Latin/IPA vertical bars design issues

moyogo commented 8 years ago

The IPA uses the rhythm punctuation marks | U+007C and ‖ U+2016 for minor and major foot break (a break in a sentence and the end of a sentence), the IPA and orthographies use the Latin letters ǀ U+01C0, ǂ U+01C2, and ǁ U+01C1 to transcribe dental, palatal, and lateral clicks. The rhythm punctuation marks should be distinct from the letters since they can be used at the same in the same text. They should also be distinct from the lowercase letter L.

Sometimes Latin letters ǀ U+01C0, ǂ U+01C2, and ǁ U+01C1 have a diagonal stem instead of a vertical stem in roman, due to a long history of /, //, ≠ being used for them. (As a side note, this long history led another click letter, ǃ U+01C3, to look like ! U+0021 when it was originally a vertical bar with a dot below.)

Because ǀ U+01C0 cannot look like the rhythm punctuation marks and lowercase L, there isn’t much room distinct designs. One solution is to have there stems narrower than the stem of lowercase L (as is done in Tinos, Noto Sans, Noto Serif). Another solution, if their stem is the same as that of lowercase L (this occurs when hinted at small sizes anyway), would be to play with the height of their top and bottom, making them shorter than ǀ U+01C0 but longer than l U+006C.

For example this snippet from Helen Eaton (2006). “Sandawe”. Journal of the International Phonetic Association, 36, pp 235-242 doi:10.1017/S0025100306002647 where the rhythm marks and the letters are clearly distinct. screen shot 2015-09-05 at 20 52 55

See the following with Arimo, Cousine, Tinos, Noto Sans, Noto Serif (in that order) in roman and then in italic: a | a ‖ aǀaǂaǁala (U+0061 U+0020 U+007C U+0020 U+0061 U+0020 U+2016 U+0020 U+0061 U+01C0 U+0061 U+01C2 U+0061 U+01C1 U+0061 U+006C U+0061) screen shot 2015-09-05 at 20 34 17

In Arimo:

rhythm mark ‖ U+2016 should match the design of the vertical bar punctuation mark | U+007C and descend below the baseline.
Latin letters ǀ U+01C0, ǂ U+01C2, and ǁ U+01C1 should be slanted in italic.
Latin letters ǀ U+01C0 should be distinct from l U+006C (lowercase L), ǂ U+01C2, and ǁ U+01C1 should follow its design features.

In Cousine:

‖ U+2016 is missing.
Latin letters ǀ U+01C0, ǂ U+01C2, and ǁ U+01C1 should be slanted in italic.

In Tinos:

rhythm mark ‖ U+2016 should match the design of the vertical bar punctuation mark | U+007C and descend below the baseline.
Latin letters ǀ U+01C0, ǂ U+01C2, and ǁ U+01C1 should be slanted in italic.

In Noto Sans:

rhythm mark ‖ U+2016 is alright.
Latin letters ǀ U+01C0, ǂ U+01C2, and ǁ U+01C1 should be clearly distinct from rhythm marks | U+007C and ‖ U+2016
Latin letters ǀ U+01C0, ǂ U+01C2, and ǁ U+01C1 should be slanted in italic.

In Noto Serif:

rhythm mark ‖ U+2016 is alright.
Latin letters ǀ U+01C0, ǂ U+01C2, and ǁ U+01C1 should be clearly distinct from rhythm marks | U+007C and ‖ U+2016

marekjez86 commented 8 years ago

This bug is from now on used ONLY for Noto issues. For Arimo, Cousine or Tinos issues please see https://github.com/googlei18n/noto-fonts/issues/686 .

Since I created bug https://github.com/googlei18n/noto-fonts/issues/686 to track Arimo, Cousine or Tinos issues, I'm removing the ArimoTinosCousine label here.

TidharC commented 7 years ago

The glyphs look generally OK. I only see a problem with the Noto Serif Display fonts, Black and Italic; the special char glyphs U+01C0 U+01C1 U+01C2 U+007C U+2016 are a much lighter weight than letters, such as "a" or "l", in the same font.

specimenNotoSans-Black.pdf specimenNotoSansDisplay-Black.pdf specimenNotoSansDisplay-Italic.pdf specimenNotoSans-Italic.pdf specimenNotoSerif-Black.pdf specimenNotoSerifDisplay-Black.pdf specimenNotoSerifDisplay-Italic.pdf specimenNotoSerif-Italic.pdf

marekjez86 commented 7 years ago

<html lang="en-phonipa"><p><span style="font-size:80.0px">a | a ‖ aǀaǂaǁala</span></p></html>

en-504-NotoSans-Regular en-504-Roboto-Regular en-504-NotoSans-Italic en-504-Roboto-Italic en-504-NotoSerif-Regular en-504-NotoSerif-Italic

en-504-NotoSans-Regular.pdf en-504-NotoSansDisplay-Black.pdf en-504-NotoSansDisplay-Italic.pdf en-504-NotoSansDisplay-Regular.pdf en-504-NotoSansMono-Black.pdf en-504-NotoSansMono-Regular.pdf en-504-NotoSerif-Black.pdf en-504-NotoSerif-Italic.pdf en-504-NotoSerif-Regular.pdf en-504-NotoSerifDisplay-Black.pdf en-504-NotoSerifDisplay-Italic.pdf en-504-NotoSerifDisplay-Regular.pdf en-504-Roboto-Black.pdf en-504-Roboto-Italic.pdf en-504-Roboto-Regular.pdf en-504-NotoSans-Black.pdf en-504-NotoSans-Italic.pdf

marekjez86 commented 7 years ago

@moyogo : I included Roboto because both Noto Sans and Serif was supposed to "go well" with Roboto - we want to Roboto look as good as Noto :-) . However, if there are problems with Roboto please file them in https://github.com/google/roboto/issues

moyogo commented 7 years ago

capture d ecran 2017-01-02 a 06 47 21 It’s still hard to see which of ǀ and l is which in Noto Sans and Roboto.

marekjez86 commented 7 years ago

note from MTI: "change glyphs uni01C0, uni01C1, and uni01C2"

nizarsq commented 3 years ago

Still hard to tell which is U+01C0 (ǀ) and which is U+006C (l) in NotoSans(Regular and Italic). The rest of the issue in NotoSans and NotoSerif looks ok now.

notofonts / latin-greek-cyrillic

Latin/IPA vertical bars design issues #177