bugUnifont - Bugs: bug #51642, Problems with Devanagari

 
 

You are not allowed to post comments on this tracker with your current authentication level.

bug #51642: Problems with Devanagari

Submitter:  David Corbett <dscorbett>
Submitted:  Wed 02 Aug 2017 01:37:18 AM UTC
   
 
Category:  None Severity:  3 - Normal
Item Group:  None Status:  Fixed
Privacy:  Public Assigned to:  unifoundry
Open/Closed:  Closed

Jump to the original submission

Sat 01 Jun 2019 06:01:52 PM UTC, comment #8: 

I have redrawn the Devanagari and Bengali glyphs with narrower consonants in the just-uploaded Unifont 12.1.02, so that vowel should appear better when superimposed.  I'm not sure if anyone is actually using these glyphs to write Hindi, Sanskrit, Bengali, etc., so the old glyphs are preserved in the font/plane00/alt/uni0009-orig.hex file -- they look much better in the FontForge sample squares that are on top of each code point.  I went over the Bengali glyphs with someone whose native language is Bengali.


Paul Hardy

Paul Hardy <unifoundry>
Group administrator
Sat 13 Oct 2018 05:32:20 PM UTC, comment #7: 

I am going to redraw the Devanagari and Bengali glyphs.  I will move the original, beautiful glyphs to the font/plane00/alt directory so they will be preserved.  The only case where this might be useful is if someone wants a BDF or PCF font for a system that does not support TrueType fonts, yet which allows for overlappping glyphs (unlike the original X Window System BDF font handling).

As for TrueType fonts, Devanagari and Bengali have many glyph substitutions that a native speaker/writer would expect to exist in a font, so I cannot see anyone wanting to use Unifont for those scripts when better free fonts are available for them (including GNU Free Font).

Thanks,


Paul Hardy

Paul Hardy <unifoundry>
Group administrator
Fri 22 Sep 2017 06:44:27 AM UTC, comment #6: 

David,

I have not yet. Devanagari was one of the earliest contributions to Unifont, and I am reluctant to just discard the hard work of one of the first contributors. I would rather adjust what is there. I have tried to reach him but so far I have not received a reply email (same with Tibetan).


Paul Hardy

Paul Hardy <unifoundry>
Group administrator
Thu 21 Sep 2017 06:19:31 PM UTC, comment #5: 

Have you considered Devanagari further?

I think that Devanagari, Bengali, Newa, and Tirhuta need to be redrawn. The base glyphs are too big and bold, which doesn’t harmonize with most of Unifont, and more importantly doesn’t leave room for combining marks.

David Corbett <dscorbett>
Sun 13 Aug 2017 11:07:19 PM UTC, comment #4: 

David,

I have gone through all of your Devanagari suggestions. I need to spend more time on them. Here are some thoughts in the meantime.

U+0903 and U+1CE2..U+1CE8 (visargas) are combining characters. I know what you mean about them not being rendered properly. I might add yet another file for Unicode combining glyphs that will not be encoded as combining in Unifont, or give them a positive offset, or do something else. I need to think about handling them one of those ways versus some other possible approach.

On Vedic Extensions in general, I need to consider how Vedic chanting marks combine with vowel marks in Unifont. For example, U+0951 DEVANAGARI STRESS SIGN UDATTA is a Vedic chanting mark that should appear on top of a right-side vertical stroke; any combining vowels (such as U+0945..U+0948) can appear just to the left of the udatta, but it isn't the end of the world if it is offset a little to the right. When writing by hand (or doing real typesetting), the udatta would be written right on top of any center vertical stroke in letters such as U+0915 KA, but that is too much for Unifont.

Making all of the chandras have the same shape is definitely a good idea.

The Vedic Extensions block and Devanagari Extended block came along after the main Devanagari block. I never spent sufficient time considering how the new additions combined with Devanagari when rendering Vedic texts. Part of that is if someone was interested in representing Sanskrit, I recommended the Sanskrit font from Omkarananda Ashram to them. It is an outstanding outline font, and Unifont will never compare.

I don't know if I can work through everything by the next release.

I will be on travel next weekend, and am hoping to issue a release in the not-too-distant future.

Thank you,


Paul Hardy

Paul Hardy <unifoundry>
Group administrator
Fri 04 Aug 2017 01:02:48 PM UTC, comment #3: 

In that case, you could add blank columns to the sides of the visarga and its diacritics so they are 16 pixels wide.

David Corbett <dscorbett>
Fri 04 Aug 2017 02:22:31 AM UTC, comment #2: 

David,

I will follow up in greater detail on this later. A quick note though: I would like to keep all the Indic combining characters as being 16 pixels wide. The Indic script ranges are hard-coded in unibmp2hex to be 16 pixels wide (for better or worse). I think it will complicate things for me if I have to remember which Indic combining characters are 8 pixels wide and which are not.

Thank you,


Paul Hardy

Paul Hardy <unifoundry>
Group administrator
Thu 03 Aug 2017 12:21:53 AM UTC, comment #1: 

U+1CF4 VEDIC TONE CANDRA ABOVE should look like the other candras.

U+0903 DEVANAGARI SIGN VISARGA should not have a combining glyph. Currently, it often overlaps its base glyph. There is not enough room for the visarga diacritics U+1CE2 to U+1CE8. My suggested glyph does not give it a headstroke, because in most of the fonts I checked, it does not have one. My suggested glyph requires its row to be removed from plane00-combining.txt. This does mean it would not get a dotted circle in the dotted circle version of the font.

All the visarga diacritics are too far left. My suggested glyphs require their offsets to be changed to -8.

