Glyphs not part of the required glyphsets

[vv-monsalve]

@josescaglione As part of the subsetting job for the per-script families, I've identified some glyphs included in the font that are not part of the required Glyphsets or are not used in the font. Please review them and indicate in which language or context they are used or if they can be ignored/deleted.

Glyph Name	VV Comment	TT Comment
acutecomb_macroncomb	Not used by any glyph	I believe we can remove
breve-cy	Used by brevecomb-cy, could only one be used?	Will enquire
brevecomb-cy	Uses breve-cy as comp, could only one be used?	Will enquire
bulletoperator	Not defined in a glyphset	Basic Multilingual Plane
commaaboverightcomb	Not used by any glyph	Usage unclear, was part of initial Afr lists and we decided to keep
commaturnedmod		I believe we can remove
dieresistonoscomb	Not used by any glyph	Greek
divisionslash	Deprecated from our definitions	Basic Multilingual Plane
Ebreve	Auxiliary glyph in multiple Latn languages, GF TransLatin Pinyin	Latin Extended-A
ebreve	Auxiliary glyph in multiple Latn languages, GF TransLatin Pinyin	Latin Extended-A
Ecedillabreve		Cleaned up African list – February
ecedillabreve		Cleaned up African list – February
Ezhcaron		I believe we can remove
ezhcaron		I believe we can remove
firsttonechinese	Not used by any glyph	Pinyin
Gcircumflex	eo_Latn Esperanto GF Latin Beyond	Latin Extended-A
gcircumflex	eo_Latn Esperanto GF Latin Beyond	Latin Extended-A
glottalstopreversed	nuk_Latn Nuuchahnulth GF_Latin Beyond for the Americas	Cleaned up African list – February
graphemejoinercomb	Not defined in any lang file	Hebrew
gravecomb_macroncomb		I believe we can remove
hmod	zap_Latn Zapotec GF Phonetics IPAStandard	Cleaned up African list – February
Ibreve	Auxiliary glyph in multiple Latn languages, GF TransLatin Pinyin	Latin Extended-A
ibreve	Auxiliary glyph in multiple Latn languages, GF TransLatin Pinyin	Latin Extended-A
Ismall	GF Latin African present in prod names but not in nice names list	Cleaned up African list – February
Jcircumflex	eo_Latn Esperanto GF Latin Beyond	Latin Extended-A
jcircumflex	eo_Latn Esperanto GF Latin Beyond	Latin Extended-A
kgreenlandic	kl_Latn Kalaallisut GF Latin Beyond	Latin Extended-A
lameddageshholam-hb		Will enquire
lamedholam-hb		Will enquire
Lbelt	hur_Latn Halkomelem, not part of any glyphset	Cleaned up African list – February
lbelt	hur_Latn Halkomelem GF Latin Beyond	Cleaned up African list – February
Ldot	Deprecated from our requirements in favor of periodcentered.loclCAT.case	Latin Extended-A
ldot	Deprecated from our requirements in favor of periodcentered.loclCAT	Latin Extended-A
longs	Auxiliary glyph in France and German	Latin Extended-A
lowlinecomb	no lang, GF Latin Beyond`	Usage unclear, was part of initial Afr lists and we decided to keep
macroncomb_acutecomb		I believe we can remove
macroncomb_gravecomb		I believe we can remove
Mhook	not found in any of the language definitions	Cleaned up African list – February
mhook	No lang, GF Phonetics IPAStandard	Cleaned up African list – February
Obreve	Auxiliary glyph in multiple Latn languages, GF TransLatin Pinyin	Cleaned up African list – February
obreve	Auxiliary glyph in multiple Latn languages, GF TransLatin Pinyin	Cleaned up African list – February
Otildedieresis	not found in any of the language definitions	Cleaned up African list – February
otildedieresis	not found in any of the language definitions	Cleaned up African list – February
rfishhook	No lang, GF Phonetics IPAStandard	Cleaned up African list – February
Rmacronbelow	not found in any of the language definitions	?
rmacronbelow	not found in any of the language definitions	?
Ubreve	Auxiliary glyph in multiple Latn languages, GF TransLatin Pinyin	Latin Extended-A
ubreve	Auxiliary glyph in multiple Latn languages, GF TransLatin Pinyin	Latin Extended-A
Udieresiscaron	zap_latn Zapotec, aux, GF TransLatin Pinyin	Pinyin
udieresiscaron	zap_latn Zapotec, aux, GF TransLatin Pinyin	Pinyin
Udieresisgrave	zap_latn Zapotec, aux, GF TransLatin Pinyin	Pinyin
udieresisgrave	zap_latn Zapotec, aux, GF TransLatin Pinyin	Pinyin
Udieresismacron	No lang GF TransLatin Pinyin	Pinyin
udieresismacron	No lang GF TransLatin Pinyin	Pinyin
verticallinebelowcomb	not found in any of the language definitions	Cleaned up African list – February
wordjoiner		Recommended for non latin scripts
yturned	No lang, GF Phonetics IPAStandard	Cleaned up African list – February
Zcircumflex	No lang, GF Latin Beyond	Cleaned up African list – February
zcircumflex	No lang, GF Latin Beyond	Cleaned up African list – February
zerowidthspace		?
Zmacronbelow	No lang, GF Latin Beyond	African reference alphabet
zmacronbelow	No lang, GF Latin Beyond	African reference alphabet

