google / emoji4unicode Goto Github PK

MarkD commenting on issue 19:

I'd suggest all the circled ones together, either before or after all the
squared ones.

e-B2B
e-B3D
e-B43
e-B50
next to the other circled ones. 

Compare again the sets of ARIB and Emoji symbols. Review currently proposed 
unifications, and review whether we should unify more symbols.

For example, in cross-references.

William Overington pointed me to a thread at http://forum.high-
logic.com/viewtopic.php?p=10544#p10544 where he made the following comment:

In the http://www.unicode.org/~scherer/emoji4u ... t/utc.html document as 
it stands as I write this post, I did notice one item straightaway.

About a fifth of the way down is an item labelled as YEOMAN OF THE GUARD 
with the note • Beefeater, British below it. Yet the illustration is of a 
Guardsman of a Foot Guard regiment, not of a Yeoman of the Guard, which is 
a different group.

The Yeomen of the Guard wear Tudor-style uniforms, with a flat hat. They 
are typically shown on the television at the State Opening of Parliament in 
England. They are not the same group as the Yeoman Warders of the Tower of 
London. Readers who know of the Gilbert and Sullivan Opera named Yeoman of 
the Guard may also know that that name is in that Opera being wrongly 
applied to the Yeoman Warders of the Tower of London.

There are pictures on the following page.

http://en.wikipedia.org/wiki/Yeomen_of_the_Guard

http://en.wikipedia.org/wiki/Yeomen_Warders

http://en.wikipedia.org/wiki/Foot_Guards

ARIB-9003=U+26CE TRAFFIC WARNING just looks like another heavy exclamation 
point. We should propose changing its name to something descriptive of the 
shape rather than its Japanese TV Symbols semantics.

Please add comments for further proposed changes of names of characters in 
the encoding pipeline.

SoftBank symbols have old and new IDs. The current data contains the old 
IDs. Add the new IDs and show them in the generated charts.

See http://www.unicode.org/charts/symbols.html
especially
Geometrical Shapes http://www.unicode.org/charts/PDF/U25A0.pdf
Arrows http://www.unicode.org/charts/PDF/U2190.pdf
and http://www.unicode.org/charts/PDF/U27F0.pdf
and http://www.unicode.org/charts/PDF/U2900.pdf
Miscellaneous Symbols and Arrows 
http://www.unicode.org/charts/PDF/U2B00.pdf
(e.g. squares like e-B6B, e-B71, etc.)

See Open Issues: https://docs.google.com/Doc?docid=ddsrrpj5_44cnr2fj64

... and report to people involved in starting the font design

Do not count unified symbols nor in_proposal="no" symbols.

Maybe print at the bottom of the HTML charts?

Per comment by KameOyaji:

http://groups.google.com/group/emoji4unicode/browse_thread/thread/b2e536ad983594
6

e-4C5: BOUTIQUE 109 should not be in the UTC proposal since it refers to a
fashion chain boutique store in Tokyo -- the one in Shibuya being most famous. 

The Softbank's name for this Emoji is: Shibuya. Not Shibuya 109. But it
still remains true that the image has 109 logo in it. 

I propose to 

1) change the name to SHIBUYA
2) move it to the Softbank specific section and move it out of the UTC proposal

OPENWAVE:
http://www.unicode.org/~scherer/emoji4unicode/snapshot/full.html#e-B89

is a business logo and should be moved to KDDI specific section and be
removed from the UTC version of the table. 

- if unified with an existing character,
  verify that the name is the same as that character's
- otherwise, verify that the name is not already used in Unicode

Optionally generate a shorter version of the chart, with fewer details,
less scrolling, fewer printed pages.

We don't seem to need the carrier codes except for examining cross-mappings.

Generate an HTML chart suitable for a font designer.

Add support for <design> sub-elements of <e> in emoji4unicode.xml.

Show Unicode code points for the new symbols font, instead of symbol IDs. 
Probably something like base+ID where base is a Unicode Private Use code 
point (Maybe U+E000 if we want to use BMP code points, or U+FF000 if we 
want to avoid collisions with other PUA uses.) The "ID" is our symbol ID. 
(A font designer will need a code point, and we can easily read our ID from 
that if we know the base.)

Hide irrelevant data, like carrier Unicode/Shift-JIS/JIS codes.

Add data (at least code points & names) for ISO 10646 AMD6 characters and 
check for character name problems.

For the font design, we want to know how many symbols use representative
images from which carrier. Also, given some existing or assumed resources
for some of the symbol sets, we want to count remaining symbols where we
may need to start from scratch.

