bugUnifont - Bugs: bug #56781, Please provide a version WITHOUT...

 
 

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

bug #56781: Please provide a version WITHOUT Indic/Semitic scripts

Submitter:  None
Submitted:  Tue 20 Aug 2019 11:10:48 PM UTC
   
 
Category:  None Severity:  3 - Normal
Item Group:  None Status:  Invalid
Privacy:  Public Assigned to:  None
Open/Closed:  Closed

Discussion locked!

Sat 02 Jan 2021 12:38:13 PM UTC, comment #3: 

Whoever filed this has not bothered to reply although more than one year has passed.  I cannot create an additional special version of the font for one anonymous person who does not have the time to provide a response, after having been given ample time to write just a single reply.  Therefore, I am closing this.


Paul Hardy

Paul Hardy <unifoundry>
Group administrator
Tue 10 Dec 2019 09:19:57 AM UTC, comment #2: 

Hi! You can compile it from source and leave those scripts out from your version.

GNU Unifont covers the entire BMP and just because you have problems this doesn't mean that everyone should receive a font which doesn't cover BMP. Your request is destructive.

Again, please compile the font from source, you have information in the source how to do it.

Greetings

Hacker <hacker>
Sun 25 Aug 2019 09:33:40 PM UTC, comment #1: 

Dear Anonymous,

Thank you for taking the time to write such detailed correspondence.  I think it merits a detailed reply, so I will answer your major points.

I’d like to start by saying that Unifont today is intended to be the future font you refer to as a “‘fallback/last resort’ font”.  For example, that is how FontForge uses it.  Unifont was not intended to be used as a stylesheet font.  So your correspondence is actually a feature request, not a bug report.  I could consider creating a stripped-down Unicode Plane 0 font that removes complex scripts, but I would not go creating 4 or 5 other font derivatives.

On the Han ideographs, most of them are from a public domain font created by the People’s Republic of China.  Unifont took them from the Wen Quan Yi font, which is maintained by native Chinese speakers who grew up using simplified Chinese and traditional Chinese writing.  Wen Quan Yi has been the standard Chinese font on free distros for the past decade.  I do not think those glyphs can be improved upon in Unifont’s charcell bitmap format.  For that font format, I think they are ideal as a consistent, comprehensive set of glyphs.

On Indic scripts, the Government of India developed character assignments in the days when PCs were using fonts with only 256 glyphs.  The lower 128 code points were used for ASCII, and the upper 128 were for a given Indic script.  What they created was an encoding that someone could read, though without substitution glyphs for common letter combinations, the result would not be what a native speaker would write by hand.  When Unicode was developed, the Government of India simply mapped those upper blocks of 128 glyphs to Unicode ranges.  The result is something that someone can read, though the lack of complex character combinations would look unusual to a native speaker.

With Arabic’s initial, medial, and final forms for letters, a font such as Unifont is never going to produce good results.  Today, much better free fonts are available on the web for Arabic, for example those available from the Arabeyes project.  I would not attempt to duplicate, let alone improve upon, the Arabeyes efforts.

I did recently adjust the position of Hebrew cantillation marks in Unifont, but for other improvements in Semitic scripts I think a modern OpenType font is the way to go.

GNU Freefont is a free OpenType font that supports Indic and other complex scripts (though not yet all).  For that reason, I would not consider creating separate “Unifont Indic”, “Unifont Semitic”, etc. fonts with context-based glyph substitutions.

Google’s Noto font family is also a wide-ranging set licensed under the free SIL Open Font License.  It has a ways to go, however.  Just for example, I noticed there was a lack of consistency in character weights and alignments between Noto’s standard ASCII Latin letters and Noto’s archaic English letter forms when I wrote my “olden” program (http://unifoundry.com/olden/).  With my UTF-8 Greek to Beta Code program (http://unifoundry.com/unibetacode/), I noticed that the Noto Greek font needs work in correctly positioning letters and accents for proper rendering of ancient (polytonic) Greek.  Ironically, I think Unifont looks better in both cases because of its consistency in those scripts.

Even so, neither GNU Freefont nor Google’s Noto font provide good support for Vedic Sanskrit.  The best effort for Vedic Sanskrit support has been from Omkarananda Ashrama and the ITRANS project, which has been in existence for over 20 years.  They are continuing to improve upon Sanskrit fonts, to be licensed under the free SIL Open Font License.  They will probably produce better results than any other organization, given their religious interest in optimum rendering of Sanskrit.  I would not attempt to provide an alternative to what they have been doing so well for so long.  And of course, modern Indic text is certainly better in any font with character substitutions than Unifont provides.

