RE: Phetsarat font, Lao unicode

From: Kenneth Whistler (
Date: Mon Jul 09 2007 - 14:11:07 CDT

  • Next message: Magda Danish (Unicode): "FW: that font-letter spacing problem - MS Word issue?"

    Kent said:

    > Likewise, reordrant vowels **should** reorder around any base
    > character. There is no need to add any characters for this, but
    > display systems may need to be updated.

    And I'm going to disagree with that assessment. It is
    certainly debatable what left-side Indic vowels should
    do when no "appropriate" base character is present, and
    there is a lot of wiggle room in determining just exactly
    what "appropriate" would mean.

    I, for one, think that reordering left-side Indic vowels
    around arbitrary base characters would be a confusing and
    problematical fallback behavior, and that the appropriate
    scope for full Indic rendering behavior (left-side reordering,
    reph formation, conjunct formation, other kinds of stacking,
    and so on) is *within* a script. So I wouldn't advocate
    even having U+09BF BENGALI VOWEL SIGN I, for example,
    reordering left around U+0915 DEVANAGARI LETTER KA. Because those
    are two different scripts, each with complex display behavior,
    it is expecting too much of rendering engines, I think, to
    insist that they mix such scripts together (even when closely
    related) and have display just work out o.k.

    Unicode *is* designed for multilingual text, where multiple
    scripts can coexist together. But I think it is too much --
    and goes beyond what is reasonable to ask -- for characters
    from different scripts to be jumbled together and get
    "correct" rendering (whatever that might mean), as opposed
    to mixing words or lines of text which internally are
    coherent instantiations of single script data.

    So I agree with John Hudson -- if there is a particular
    Latin letter (or symbol) that Lao vowels need to be
    displayed on for particular dictionary or pedagogical
    purposes, then that should be treated as a special case,
    rather than expecting such behavior as the general rule.


    > /kent k

    This archive was generated by hypermail 2.1.5 : Mon Jul 09 2007 - 14:16:52 CDT