Re: Representative glyphs for combining kannada signs

From: Vinod Kumar (
Date: Thu Mar 16 2006 - 22:16:58 CST

  • Next message: Peter Constable: "RE: Representative glyphs for combining kannada signs"

    On 3/15/06, Philippe Verdy <> wrote:

    > So the "representative glyphs" of the following dependant vowel signs are not very significant: the dotted circle below them should be decentered to the left, the glyph for the sign being shown attached to its upper right corner:

    The issue here is how accurate the positioning of the vowel signs
    should be. In a representative description, placing the sign 'above'
    the consonant is good enough rather than a more accurate but more
    detailed 'top right corner' specification. The tradeoff arises even in
    Devanagari. The vowel sign E (U+947> will appear centered over centre
    stemed shapes (like Ka U+915), but will be towards right for
    consonants with a right stem (like Kha U+916).

    So I feel that the positioning of the signs need not be specified much
    more accurately than just 'top', 'bottom', 'left' or 'right'.

    In most fonts, the glyph corresponding to a sign, will be placed in
    the EM frame such that when a consonant and the sign are rendered one
    after another, the sign will reasonably be ok. Please see a Unicode
    code chart for Kannada KannadaVisualSylCh9.pdf, produced for IndiX
    using the RaghuKannada font at:

     Whatever Phillipe Verdy has asked for (positioning of vowel signs) is
    present in this chart. When the chart is not made from a real font but
    put together as bits and pieces then such positioning may not be
    present. That is why TUS4 code chart for Kannada shows these signs
    centered over the dotted circle.

    > Note that the representative glyph of the virama/halant sign uses an exagerated horizontal stroke all above the dotted circle. Keeping only the little 6-shaped curl attached on the right should be enough.

    When the virama is attached to a consonant with top horizontal line,
    we may think that the 6-shaped curl is enough and that the horizontal
    line comes from the consonant. But consider Kannada Letter Kha (U+c96)
    that does not have a horizontal line but left leaning vertical at its
    right. When Virama is attached to this Kha, there appears a horizontal
    line joining the Kha to the 6-shaped curl. So the horizontal line can
    be thought of as a part of the nominal shape of the virama in Kannada.

    > But these considerations remain valid for other base consonnant letters, even if the ligature is more "destructive" for their base shape...
    TUS4 shows sample half forms for Devanagari (Table 9-1). Similar
    Sample subscript forms for Kannada, Oriya etc would help to widen the
    standard to other Indian scripts. Currently, Devanagari garners most
    of the attention and other scripts get an introduction that is
    "abbreviated but highlight any differences from Devanagari where
    appropriate"! The treatment of ZWJ/ZWNJ for the other scripts has been
    brought out and solved by Peter Constable.

    All the Indic scripts should have tables for 1) Sample half or
    subscript forms, 2) Sample conjuncts 3) conjuncts of consonant and

    Vinod Kumar
    Project IndiX

    PS: Message resent with attached pdf replaced by link.

    This archive was generated by hypermail 2.1.5 : Thu Mar 16 2006 - 22:34:43 CST