Unifont is a BDF font and a TrueType font.  It has never had a claim of being an OpenType font.  I consider it a serious bug if any browser or OS cannot properly handle TrueType fonts (or PostScript Type 1 fonts for that matter, with the thousands of Type 1 fonts  that are available).  It is less serious if a browser or distro cannot handle BDF fonts, though such lack of support would still be a bug.

So to summarize, I can consider a feature request of creating a stripped-down Unifont variant in the future that does not contain complex scripts or scripts where the ability for character substitution (as with Han ideographs) is beneficial and available in other free fonts.  I believe that efforts to support complex scripts should be put into GNU Freefont and other such free vector fonts, not GNU Unifont.

Thank you,


Paul Hardy

Paul Hardy <unifoundry>
Group administrator
Tue 20 Aug 2019 11:10:48 PM UTC, original submission:  

Unifont is unsuitable for use in any stylesheet because of its coverage of Indic/Semitic which is completely incorrect without the REQUIRED OpenType features for contextual shapes and positioning.

I request the generation of a version WITHOUT ANY coverage of any script that requires the OpenType feature.

Instead, move the existing Indic/Semitic script into a separate "fallback/last resort" font.

As well, you should probably extract the Han ideographs, which are very poor and almost unreadable on display in this version without hinting. There are much better fonts for Han, but if you isolate the Han ideographs in a separate font, may be you'll start hinting them.

So "Unifont" should become "Unicode Alphabetic" (also including symbols), "Unifont Indic", "Unifont Semitic" (these two may also be merged into one "Unifont Complex" font), and "Unifont Ideographic".

This way you can also include in these 4 fonts the characters from supplementary planes, including historic scripts (except possibly Supplementary Ideographs, notably because it will soon add a third plane in Unicode 13.0, so at least 2 or 3 fonts will still be needed to cover all ideographs).

Finally, a 5th font may be designed for Emojis because of their colorful glyphs and specific designs (but keep the non-Emoji variants in the "Unicode Alphabetic" font is possible, in their basic monochromatic variant).

Ideally, you may use "font linking" in the base "Unicode Alphabetic" font, to reference the name of the other fonts to lookup for missing scripts.

This will make easier life for everyone and will allow many improvements including font hinting and additional variants for Alphabetic scripts without being blocked by the limit of 65535 glyphs per TrueType/Opentype font (65535 is enough to offer really full coverage of all alphabetic scripts, symbols and punctuation in a single font, or to cover all indic scripts in a separate font with the requires OpenType features, or to cover all modern Ideographs in a separate font (which can also have separate font variants fonts for Kanji, Simplified Han, and Traditional Han).

Note also that Arabic has two major styles (Naskh and Nastaliq, and the Nastaliq style may be provided in a separate font, while maintaining Nakh and other Semitic scripts like Hebrew together in the same Semitic font).

I no longer see any reason to maintain a font with the intent to cover a single plane only, but with incorrect layouts caused by the one glyph per character limitation. Today we DO need the required OpenType features.

For now Unifont is no longer a viable option, when the obvious better choice is offered with the "Noto" family and its excellent coverage, appropriate shaping, and which already covers all scripts encoded up to Unicode 10.0 (plus one in Unicode 11.0) and further scripts coming soon (along with the continued support of all extensions of existing scripts up to Unicode 12.0).

"Unifont" is no longer recommandable, and in fact it's best today to uninstall this font which causes more damages than helping when it breaks ALL Semitic and Indic scripts and so poorly renders Ideographs, compared to all other installable free fonts and other preinstalled fonts.

If you don't make this split, the "Unifont" project will soon die as nobody will ever want to use it (and probably soon, the font will even be rejected completely by browsers if it does not have the required OpenType features for the scripts it is "supposed" to support in its mapping, and probably some OS update will later cleanup and erase this broken font from our systems).

One good way to renew the interest in this project is to continue providing the existing "Unifont" font but only as a legacy (with its existing limitations), but distribute and develop/improve separately the 4 separate fonts without the strong limitation that can then be avoided): existing users may continue using "Unifont", but new users will prefer using the separate 4 fonts that will finally be correct for Indic and Semitic scripts. This will also be simpler to update, will use less memory in browsers (in most pages) and will allow web site designers to reintroduce the "Unifont *" 4 families in their stylesheets without fearing a very erroneous representation of their text.

Anonymous

 

(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 hacker (Posted a comment)
  • -email is unavailable- added by unifoundry (Posted a comment)
  •  

    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
    2021-01-02 unifoundry StatusNone Invalid
        Open/ClosedOpen Closed
        Discussion LockNone Locked

    Back to the top

    Powered by Savane 3.13-4b48.
    Corresponding source code