Re: Grapheme clusters

From: Deborah Goldsmith (goldsmit@apple.com)
Date: Tue Oct 05 2004 - 13:54:28 CST

  • Next message: Philippe Verdy: "Re: Grapheme clusters"

    UAX 29 provides for language-specific tailoring of break behavior, and
    this seems like a situation where you'd want grapheme break to be
    tailored. See section 3 of UAX 29 for a discussion of this.

    Which language are we discussing here?

    Deborah Goldsmith
    Internationalization, Unicode liaison
    Apple Computer, Inc.
    goldsmit@apple.com

    On Oct 5, 2004, at 9:09 AM, Chris Harvey wrote:

    >
    > Ysgrifennodd Christopher Fynn <cfynn@gmx.net> ar y 05-10-2004 am 10:42:
    >
    >> It would of course be possible to have these pair combinations
    >> replaced by a single ligature glyphs using the "locl" feature in
    >> OpenType under a specific language tag.
    >
    > Are ligatures what Iím looking for? The letters of the consonant
    > cluster like "kw" are not joined together visually in any way. Also,
    > when I use OpenType for ligatures like ffi st etc. the parts of the
    > ligature are deleted one at a time. I think the UAX#29 or UAX#28
    > discusses the differences between grapheme clusters and ligatures.
    >
    >> Is there a locale setting for this language? - Many applications now
    >> automatically tag documents with the current input locale.
    >
    > There is no locale setting. The Native American language has a very
    > small speaker base.
    >
    >> You can tell these people that the PUA is no real solution since you
    >> can get some very unexpected glyphs displayed for PUA characters.
    >> (Microsoft Windows automatically maps a bunch of non-BMP CJK
    >> characters to PUA codepoints and sometimes these will display instead
    >> of the glyphs in your font.
    >
    > In fact, I have been continuously sending them lists of all the
    > reasons not to use PUA, the CJK problem being only one of many. The
    > problem is, they feel the PUA solves their two biggest issues,
    > backspacing the clusters and collation, without realising that a whole
    > new and bigger batch of problems arises.
    >
    > Thanks,
    >
    > Chris Harvey
    >
    > --
    > Gwlad heb iaith, gwlad heb galon
    >
    > www.languagegeek.com
    >
    >



    This archive was generated by hypermail 2.1.5 : Tue Oct 05 2004 - 13:57:53 CST