Re: which scripts are written vertically

From: Philippe Verdy (
Date: Mon May 16 2011 - 18:49:19 CDT

  • Next message: fantasai: "Re: which scripts are written vertically"

    2011/5/17 fantasai <>:
    > I'd drop the "using digital layout" part but otherwise, yes, that's a
    > good definition. If we had such a property it would need to state for
    > each script:
    >  - what vertical directionality it has (if any). Horizontal scripts
    >    would be none. Most vertical scripts would be top-to-bottom. Going
    >    by Michael Everson's answer, Ogham would be bottom-to-top.
    >  - If it has vertical directionality, how the text is transformed from
    >    horizontal to vertical, i.e. are grapheme clusters rotated (laid
    >    sideways wrt horizontal) or translated (kept upright like CJK).

    For CJK, it's actually a bit more complex : sinograms are typically
    not rotated, but *some* CJK punctuations are rotated (notably
    parentheses, brackets, braces, and similar). Additionally, even in the
    vertical layout, some punctuations may still be laid out horizontally
    as if they were attached to the sinogram.

    Finally, there's not just the problem of rotation or not, but
    conversion of metrics and alignment: the typical regular "grid"
    observed in the horizontal layout is not always followed in the
    vertical layout where characters are tiled possibly with variable
    height. The glyph itself also often has to be centered horizontally or
    moved vertically (e.g. dots and commas), implying an additional shift.
    So for the same apparent glyph, even without a rotation, there's a
    distinct positioning between the horizontal and vertical layout.

    A font capable of rendering CJK vertically will contain additional
    metrics adjustments for the vertical layout, and some glyphs rotated,
    and quite often redrawn to take into account the variable weight of
    strokes or simply the drawing position of the painting brush.

    Most of this complexity however is avoided in the most numeous
    characters, i.e. sinograms, but not for more basic subsets like

    Interlinear annotations (uncluding ruby) will also use a distinct
    positioning and orientation between the two layouts.

    Note: I largely prefer the term "sinogram" to the term "ideograph"
    used in Unicode and ISO/IEC 10646, because the latter is misleading as
    in fact very few sinograms are ideographs, or even phonograms, but a
    combination of both types, or change their role from one type to the
    other depending on the text composition ; I've seen also "Han
    ideographs" to designate the script, but I much prefer the correct
    expression "sinographic script", instead of even the "CJK script"
    expression which is in fact partly wrong and would cover too many
    things that are used in Chinese, Japanese, Korean and other related
    (but non-Han !) languages. But now it's too late to discuss a change
    of this terminology in the stable encoded character names and block
    names. There does exist some pure ideographic scripts, however this is
    definitely not for Chinese at least (except possibly in very old
    eras), but for many other modern uses (where their existence has now
    exploded and is in constant growth in lots of real-life domains).

    This archive was generated by hypermail 2.1.5 : Mon May 16 2011 - 18:51:28 CDT