Do you know any app online that can list the languages covered in the character set of a given font?

Options
Laura Meseguer
Laura Meseguer Posts: 11
edited June 3 in Font Technology
Do you know any app online that can list the languages covered in the character set of a given font? 

Thanks! 


Answers

  • Laura Meseguer
    Laura Meseguer Posts: 11
    I just found FontDrop! https://fontdrop.info/ :) 
  • Jens Kutilek
    Jens Kutilek Posts: 377
    There are a few ...
    https://hyperglot.rosettatype.com/
    https://googlefonts.github.io/shaperglot/
  • Andreas Stötzner
    Andreas Stötzner Posts: 812
    The proprietary Fonts management program on the Mac shows you which languages might be supported by a given font.
  • Giuseppe Salerno
    Giuseppe Salerno Posts: 4
    I also recommend https://fontanalyzer.app/
    by Jan Charvat
  • Denis Moyogo Jacquerye
    Denis Moyogo Jacquerye Posts: 220
    Those tools are only as good as the work put into them.
    Which means they shine at a few things but do poorly at others.
    Not pointing fingers, that’s just inherent to the tools and the data they use.

    For example, most tools incorrectly report on Malagasy since the data they’ve used is incorrect in the first place but coming from a trusted source.

    In the case of Ă, it’s interesting as "Călin" has been a lot in the news lately. But I totally agree, "auxiliary" is messy, not well defined and too often unreliable.

    Shaperglot is outstanding and detected 716. It also informs what is lacking in a font to support any given language. But it fails in several combinations of diacritics and base letters that are actually available, what causes the difference I counted
    Do you have an example?

  • Yves Michel
    Yves Michel Posts: 206
    What about the Bulletproof site Bulletproof Font Tester ? I found there the missing characters in my fonts (per language). Very useful!
  • Kent Lew
    Kent Lew Posts: 1,000
    I’ve always been wary of lists of “language support”. Especially in marketing contexts. Definitions can be fuzzy. One person’s auxiliary character is another’s required. One person’s language is another person’s dialect.
    (I always found Hoefler’s list to be artificially inflated — perhaps to appear more comprehensive, even though supporting the exact same character set as others).
    To me it is always more useful to display the entire character set repertoire and provide testers and let users determine if everything they need for their language, as they understand it, is present. 
    On the other hand, in a corporate environment, designers/licensors only get told “we need to support X, Y, and Z languages”, and so rely on such lists for better or worse.
    As Denis states, the various tools available to assemble such lists from a font’s contents will vary in their comprehensiveness, accuracy, and reliability — depending on the scope and quality of the source data used. (Data which can vary even within linguistic communities!)


  • John Hudson
    John Hudson Posts: 3,434
    I’ve always been wary of lists of “language support”. Especially in marketing contexts.
    Agreed. We’ve relegated language lists to our glyph set support documents, but I understand why a distributor site might encourage their use, because it provides a search vector to find fonts that support — or claim to support — a given language.

    The biggest problem with language lists is that they’re usually just a big text string, and what gets listed is one particular foundry or list tool’s set of names spelled in a particular way. It also is most often limited to English language exonyms.

    A better system would refer to an external database that includes variant spellings, endonyms, localised names in different languages, ISO tag mappings, etc.. This could be linked to something like Shaperglot and implemented on a foundry site dynamically, such that as new information is added to the database or as fonts are extended and updated, the information on the site is automatically updated. This would enable a potential customer from anywhere in the world to search for a font that supports language X under whatever name and in whaterver script he or she normally uses, rather than needing to know what the language is called in English and as written in Latin script.
  • Simon Cozens
    Simon Cozens Posts: 792
    Igor Freiberger said:
    But it fails in several combinations of diacritics and base letters that are actually available, what causes the difference I counted. Probably some minor issue reading the mark and mkmk tables from the font.

    Given how it works, this is unlikely - but if you can give me some more details, or ideally file an issue, I can get it fixed.
  • Igor Freiberger
    Igor Freiberger Posts: 293
    Simon and Denis, thanks for the questions. Due to your asks, I was able to identify there was some problem in my side when generating <mkmk> tables. After a new, revised font export, I got even more languages than I have catalogued:


  • mitradranirban
    mitradranirban Posts: 73
    Shaperglot also gives false positive results. I just added Tamil code points but no Tamil OT table but it says Tamil is 100% supported, whereas the vowel marks are not aligned properly yet. 
  • Thomas Phinney
    Thomas Phinney Posts: 3,013
    Indeed, there is a difference between checking for language support in terms of codepoints, versus checking the font total functionality for that language.

    I can understand how, with a name like Shaperglot, you are expecting more overall “shaping” testing.
  • Dave Crossland
    Dave Crossland Posts: 1,475
    mitradranirban said:
    Shaperglot also gives false positive results. I just added Tamil code points but no Tamil OT table but it says Tamil is 100% supported, whereas the vowel marks are not aligned properly yet. 
    Hmm. I mean, the shaperglot concept is to go beyond simple list of unicode characters - for which my pyfontaine project was I thought the most comprehensive until hyperglot, mostly thanks to Thomas Phinney's work at Extensis WebINK, some 15 years ago - and to also additionally account for open type features required to support a language. 

    But shaperglot is still rather early in its lifecycle of developing such definitions, and I rather suspect that there are no shaping definitions for Tamil, only inherited unicode character definitions. 

    The Turkish dotless i opentype feature is a classic failure case for the unicode based checkers, where glyphsapp autogenerated the opentype code necessary, but many fonts without that code but with the necessary but insufficient Unicode characters would report Turkish as supported. That language has the additional feature definition. 

    I suppose that shaperglot could get a quick boost from simply upcycling Noto fonts' shaping data into language shaping requirements....
  • Simon Cozens
    Simon Cozens Posts: 792
    Indeed, I had a meeting about that yesterday. :-) We are planning to analyse the requirements for Indic languages later in the year.

Categories