

    maybe good to have a property in a cache to indicate if a font has
    an emoji or not, and we could leave the way to use it to
    applications perhaps.

    On Wed, Sep 7, 2016 at 4:28 PM, Guo Yunhe <guoyunhebrave@xxxxxxxxx
    <mailto:guoyunhebrave@xxxxxxxxx>> wrote:

        Hi, I recently studied some emoji fonts. These emoji fonts
        have fontconfig difficulties when packaging. They try to set
        the font as default emoji font but do not affect others.
        Usually the font has a separated configure file.

        <match>
          <test name="family">
            <string>sans-serif</string>
          </test>
          <edit binding="strong" name="family">
            <string>Nimbus Sans L</string>
            <string>EmojiOne Color</string>
          </edit>
        </match>

        However, this will affect sans-serif font settings of other
        font packages or users' setting, because the package do not
        know which sans-serif font users want to use.

        I suggest maybe we can map the Unicode emoji block as test
        condition. Just like when we set a Japanese font, it won't
        affect English and Arabic fonts.

        <match>
          <test name="family">
            <string>sans-serif</string>
          </test>
          <test name="lang">
            <string>emoji</string>
          </test>
          <edit binding="strong" name="family">
            <string>EmojiOne Color</string>
          </edit>
        </match>



        -- 
        Guo Yunhe
        _______________________________________________
        Fontconfig mailing list
        Fontconfig@xxxxxxxxxxxxxxxxxxxxx
        <mailto:Fontconfig@xxxxxxxxxxxxxxxxxxxxx>
        https://lists.freedesktop.org/mailman/listinfo/fontconfig
        <https://lists.freedesktop.org/mailman/listinfo/fontconfig>




    -- 
    Akira TAGOH

    -- 
    Guo Yunhe




-- 
Akira TAGOH

--
Guo Yunhe

_______________________________________________
Fontconfig mailing list
Fontconfig@xxxxxxxxxxxxxxxxxxxxx
https://lists.freedesktop.org/mailman/listinfo/fontconfig