Re: [unicode] Re: Ligatures For Indic languages

Date: Fri May 30 2008 - 09:46:45 CDT

  • Next message: John H. Jenkins: "Re: Understanding PUA - (was) Re: Re: Ligatures For Indic languages"

    Dear Vinod,

    On Fri, 30 May 2008 14:37:41 +0530
    "Vinod Kumar" <> wrote:

    >Most Indian script enthusiasts are quite satisfied with the current state of
    >generating ligatures. Unicode encoding and shaping rules have found their
    >way to common usage as well as high quality printing. There is no reason,
    >incentive nor inclination for creating any standard for a ligature
    >collection. Nor does it gurantee the quality of Indic script printing.

    Excuse me, I want to devide the request from Manik in two parts.

    R1. a standardization of ligature glyph collection for Indic scripts.

    R2. a character encoding scheme to point a precomposed ligature glyph
        by single character codepoint.

    I think, you tell R2 is not required. I guess most people using
    Unicode-(or ISCII-)oriented software agree with you. Although
    some rendering systems, like OpenType or TrueType/GX, use glyph
    index to point a precomposed ligature glyph, it is just internal
    mechanism (there's no need to define font-independent glyph index),
    and the rendering quality of them are sufficient, so no need to
    give further restrictions to character encoding by the insertion
    of precomposed ligatures in PUA.

    I don't have strong objection. I think R2 comes from the legacy
    rendering system that cannot separate the character encoding
    codepoint and glyph index of the font. The discussion based on
    the prioritization of such legacy systems is difficult, I'm afraid.

    # if the prioritization is permitted, 8bit encoding system should
    # be prioritized, a rendering system without proportional font
    # support should be prioritized, etc etc.

    I'm not sure if R1 is unreasonable request /or not. I think,
    even a Devanagari OpenType font including only "unshaped"
    glyph for each codepoint of ISO/IEC 10646 (without ligature
    glyphs or glyphic components to be layout) can insist "this
    is Unicode OpenType font". I think it's not false, because
    Unicode does not guarantee the glyph shaping feature, and
    what kind of glyphs should be displayed. But I don't want to
    use such poor font. So, I think it looks natural to hit upon
    the idea of "standard collection of ligatures". I'm not the
    expert of the orthodoxy of Indic scripts, and I don't know
    such idea is possible /or not.

    However, considering the approach of INSFOC collecting the
    decomposed glyphic components, the standardization of
    precomposed ligature would be considered as too expensive
    approarch, at least for Devanagari, Gujarati and Punjabi:
    INSFOC charts for these 3 scripts are published. Vinod, how
    do you think of other Indic scripts? Almost same?

    >The CJK approach of pages after pages of encoded letters or ligatures or
    >conjuncts is not needed for Indic scripts. We are proud of the logical
    >structure of our scripts that allows thousands of shapes to be generated
    >from a couple of hundred basic codes. The enhancements needed to achieve
    >this have been researched well, standardized into Unicode and implemented
    >for daily and specialist use.

    I agree with your pointing out: the character encoding of Indic
    scripts for information interchange is not needed to include
    precomposed ligatures. The solution how to display Unicode/
    ISCII coded text had been researched since 80s, and the straight
    solutions were already implemented and used before official
    adoption in ISO/IEC 10646.


    >Vinod Kumar
    >On 5/30/08, Mahesh T. Pai <> wrote:
    >> said on Fri, May 30, 2008 at 04:02:53PM +0900,:
    >> > I'm interested in if Bureau of Indian Standards has any plan
    >> > to define a standard of precomposed ligature collection (to
    >> > guarantee the quality of Indic script printing), and asked
    >> > such question to BIS, but I couldn't receive any comments.
    >> This is something I have raised in more than one forum, but
    >> unfortunately, I have no affiliations, and I am not a developer, so
    >> have been unable to follow through. :(
    >> The TDIL may be the more appropriate forum - but it is part of a
    >> government behemoth. But they have defined a minimal set of glyphs for
    >> Malayalam - and I guess the position is same for Devanagari
    >> too. Devanagari is described in one of the files downloadable from
    >> .
    >> --
    >> Mahesh T. Pai <<>>


    This archive was generated by hypermail 2.1.5 : Fri May 30 2008 - 09:51:12 CDT