[vv-monsalve]

@moyogo, could you please take a look at this list and help us determine whether any of them should be listed in the Latin African glyphs? E.g. Mhook, rfishhook.

[josescaglione]

Hi Viviana, you are asking very hard questions :)

Your list features a mix of things and I cannot easily know what languages use which glyphs. Your list does feature a mix of things spaning from characters present in Latin Extended A which we suggest to support almost fully, Pinyin, characters that we identified as used in African languages back in February this year (we submitted all of the documentation back then and later on directly to you), and a few glyphs I am personally not familiar with.

I will try to reply to your specific questions and add any notes we can. And of course I am happy to remove anything you do not want in the font.

[vv-monsalve]

Under the strategy of using a single source file to create different script families, a subsetting process is essential. This requires us to specify which elements will be included in each family.

I began this process by utilizing our Glyphset definitions, as they outline our font requirements. However, I noticed that many glyphs were omitted when compared with the previous fonts.

This review's purpose is to gain a clear understanding of the what is in the source/font and the reasons behind those choices. This way, we can make informed and strategic decisions about what to include in the final output fonts.

[josescaglione]

did not manage to find answers for them all but I think you will find some of the information you need here: Lista viviana.numbers.zip

[vv-monsalve]

Thank you for the information. I've added it to the table above as well as additional info from our languages database.

characters present in Latin Extended A

We have detached from the Unicode blocks and moved to a language-centered approach to define our Glyphsets.

Pinyin, characters that we identified as used in African languages back in February this year (we submitted all of the documentation back then and later on directly to you)

When we discussed revising the Glyphsets, I mentioned that the Cyrillic Plus set needed to be reviewed and redefined. Therefore, the information you provided would be taken into account for that task (which we did). However, the Latin African glyph set is considered defined after @moyogo extensive work over the past year and early this year. Nevertheless, I'm also asking him to review the list and provide feedback.

[moyogo]

I double checked some of these since the language data added to gflanguages was much smaller than what I would have liked to cover, so there’s likely gaps for less common cases.

In short, I’d recommend keeping:

• /Ismall and its capital /uniA7AE
• /lowlinecomb as it sometimes used instead of /macronbelowcomb,
• /commaturnedmod as it is a transformed composite glyph of /apostrophemod used in English in Hawaii, mostly in toponyms or people’s names. While looking for African languages that might use /commaturnedmod for a pharyngeal stop like in Arabic transliteration, it turns out instead that ʕ /glottalstopreversed is in use Ethiopa (https://github.com/googlefonts/lang/pull/179).

Here are long comments with short comments in the table at the end.

While /Abreve/abreve is used in Tuareg languages (Tamashek [tmh] and Tamasheq [taq_Latn] or Tawallammat Tamajaq [ttq_Latn]), the other vowels with breve /Ebreve/ebreve/Ibreve/ibreve/Obreve/obreve/Ubreve/ubreve are not used in current African orthographies. They have been used in really old African orthographies but those are out of scope. /Ubreve/ubreve is used in Esperanto. /Abreve/abreve/Ebreve/ebreve/Ibreve/ibreve/Obreve/obreve/Ubreve/ubreve are used in Jarai in Vietnam.

