Re: uniscribe engine & opentype font

From: Richard Wordingham (
Date: Mon Feb 27 2006 - 18:04:01 CST

  • Next message: Travis Griggs: "Re: CLDR/LDML questions"

    Ngwe Tun wrote:

    > It's might be off-topic; If you think off-topic, please drop off to
    > another
    mailing list.

    Well, if the discussion goes into much detail on SE Asian scripts, it might
    more properly belong on the seasia forum - subscription
    details at .

    > I would like to know about how does it work relationship of Uniscribe &
    > Frankly, Microsoft uniscribe engine was not supported Burmese shaping. But
    some persons implemented opentype with default script features rlig and
    kern. It's properly rendered in Windows XP SP2. But We need to patch
    uniscribe engine.

    It's 'just' a matter of replacing USP10.dll with a more capable (and
    well-tested!) version. It's not easy to do yourself globally (unless you
    have a backdoor through a dual boot system).

    > It's easy but I doubted in their relationship. let me have
    any related article about uniscribe and opentype.

    The introductory text is at . I am not
    sure how accurate it is - I still can't type Sanskrit sim.ha (U+0E2A,
    U+0E34, U+0E4D, U+0E2B), and I've never seen the dashed circle in Thai.
    (Even if I should type sara ue (U+0E36) for U+0E34, U+0E4D, that still
    doesn't help me with U+0E38, U+0E4D, e.g. for pum.liGga (U+0E1B, U+0E38,
    U+0E4D, U+0E25, U+0E34, U+0E07, U+0E3A, U+0E04) 'masculine gender'. The
    Thai Royal Institute Dictionary has sara u below the consonant and nikkhahit
    above, but the on-line version has the undefined character entity &plaum;)
    It doesn't work even if I set my language to Sanskrit (with the Kedmanee
    keyboard) in Windows XP SP2.

    > Thai, Burmese, Laos and
    Khmer are difficult to detect clustering/syllable boundry, it is work done
    in uniscribe, right?

    The clustering isn't difficult to detect - clusters start at each
    non-subscript consonant code not immediately preceded by a virama/coeng. Of
    these four scripts, only Lao has subscript consonant codes, only Burmese has
    a virama, and only Khmer has a coeng.


    This archive was generated by hypermail 2.1.5 : Mon Feb 27 2006 - 18:08:45 CST