Re: unicode Digest V4 #3

From: Philippe Verdy (verdy_p@wanadoo.fr)
Date: Mon Jan 05 2004 - 08:53:43 EST

  • Next message: Michael Everson: "LATIN SOFT SIGN"

    From: "Peter Kirk" <peterkirk@qaya.org>
    > I would think that the issue here is whether "i with retroflex hook
    > below" is a suitable description of this character. It may be a
    > reasonable match for the glyph. But this is not a mark of retroflection
    > (although arguably of back articulation (cf. U+0320)), and to call it
    > one is probably an anachronism. There is probably no historical link
    > with the retroflex hook.
    >
    > I would prefer a new character with no compatibility decomposition; or
    > if there is any compatibility decomposition it should be directly to
    > dotless i which is the modern equivalent.
    >
    > It seems that we do actually need two new character pairs, this one and
    > also the soft sign lookalike - unless it is considered acceptable to use
    > the Cyrillic characters in Latin text cf. the use of Latin Q and W in
    > Cyrillic Kurdish.

    Note that the name I gave just suggests its approximate look. It does not
    necessarily mean its is semantically correct. So of course, if the only use
    of
    this i with lower-right hook has a better traditional name, it should have a
    name that matches this tradition if it is ever encoded. But for now, in
    absence of this character in Unicode, the composition:
        <Latin small letter dotless-i><combining retroflex hook below>
    or
        <Latin capital letter I><combining retroflex hook below>
    is quite good to represent it, and it works with Turkish/Azeri case
    mappings.

    That's why it could be a compatibility decomposition (not a canonical one,
    because it would break the stability policy, if the canonical decomposition
    was not also excluded from recomposition, making the new character simply
    unneeded, as excluded compositions are a pollution to avoid in all new
    Unicode assignments)

    So if new characters are assigned, they will have better names than the name
    suggested by this compatibility decomposition, which is for now the only way
    to represent the character (I think it is beter to use this combining
    sequence
    than to approximate it with a soft sign, or a B/b letter, or a L/l letter)

    Note that the Latin soft sign you want for the second Latin version of Azeri
    (that looks like "b" not like the cyrillic "bI") is very near from the
    "Latin letter
    sixth tone" which already exists in Unicode (U+0184/U+0185), except that
    the bottom of the left stem is rounded before the loop (more or less like
    the
    "6" digit) instead of straigth in the normal soft sign (which is extremely
    near
    from the small letter "b"). You'll note how the imported cyrillic soft signs
    are
    hard to read distinctfully from the latin letter "b" or the two lattin
    letters "bl"
    if writing the soft-sign in lowercase text. So the form used in the "sixth
    tone"
    letter is really helpful to make it read distinctly.

    So, could this Latin soft sign be considered instead as a glyph variant of
    the
    Latin sixth tone letter ?



    This archive was generated by hypermail 2.1.5 : Mon Jan 05 2004 - 09:27:55 EST