Transliteration into Latin characters of Cyrillic characters

michele casanova
michele casanova Posts: 79
in Type Design Technique & Theory
Is it reasonable to include two character variations (cvXX) in a font to transliterate Latin characters into Cyrillic and vice versa? Or is it just a waste of time?
I know there are numerous variations, but the ISO 9 table appears to be bidirectional; there seems to be only a case issue for ъ and ь.
Tagged:

Comments

  • John Hudson
    John Hudson Posts: 3,603
    Yes, a waste of time for the reasons Tural identifies, and also contrary to the general character/glyph distinction.
  • Nick Shinn
    Nick Shinn Posts: 2,328
    edited February 10
    I made a transliteration typeface, Auld English Spell. It’s very language specific and flouts the character/glyph distinction.
    For instance, “life of the party” transliterates to “lyfe of the partie.”
    (I also included some translation, such as “before” becoming “ere,” but that’s another story.)

    If it is clearly identified for what it is, and users are aware of the nature of the beast, I don’t see anything wrong with going down this intriguing path and seeing where you end up.
    Who knows what you might discover?!
    Certainly, there will be “fails,” but that may be acceptable—like smart quotes, for instance.

    What is the history of such Cyrillic ↔ Latin automation?
    Has it been previously attempted in a font? 
    Are there online apps that do it?
    Sounds like something AI could do. 
  • John Hudson
    John Hudson Posts: 3,603
    edited February 10
    Are there online apps that do it?
    There are lots of online transliteration tools, many offering a choice of different transliteration systems or standards. A lot are language or script specific, e.g. the Anbani tool to switch text between the various Georgian writing systems. One of my favourite more general tools, which I use very often, is the Aksharamukha Script Converter. What they all have in common is that they work in character space, not glyph space, so the output is stable, can be copied and pasted preserving both encoding and script, can be searched and sorted according to the script, and is not font dependent.
  • Nick Shinn
    Nick Shinn Posts: 2,328
    I should point out that in the transliterated “lyfe of the partie,” the y glyph is coded as i.alt, and the “ie” as y.alt—so the canonical text is not changed.
  • John Hudson
    John Hudson Posts: 3,603
    edited February 11
    Is that what the user wants, though? Maybe what they want is text in auld — or maybe olde — English spelling, and are disappointed when they copy and paste and get something different from how it looks? And what if the input text is already in something other than the modern English standard you presume? Won’t the glyph substitutions garble it?

    [To be clear, Nick, I consider your Auld English Spell font a form of play: it’s fun, and cleverly done. I think what Michele was asking about is more serious, and more problematic from a text processing perspective.]
  • michele casanova
    michele casanova Posts: 79
    Thanks for the guidance. I understand this isn't a viable solution.
    My idea was to use an international standard, but I know there are many local variations.
  • Nick Shinn
    Nick Shinn Posts: 2,328
    edited February 11
    Maybe what they want is text in auld — or maybe olde — English spelling, and are disappointed when they copy and paste and get something different from how it looks?
    Well then, they can go to a any number of web sites that do that sort of re-writing.
    The Auld English is designed to simulate the effect of reading English 400 years ago, by a variety of means, not to re-write text.
    I think what Michele was asking about is more serious, and more problematic from a text processing perspective
    I was very serious in my investigation, which was an open-ended design process. I realize Michele is addressing a “harder” issue, but in mentioning my perhaps irrelevant design I was attempting to provide some broad context and encouragement to continue, because who knows what discoveries experimenting will produce? 
  • John Hudson
    John Hudson Posts: 3,603
    ...who knows what discoveries experimenting will produce?
    We can, however, anticipate where in the digital text stack specific kinds of operations are most appropriate and where not, because the kinds of operations are known even when the specifics of individual operations are not.

Categories