0903:00000000103810000000103810000000
1CE2:00000000000000007E00000000000000
1CE3:00001008040202023C00000000000000
1CE4:00001020408080807800000000000000
1CE5:00000000000000007880808040201000
1CE6:00000000000000003C02020204081000
1CE7:00060908040202023C00000000000000
1CE8:00000000000000007880808040209060
1CF4:004C003C00000000000000000000000000000000000000000000000000000000

David Corbett <dscorbett>
Wed 02 Aug 2017 01:37:18 AM UTC, original submission:  

U+090D DEVANAGARI LETTER CANDRA E, U+0911 DEVANAGARI LETTER CANDRA O, U+0945 DEVANAGARI VOWEL SIGN CANDRA E, U+0949 DEVANAGARI VOWEL SIGN CANDRA O, U+0955 DEVANAGARI VOWEL SIGN CANDRA LONG E, and U+0972 DEVANAGARI LETTER CANDRA A render the candra differently. My suggested glyphs standardize on that of U+0911, except for U+0955, where I omitted one pixel for clarity.

The vowel signs above (U+093A, U+0945, U+0946, U+0947, U+0948, U+0955) should be aligned with the vertical stroke of most consonants.

The vertical strokes of vowel signs to the right (U+093B, U+093E, U+093F, U+0940, U+0949, U+094A, U+094C, U+094E) should be the same length as the vertical stroke of most consonants.

The headstroke of U+094F DEVANAGARI VOWEL SIGN AW is too low.

U+0971 DEVANAGARI SIGN HIGH SPACING DOT should be at the height of the headstroke. There is no reason for it to be 2 cells wide.

U+0976 DEVANAGARI LETTER UE and U+0977 DEVANAGARI LETTER UUE have an extraneous candra.

There is no reason for U+1CE1 VEDIC TONE ATHARVAVEDIC INDEPENDENT SVARITA, U+1CF2 VEDIC SIGN ARDHAVISARGA, U+1CF3 VEDIC SIGN ROTATED ARDHAVISARGA, and U+1CF7 VEDIC SIGN ATIKRAMA to be combining characters. They overlap their bases illegibly and U+1CF3 is cramped. They should be 1 cell wide and should not have combining.txt entries.

U+A8FA DEVANAGARI CARET is a “zero-advance character centered on the point between two orthographic syllables”; that is, it should be a combining character with an offset of -8. It is also too low: it should balance on the headstroke.

U+A8FB DEVANAGARI HEADSTROKE should not join adjacent headstrokes. This is to make it obvious how many headstrokes there are in a lengthy lacuna.

090D:009800780000FFFF0C300C300C300C300C600400020001800060001000180018
093A:0018001800180018000000000000000000000000000000000000000000000000
093B:000C000C000C003F000C000C000C000C000C000C000C000C000C000C00040000
093E:000000000000003F000C000C000C000C000C000C000C000C000C000C00040000
093F:3C0062003100FC00300030003000300030003000300030003000300010000000
0940:00F001080108003F000C000C000C000C000C000C000C000C000C000C00040000
0945:004C003C00000000000000000000000000000000000000000000000000000000
0946:020001F000080000000000000000000000000000000000000000000000000000
0947:00E0001000080000000000000000000000000000000000000000000000000000
0948:00E001D000280000000000000000000000000000000000000000000000000000
0949:0026001E0000003F000C000C000C000C000C000C000C000C000C000C00040000
094A:008000780004003F000C000C000C000C000C000C000C000C000C000C00040000
094B:00F000080004003F000C000C000C000C000C000C000C000C000C000C00040000
094C:00F001E80014003F000C000C000C000C000C000C000C000C000C000C00040000
094E:0000000000003C00300030003000300030003000300030003000300010000000
094F:010400F2000A00FF000600060006000600060006000600060006000600020000
0955:004C0038007C0000000000000000000000000000000000000000000000000000
0971:00000018180000000000000000000000
0972:013000F000001CFF2260026006601C60046043E0436027601E60006000200000
0976:0000000000001CFF2260026006601C60046043E0436027601E60002010800F00
0977:0000000000001CFF2260026006601C60046043E027601E6010E00F0020401F80
1CE1:0C1211101010101010101010101010F0
1CF2:000000000000423C003C420000000000
1CF3:00000000000042242424244200000000
1CF7:06081010101010101010101020202000
A8FA:0220014000800000000000000000000000000000000000000000000000000000
A8FB:0000000000003FFC000000000000000000000000000000000000000000000000

The glyphs of U+0936 DEVANAGARI LETTER SHA and U+096F DEVANAGARI DIGIT NINE are not the usual glyphs. They are not wrong and don’t need changing! I only mention it in case it was unintentional. U+A8E9 COMBINING DEVANAGARI DIGIT NINE is inconsistent with U+096F.

David Corbett <dscorbett>

 

(Note: upload size limit is set to 16384 kB, after insertion of the required escape characters.)

No files currently attached

 

Depends on the following items: None found

Items that depend on this one: None found

 

Carbon-Copy List
  • -email is unavailable- added by unifoundry (Posted a comment)
  • -email is unavailable- added by dscorbett (Submitted the item)
  •  

    There are 0 votes so far. Votes easily highlight which items people would like to see resolved in priority, independently of the priority of the item set by tracker managers.

     

    Follow 3 latest changes.

    Date Changed by Updated Field Previous Value => Replaced by
    2019-06-01 unifoundry StatusNone Fixed
        Assigned toNone unifoundry
        Open/ClosedOpen Closed

    Back to the top

    Powered by Savane 3.13-d3ae.
    Corresponding source code