Re: Representative glyphs for combining kannada signs

From: Vinod Kumar (
Date: Thu Mar 30 2006 - 01:11:30 CST

  • Next message: Kent Karlsson: "RE: Malayalam vowel sign AU (was Re: Representative glyphs for combining kannada sign)"

    On 3/29/06, Kent Karlsson <> wrote:
    > Antoine Leca wrote:

    > >
    > > Finally, if I am also allowed that, how is it different for the position of
    > > the I matra in the rendering of Nagari conjunct NG.K.I ङ्कि?
    > I'm not sure what your point is here. But if you mean whether the display
    > order should be I:NG:K or NG:I:K, I think that should depend *only*
    > on the underlying characters. In this case NG.K.I should have the display
    > order as I:NG:K and NG.ZWNJ.K.I should have the display order as NG:I:K,
    > regardless of which font is used for display.

    The underlying character sequence <Nga Halant Ka MatraI> should have the
    display with the glyph gMatraI preceding the glyph corresponding to Nga
    only if the <Nga Halant> has generated a half form or has joined with gKa.
    So <gMatraI gNgaHalf gKa> and <gMatraI gNgKa> are correct.
    But if the font does not have the half Nga or conjunct glyphs as is normal,
    then the correct rendering is <gNga gHalant gMatraI gKa>.
    I am sick and tired of people glibly claiming that
    <gMatraI GNga gHalant gKa> is valid display for the character sequence.

    > However, an *editing program*
    > may insert ZWNJ at suitable places to avoid moving the I too far (either for
    > a specific font, or depend on some compromise data for several
    > common fonts). The editing program could even be something that
    > interjects itself just prior to display of a text; but should not be part
    > of the basic (Indic capable) display engine itself.
    The editing program should not be burdened with the task of looking at the
    fonts, determining that for <cons1 halant cons2>, neither gCons1Half nor
    gCons1Cons2 are present and insert a ZWNJ into the character sequence
    <cons1 halant ZWNJ cons2>. The 'something' that Kent refers to has its
    incarnations in Uniscribe that stands between the font and the upper layer
    (editor for example), carries out a back and forth conversation with the
    font and arrives at the correct rendering. But it is part of the basic
    Indic capable engine and not part of an editor.

    Just as the editor should not be burdened with such font or script dependent
    activities, intermediate layers too become simpler if they do not have to
    solve the problems of the lower layer. Moving the gMatraI right after a
    <consonant halant> sequence that refuses to combine with the halant
    should better be done at the point where the refusal occurs- in the font.

    Vinod Kumar

    This archive was generated by hypermail 2.1.5 : Fri Mar 31 2006 - 01:06:25 CST