Consider adding per-symbol anchors into the generated HTML charts to make 
it easy to jump to a symbol row via #e-4B0 or similar.

We received data with the official names.

covered by round-trip mappings with symbols

The ARIB-9138 (U+26F7) character has a ski with a person on it.
All 3 carriers have just a ski and a boot on it. If the ARBI name were
something like "SKIING", this would be acceptable but the proposed name is
"SKIER". Unless the ARIB name can be changed, I suggest that we keep it as
SKIING or something like "A SKI AND A BOOT".

e-4B4: BLACK CROSS ON SHIELD

This unification was made in r65: 

http://code.google.com/p/emoji4unicode/source/detail?r=65

There is a new Unicode character ARIB-9109 (U+26E8) but this is a black
cross within a shield. If you look at Emoji for 3 carriers, you see that
all of them have a house and a cross on it. 

I feel that we should not unify this unless the ARIB proposal name can be
changed to something like "HOSPITAL CROSS".

While DoCoMo and KDDI have just a single "clock symbol", SoftBank has 12 
symbols, one per full hour. We currently round-trip e-027 "10 oclock" with 
SoftBank #239=#old45 which makes sense given the continuity in the old 
numbers. However, the current SoftBank #239 image is quite different from 
their other clocks, and both the (new?) image and the new numbers (they now 
jump from #369 "9 oclock" to #370 "11 oclock") suggest that they have 
changed their preference.

It seems like we should have a round-trip between e-02A "clock symbol" and 
SoftBank #239=#old45, and a fallback from e-027 "10 oclock" to that same 
SoftBank symbol.

Unify e-B82 key/password with ARIB-9071=U+26BF, which is a key inside a 
square?

A suggestion from: Werner LEMBERG <[email protected]> to the [email protected].

"Maybe a typo: Ophiucus is normally written as Ophiuchus (at least in
the Astronomical world)."

Sometimes we see the spelling OPHIUCUS but the majority of web references use:

OPHIUCHUS

I propose to rename this entry to OPHIUCHUS.

I think e-4BC FOUNTAIN can be unified with ARIB 9125 = new U+26F2

James Kass reports the following misspellings:
e-355 SPEAK NO EVIL MONEY
e-502 BOOK WITH VERICAL FILL

From Asmus Freytag 2008-feb-01:

Clock faces, computer/document icons, as well as a rather significant
number of other symbols are present in the suite of wingdings fonts
distributed by Microsoft. A cross mapping to these would be a useful
exercise - not the least because these fonts represent existing black
and white interpretations of the glyph shape(s) for such symbols. These
glyphs might represent possible starting points for representative
glyphs, should these characters be encoded.

See Open Issues: https://docs.google.com/Doc?docid=ddsrrpj5_44cnr2fj64

For correctness sake, add the name of "Satoru Takabayashi (revisions)"
under "Main authors:" in trunk/data/emoji4unicode.xml

In discussing issue #16 Mark/Kat/Markus looked at the adjacent e-B98..e-B9C
symbols with hands and index fingers pointing in various directions.

We noticed that e-B98 shows the palm of the hand while the other four show
the back of the hand.

We agreed to unify e-B98 -- rather than e-B99 -- with U+261D WHITE UP
POINTING INDEX because the existing characters all show the palm of the hand.

We agreed to disunify e-B99..e-B9C from existing characters and give the
new symbols names like the existing ones but with "BACKHAND" inserted.

e-1E3 does not round-trip to any carrier and does not have any
representation. I assume it's a Google-created symbol. If so, then it
should be in_proposal="no".

They look to me like they are phone IME status indicators and belong into
"Communication (4. Artifacts)" or "Phone Specific (5.
Activities/Work/Entertainment)".

The charts should show whether we are unifying with an existing (Unicode 
5.1) character or an upcoming one (Unicode 5.2/AMD6). For upcoming ones we 
should not show the actual characters because there won't be fonts for 
them.

Its description says "Personal Digital Cellular Symbol/Logo. ?"

Notes from feedback from UTC meeting 2008-feb-05:
KDDI 279 Top Secret Sign =U+3299?
Softbank 201 Existence Sign =U+3292
Softbank 203 Monthly Sign =U+328A (there may be additional characters in 
the vicinity of U+328A that can be unified with other symbols in question 
here)
KDDI 384 Service Sign =U+32DA?
KDDI 402 Celebration Sign =U+3297?

TODO: Need to verify unifications and apply them in the table.

Discuss: Any further feedback on the proposed unifications? Characters like 
U+3299 are much less styled (they look like the Han characters with 
enclosing circle) than the symbols in the Emoji context.

