Need a mapping of glyph coverage
The system needs a key of each icon in the glyph coverage and how it is represented in the database.
- Open Font Library Developers
- Dave Crossland
- Christopher Adams
- Series goal:
- Accepted for trunk
- Milestone target:
- Started by
- Christopher Adams on 2011-03-17
- Completed by
- Christopher Adams on 2011-03-31
The input to the database is all provided by fontaine: http://
The relevant tables and values are:
font_metadata: id, file_id, file_name, upload_user, date, common_name, native_name, sub_family, style, weight, fixed_width, fixed_size, copyright, license, license_url, glyph_count, character_count, fontaine_report
orthographies: id, font_id, common_name, native_name, support_levelchar, percent_coverage, missing_values
Fontaine will check the font's glyph coverage for many language categories such as Western Latin, Chinese, Euro, etc.
It will then indicate whether the support for each language is Full, Partial (>80%), Fragmentary or None. For fonts with Partial or Fragmentary support, Fontaine will list the missing glyphs.
The purpose of this feature is to not only list what languages are supported, but to show other type designers what work still needs to be done to give the font better glyph coverage.
However, I am trying to weigh the utility of displaying missing glyphs for orthographies with fragmentary support.
This is why:
According to Fontaine, Open Baskerville has full support for Basic Latin and partial support for Western Latin. For the latter, the support level is 96% and the missing glyphs are U+00c7 (Ç), U+00e7 (ç).
Excellent, I think! Now some budding type designer can help push Open Baskerville to 100% glyph coverage for Western Latin by drawing a nice cedilla and hanging it on two glyphs!
But now we turn to Open Baskerville's support for Pan African Latin, which stands at 19%. The task of filling in the other 81% might not seem so daunting until we realize it calls for another 194 glyphs!
From the perspective of usability, what we are confronted with in this last case is a great deal of data of eminently little value (even to other type designers). The original OFLB avoided this problem by hiding exact glyph coverage inside a drop-down widget. But in my mind this is merely sweeping the data under the proverbial rug.
I expect that many of you will react with, "If we have all that data on hand, then we should show it." I want to counter with the proposal that we can afford to be a bit more stringent, and only display missing glyphs for Partial orthographies.
Not to put too fine a point on it, but many fewer additional glyphs are needed to give Open Baskerville full Dutch support than Pan African Latin support. What is the argument for not displaying those, while displaying the Pan African ones?