Re: Ligatures fi and ffi

From: Hans Aberg (
Date: Thu Jun 02 2005 - 04:17:54 CDT

  • Next message: Theo Veenker: "UCS-2/4 & BOM"

    At 01:28 +0000 2005/06/02, Rajeev J Sebastian wrote:
    >On Wednesday 01 June 2005 5:24 pm, Hans Aberg wrote:
    >> At 17:44 +0100 2005/06/01, Jon Hanna wrote:
    >> >>If one so wants, one can add all the glyphs one wants, adding a
    >> >>property field saying that it is a rendering character.
    >> >
    >> >No, we cannot. That would completely change the scope and purpose of
    >> > Unicode.
    >> It would change the scope of the current character set in that respect.
    >> So what do you feel is the purpose of Unicode? -- In the context the
    >> quote above is taken from, I am speaking about glyphs that are used
    >> to give proper rendering to semantic characters.
    >I may not fully understand this thread, but what you are saying, has already
    >been implemented: TrueType Open and OpenType Font handle ligatures (for any
    >language) by tagging a ligature glyph with the semantic (unicode) codepoints.
    >So the ffi ligature is tagged with "f+f+i". Similiar things are done for
    >placement of diacritics, and also optionally using positioning instead of
    >precomposing the ligatures into unencoded positions in the font file.

    It is mainly a discussion of where to put the rendering information,
    it seems me. One can put it into the Unicode character set, as is
    already done in some cases, and one put in a higher level, like a
    computer language. The discussion is complicated by the actual usage
    of such ligatures. In some cases it is semantically acceptable to
    break the ligature up; in other cases, it is not. Example: the "ae"
    ligature. The latter case may depend (sub-)script or even (in
    principle) the author. This complicates the decision choices one will
    have to make.

       Hans Aberg

    This archive was generated by hypermail 2.1.5 : Thu Jun 02 2005 - 04:19:04 CDT