/firsttonechinesemay may be used to illustrate the use of /macroncomb, for example "ˉ is used to indicate middle tone on vowels".

/lowlinecomb may be used instead of /macronbelowcomb in particular when practical orthographies use underlining, like in Gabon.

/commaturnedmod is used in Hawaiian and other Austronesian languages like Tongan, Wallisian and sometimes in Tahitian. It can also be used in Uzbek. Semitic languages transliteration systems use it as well, for examples transliteration of Arabic or Ge’ez (Gəʻəz). For Hawaiian, it occurs in toponyms in English in Hawaii, for example on Apple Maps or Google Maps, and several documents. It may be used in Gawwada or other languages of the Dullay language cluster in Ethiopia (alternatively ʕ may be used instead), but it’s not clear if that is in a practical orthography or only in linguistic works, futher investigation is needed. Design-wise, it’s a turned /apostrophemod, so it’s a low hanging fruit.

/Ezhcaron/ezhcaron is used in Skolt Sami mostly in Finland. It is used in Laz, for example in the orthography used in school manuals in Turkey.

/Gcircumflex/gcircumflex is used in Esperanto. It is also used in Aleut in the USA. In Guinea, Ĝ ĝ is sometimes used as an adhoc substitute for Ɠ ɠ /Ghook/ghook (the same is true for b̂, d̂, ĥ and ɓ, ɗ, ɦ, or others).

/hmod is used in several North American orthographies, or phonetic transcriptions like IPA.

/uniA7AE/Ismall is used in one of the Koulango orthographies in Côte d’Ivoire.

Ĵ ĵ /Jcircumflex/jcircumflex is used in Esperanto.

ĸ /kgreenlandic was used in Kalaallisut (Greenlandic) in the old orthography and was replaced by q in the 1973 orthography. It’s not clear if Inuttitut still uses it as K is used instead in practice or in documents using the 1980 orthography, however the Labrador Inuttitut Heritage Bible uses ᴋ /Ksmall in the digitized version.

Ɬ ɬ /Lbelt/lbelt is used in several North American orthographies, sometimes as an alternate of /Lslash/lslash or Ƚ ƚ /Lbar/lbar or Ɫ ɫ /Lmiddletilde/lmiddletilde.

/Ecedillabreve/ecedillabreve is used in the ISO 259 transliteration of Hebrew and may appear in library catalog or academic works. ISO 259 also uses /abreve/obreve, other transliteration systems of Hebrew may use those and other vowels with breves.

Ɱ ɱ /Mhook/mhook are not used in African orthographies as far as we know. ɱ is used in various phonetic transcriptions including IPA, Ɱ is used in some Americanist phonetic transcription system for a voiceless equivalent. Until we find attestations of use, it can be removed and kept for a phonetic glyphset.

/Otildedieresis/otildedieresis isn’t used in orthographies as far as I know. It seems to be used in some academic phonetic transcriptions. If tilde and dieresis are used together in orthographies, they’d like be in a different order ö̃ instead.

ɾ /rfishhook is used in the IPA, not in orthographies.

/Rmacronbelow/rmacronbelow is used in Pitjantjatjara in Australia and in Mitla Zapotec in Mexico, and in some transliteration system of Tamil.

/Udieresiscaron/udieresiscaron and /Udieresisgrave/udieresisgrave are used in Hanyu Pinyin transcription of Chinese, and can be used in Southern Tutchone in Canada. /Udieresismacron/udieresismacron was used in Southern Tutchone but is now written with ü̂ /udieresis/circumflexcomb.

/verticallinebelowcomb is an alternate to /dotbelowcomb. Some Yoruba works use that shape for the dotted letters for example. Having a stylistic set that substitutes the /dotbelowcomb (after decomposing composite glyphs) by the /verticallinebelowcomb would be interesting, but it’s not clear if that’s easily accessible to users who could just use the character instead if they want to display that shape instead.

/yturned is an IPA symbol. It is not used in orthographies. It may appear in text in Heiltsuk or Kwakwala (both North American languages) as a substitute for λ /lambda or the Unicode 16.0 ꟛ /lambda-latin. A lambda was also used in the 1982 African Reference Alphabet proposed by Michael Mann and David Dalby, but there’s no evidence it has been used in orthographies with the turned y shape or with the lambda shape.

