Inconsistent styling of Chinantec tone marks

[dscorbett]

Font

NotoSans-Regular.otf
NotoSansMono-Regular.otf
NotoSerif-Regular.otf

Where the font came from, and when

Site: https://github.com/googlefonts/noto-fonts/blob/fcf8d5d5568fc96efcc4cbe6f9fbd402f62de98a/phaseIII_only/unhinted/otf/NotoSans/NotoSans-Regular.otf
Site: https://github.com/googlefonts/noto-fonts/blob/fcf8d5d5568fc96efcc4cbe6f9fbd402f62de98a/phaseIII_only/unhinted/otf/NotoSansMono/NotoSansMono-Regular.otf
Site: https://github.com/googlefonts/noto-fonts/blob/fcf8d5d5568fc96efcc4cbe6f9fbd402f62de98a/phaseIII_only/unhinted/otf/NotoSerif/NotoSerif-Regular.otf
Date: 2020-05-01

Font version

Version 2.003

Issue

In Chinantec (language system tag CCHN) the modifier letters U+02C8..U+02CB should be made consistent in length, vertical position, and stroke width with U+A717..U+A719. According to L2/04-349R, they are all part of the same series of tone marks.

Character data

ˈˉˊˋꜗꜘꜙ
U+02C8 MODIFIER LETTER VERTICAL LINE
U+02C9 MODIFIER LETTER MACRON
U+02CA MODIFIER LETTER ACUTE ACCENT
U+02CB MODIFIER LETTER GRAVE ACCENT
U+A717 MODIFIER LETTER DOT VERTICAL BAR
U+A718 MODIFIER LETTER DOT SLASH
U+A719 MODIFIER LETTER DOT HORIZONTAL BAR

Screenshot

[nizarsq]

I can see the vertical position of U+02C8 is lower and U+A717 is higher comparing with the rest, I also can see U+02CA and U+02CB length are not matching the rest, I'm not sure though if this the issue reported.
Now the question is, what is the right way to test those tones? will a sequence of standalone tones (modifier) is sufficient to tell if there is an issue? or we need them together with Chinantec script? I copied some text that have tones from this pdf to generate test case to be able to see if they are inconsistent, but unfortunately I can't tell.
@marekjez86, maybe language/script expert opinion will be helpful on this issue.

[dscorbett]

You either need to set the language system tag to CCHN (as I did) or set the HTML lang tag to a Chinantec language, e.g. lang="cco". Either way, the inconsistency will be evident.

[nizarsq]

You either need to set the language system tag to CCHN (as I did) or set the HTML lang tag to a Chinantec language, e.g. lang="cco". Either way, the inconsistency will be evident.

In the screenshots I attached (previous comment), lang tag is "cco". Could you test it again? Latin Greek Cyrillic fonts updated recently.

[moyogo]

There’s no CCHN language system tag in the sources or in the new TTF.

[dscorbett]

Exactly. @nizarsq, I’m not sure why you want me to retest this: the fonts cannot possibly support this properly without CCHN language system tags, and your screenshot clearly shows that the tone marks are still inconsistent.

[nizarsq]

Exactly. @nizarsq, I’m not sure why you want me to retest this: the fonts cannot possibly support this properly without CCHN language system tags, and your screenshot clearly shows that the tone marks are still inconsistent.
code.

There are 2 reasons why I asked to retest:
1- to see how tone marks are displayed with actual text (Not stack of tone marks), this will give more realistic result
2- to test tone marks display using the updated font

In regard to language code, I confirmed that I used the language code cco.

[moyogo]

@dscorbett This cannot be resolved properly with OpenType as it currently defines the Chinantec Language System.

The OpenType Language system CCHN is for Chinantec CCHN is for the ISO 639-3 language codes cco, chj, chq, chz, cle, cnl, cnt, cpa, csa, cso, cte, ctl, cuc, cvn.

Ozumacín Chinantec [chz] is documented as using the modifier letters U+02C8..U+02CB with U+A717..U+A719, and as seen in the L2/04-349R, the Ozumacín Chinantec documents it references or other documents, they are made consistent in length, vertical position, and stroke width there.

But other Chinantec languages do not follow the same style for U+02C8..U+02CB and they do not use U+A717..U+A719.
For example Comaltepec Chinantec [cco], uses U+02C9..U+02CB along with U+02C6..U+02C7, U+02DC.

If fonts were to have U+02C8..U+02CB designed to be consistent in length, vertical position, and stroke width, then they would not be consistent in length, vertical position, and stroke width with U+02C6..U+02C7.

It would be probably be best to offer stylistic alternates for U+02C8..U+02CB to match U+A717..U+A719 rather than having a locl feature that applies to all Chinantec languages.

[dscorbett]

In that case 'CHZ ' should probably be registered as a language system tag. For now, I agree a stylistic set is the best solution.