Peter Edberg: Docs use KDDI #279 & U+3299 together
Mark: Consider whether plain text distinction
Peter: Maybe VS?
Markus: JIS source separation but otherwise unify

See Open Issues: https://docs.google.com/Doc?docid=ddsrrpj5_44cnr2fj64

no intersection of Emoji unifications with Shift-JIS round-trips

not clear if this is the right thing to do if the Shift-JIS round-trip is 
outside the JIS X 0208 part

e-B94 SCISSOR IN HAND GAME is currently disunified from U+270C VICTORY HAND

According to my notes from the 2008-aug UTC meeting:
Ken agreed with the disunification because they are distinct gestures in 
the real world.
Mark recommended to unify now, and if the distinction is really needed 
later, then add a specific character at that time.
The meeting did not end conclusively on this topic.

For the encoding proposal, new symbol characters will need proposed Unicode 
code points. We should try to fill existing blocks with similar symbols, 
and otherwise add symbols on plane 1.

In the HTML chart, in a per-carrier-per-symbol cell, if there is no 
corresponding carrier symbol and there is also no text_fallback for the 
symbol, show the geta mark (U+3013 〓) in text fallback style, rather than 
"-".

This is how charts used to show this information with previous tools. 
(Before this project was created.)

According to my notes from the 2008-aug UTC meeting, the discussion ended 
with Ken's recommendation to unify e-B07 wavy length mark with U+3030 wavy 
dash (without change in properties), and to encode e-B08 looped length mark 
as a new compatibility character.

Look for "May unify with..." or "Possibly unify with..." and similar 
comments.

For example, one of the happy faces should be unified with U+263A.

emoji4unicode/src$ ./emoji4unicode_test.py
source separation error: e-AF8 = U+2191 = Shift-JIS-81AA
source separation error: e-AF9 = U+2193 = Shift-JIS-81AB
source separation error: e-AFA = U+2192 = Shift-JIS-81A8
source separation error: e-AFB = U+2190 = Shift-JIS-81A9

They are e-AF8 UPWARDS ARROW .. e-AFB LEFTWARDS ARROW.

Disunify and rename with prefix "HEAVY"?

Check that our <ann>otations follow the syntax of a small subset of the 
Unicode NamesList.txt file format. (Except that we don't use the TAB which 
starts most CHAR_ENTRY lines; in an <ann> element that's implied.)

Format the recognized types of annotations as described in NamesList.html. 
For example, it says there:

COMMENT_LINE: TAB "*" SP EXPAND_LINE
    // * is replaced by BULLET, output line as comment

Should the SoftBank symbols #205, #206, #207 (見出しボタン 1..3) map to e-
B63=KDDI #40, e-B64=KDDI #41 and e-B67 as the data currently does?

These symbols are used as menu selectors.

Also review the name of e-B67 with regard to being in a series with the 
other two SoftBank symbols.

Should we rename the 4 operators to "HEAVY" followed by exactly the names
of the corresponding normal characters?

That would mean
e-B51 HEAVY PLUS -> HEAVY PLUS SIGN
e-B52 HEAVY MINUS -> HEAVY HYPHEN-MINUS
e-B53 HEAVY TIMES -> HEAVY MULTIPLICATION SIGN
e-B54 HEAVY DIVISION -> HEAVY DIVISION SIGN

For example, we have HOURGLASS and HOURGLASS 2. It would be good to have 
somewhat more interesting name variations than appending "2".

e-7E5: The original Softbank data says the name is ｢RV者」. 

This is clearly a typo. I propose to correct this to 「RV車」.

... isn't it always called a Four Leaf Clover?

asks David Starner

According to my notes from the 2008-aug UTC meeting, there was consensus to 
rename the flags from
FLAG OF JAPAN
to
FLAG SYMBOL JP
etc.
(using 2-letter country codes)

There was an open issue about what to use for representative glyphs -- real 
flags, or flags with the 2-letter codes in them, or something else.

We should post a public review of the chart by mid-December 2008 and invite 
feedback. I will start a cover page on the unicodesymbols site.

This issue is intended for the collection of various additions of 
font/glyph design instructions via <design> sub-elements of appropriate <e> 
elements in emoji4unicode.xml. We can check in several of these together.

e-83B KEYPAD 10: Should look like digits 10 enclosed by a keycap like 
U+20E3.
(We should move KEYPAD 10 to immediately after KEYPAD 0.)

We have annotations on symbols that are unified with existing characters. 
We should propose those annotations as additional annotations on their 
characters.

We can also collect additional annotations for other existing/upcoming 
characters in this issue.