Cyrillic guillemotleft, multilingual texts.

From: Peter R. Mueller-Roemer (
Date: Sat May 14 2005 - 09:52:55 CDT

  • Next message: JFC (Jefsey) Morfin: "what is Latn?"

    John H. Jenkins wrote:

    > On May 13, 2005, at 10:33 AM, Hans Aberg wrote:
    >> Now, to the guillemots. Assume that <<...>> are the Cyrillic
    >> guillemots and that [[...]] are the French guillemots. If I write
    >> [[Russian text]], is the semantic meaning from that different from
    >> <<Russian text>>, assuming the quoted text is the same? The first
    >> case is probably part of a French text, quoting some Russian, and
    >> the second case, some Russian text quoting Russian.
    > And the standard Unicode response would be that in the former case,
    > you use a font specifically designed for (or optimized for) French,
    > and in the latter one designed for Russian.

    I hope 'standard Unicode response' is not this oldfashioned typewriter
    approach. We just had a conference in Berlin to support multilingual
    communication with unicode.
    Not that I am happy with the present implementations. The universal
    keyboard might be a little farther away, but I plead for some small
    improvements that would allow for at least most European + some
    near-Eastern languages and technical Symbols to have one easy to
    remember keyboard-layout.

    Combining diacritical marks (CDM) point the way. Graphite-technology
    schows that sequences of CDMs can be handled much more sensible than by
    using a default rule of overstriking in the OT-font Arial Unicode MS.
    But I want narrow Greek diacritcals (e.g. strong or weak h + accute) to
    combine side by side and be centered over the narrow iota and the wide
    o-mega. Similarly with some Hebrew accents, Arabic .. Etheopian (make
    cononant-vowel-versions enterable as simple sequence of 2 components).

    > This sort of unification issue is common in Unicode -- Japanese vs.
    > Chinese glyph design for Han, French vs. Polish glyph design for
    > accented Latin letters -- and has been from the first. Unicode
    > considers these glyph issues to be addressed through some other
    > mechanism. I should point out specifically that OpenType allows for
    > the same font to have different glyphs for the same character
    > depending on the language, if the application supports this, as do
    > other high-level font formats such as AAT and Graphite.
    > ========
    > John H. Jenkins

    This archive was generated by hypermail 2.1.5 : Sat May 14 2005 - 09:54:09 CDT