RE: Uppercase ▀ is coming? (U+1E9E)

From: Philippe Verdy (
Date: Fri May 18 2007 - 12:46:44 CDT

  • Next message: Gerrit Sangel: "CJK variation modifier"

    Frank Ellermann
    > EnvoyÚá: vendredi 18 mai 2007 05:50
    > └á:
    > Objetá: Re: Uppercase ▀ is coming? (U+1E9E)
    > Asmus Freytag wrote:
    > > I agree that almost anything looks better than gamma-shaped
    > > proposals, but as I wrote before, what's considered acceptable
    > > will be influenced over time by what's deployed.
    > While I still don't see the need of a code point for this glyph
    > variant, GROSZSTADT is an instructive example, like MASSEMASZE.
    > MASZSTAB might also exhibit any issues with SS-based designs.

    My opinion is that the two shapes (capital long S + capital Ezh ligature, or
    capital long S + capital S ligature) are possible, and one could preferable
    to the other, depending on the ethymological meaning (SZ or SS).

    The issue in fact is similar in case of lowercase writing, where the two
    possible glyph ligatures are also possible.

    But given that the esszet letter does not specify by itself which meaning is
    intended in some cases ("Masse"/"MASSE" vs. "Masze"/"MASZE"), the capital
    version will always be ambiguous (and dependant of which national
    orthographic reform is used), except in common words for which the intended
    meaning is clear (this is not clear for proper names, but for "muss"/"MUSS"
    the interpretation of the letter as a dopelt-ess and not as an ess-tzett is

    If one needs to make an explicit preference for the interpretation (for
    example in proper names), I think that the only way to solve the problem of
    which ligature would be appropriate would be to encode <S, ZWJ, Z> or <S,
    ZWJ, S> (using capital or small versions, of long S or S or Z or Ezh) when
    appropriate) and let the font display the correct ligature if supported, or
    display the two separate letters if the ligature is not supported (keeping
    the intended meaning).

    The only issue is in word-processors and word-breakers, that should not
    insert hyphens within the three-characters sequence, but this would not
    break the rules for case conversion or word-wrapping.

    Given the two possible interpretations of the esszet, a font designer could
    also make fonts that offers two glyphs for the letter, and could desire that
    Unicode accepts the registration of one or two variant selectors to specify
    the intended meaning and associated letter form (at least for common words,
    which have an already accepted standard orthography if the sharp s letter is
    present in the encoded text, but not for proper names which do not depend of
    the national orthographic variant): given the Swiss orthography which
    promoted the dopelt-ess meaning by default, the existing lowercase ess-tsett
    adopts the rounded <long s + s> ligature, so this remains the default (and a
    variant selector could be used only for the ess-tsett meaning using a <long
    s + z> or <long s + ezh> ligature).

    In fact, it could ease the conversion between the Swiss and German
    orthographies for rendering documents, if the existing esstsett was either
    deprecated (in favour of the use of a ZWJ) or qualified by using an optional
    variant selector (automatically selectable in word-processors using a
    proofing tool, or using a GUI feature selection helper (like right clicking
    on words where the two options are possible).

    This archive was generated by hypermail 2.1.5 : Fri May 18 2007 - 12:50:03 CDT