/Zcircumflex/zcircumflex The 1966 Niger national alphabet and the 1978 African Reference Alphabet have a z with right mid hook for Tamasheq, it was replaced by ẓ in the 1999 Niger national alphabet. In some documents hooked letters are substituted with letters with circumflex, it’s possible ẑ was used. Otherwise used in Chilcotin (Canada) or ISO 9 transliteration of Cyrillic.

/Zmacronbelow/zmacronbelow is used in Tahltan and Yatee Zapotec (both in Mexico), or some transliteration systems like for Persian or Hebrew.

Glyph Name	Comment
acutecomb_macroncomb	Can be removed.
breve-cy
brevecomb-cy
bulletoperator	Can be removed.
commaaboverightcomb	Can be removed.
commaturnedmod	Recommended to keep. Used in Hawaiian, Tongan, Samoan. Transformed composite of /apostrophemod (or /quoteright, but better rounder and slightly larger).
dieresistonoscomb	Greek, may be the same as dieresiscomb_acutecomb in some fonts
divisionslash
Ebreve	Can be removed. Old orthographies, transliterations, phonetic transcripts or Jarai (Vietnam) are out of scope.
ebreve	""
Ecedillabreve	ISO 259 transliteration of Hebrew
ecedillabreve	""
Ezhcaron	Used in Skolt Sami (Finland) and Laz (Turkey)
ezhcaron	""
firsttonechinese	May be used to illustread use of /macroncomb. Simple composite of /macroncomb or duplicate of /macron.
Gcircumflex	Esperanto. Aleut (USA). Ad hoc substitution for Ɠ ɠ in Guinea.
gcircumflex
glottalstopreversed	Can be removed. North American languages out of current scope.
graphemejoinercomb	Can be removed.
gravecomb_macroncomb	Can be removed.
hmod	Can be removed.
Ibreve	Can be removed. Old orthographies, transliterations, phonetic transcripts or Jarai (Vietnam) are out of scope.
ibreve	""
Ismall	Used in one of the Koulango orthographies in Côte d’Ivoire, along with capital /uniA7AE.
Jcircumflex	Esperanto
jcircumflex	""
kgreenlandic	Old Kalaallisut orthography, maybe still in Inuttitut (both North American)
lameddageshholam-hb
lamedholam-hb
Lbelt	North American orthographies
lbelt	""
Ldot	Legacy encoding
ldot	""
longs	Historical
lowlinecomb	Recommended to keep. May be used instead of /macronbelowcomb or for orthographies that used underline on typewritters. Some orthographies in Gabon might use it as such.
macroncomb_acutecomb	Can be removed.
macroncomb_gravecomb	Can be removed.
Mhook	Can be removed. Phonetic symbol.
mhook	Can be removed. Phonetic symbol.
Obreve	Can be removed. See Ebreve
obreve	""
Otildedieresis	Can be removed. May be used in some phonetic transcriptions.
otildedieresis	""
rfishhook	Can be removed. IPA Symbol.
Rmacronbelow	Pitjantjatjara (Australia), Mitla Zapotec (Mexico), or transliteration Tamil.
rmacronbelow	""
Ubreve	Can be removed. See Ebreve
ubreve	""
Udieresiscaron	Southern Tutchone (Canada), Hanyu Pinyin
udieresiscaron	""
Udieresisgrave	Southern Tutchone (Canada), Hanyu Pinyin
udieresisgrave	""
Udieresismacron	Southern Tutchone (Canada), Hanyu Pinyin
udieresismacron	""
verticallinebelowcomb	Variant of /dotbelowcomb
wordjoiner	Likely not used in Latin script.
yturned	Can be removed. IPA.
Zcircumflex	Can be removed. Chilcotin (Canada) and ISO 9 transliteration.
zcircumflex	""
zerowidthspace	Likely not used in Latin script.
Zmacronbelow	Tahltan and Yatee Zapotec (Mexico), or transliterations.
zmacronbelow	""

[josescaglione]

This is great!

Please let me double check the following: zerowidthspace wordjoiner graphemejoinercomb ...as well as the two Hebrew glyphs

For the rest, @vv-monsalve please confirm and I will remove glyphs following the above instructions.