From: Mark Davis (mark.davis@jtcsv.com)
Date: Thu Jun 26 2003 - 21:34:32 EDT
1. I agree with Ken about the current lack of precedent for Cfs before
combining marks. Interestingly, that we do have a proposal to do just
that, in
http://www.unicode.org/review/pr-9.pdf
However, note that the whole purpose of putting the Cf after the Ra is
to separate it from the halant, so that the halant will ligate with
the following character rather than the preceding. So in that sense,
PR#9 is entirely consistent with breaking a combining character
sequence into two parts.
2. Because Cfs do break combining sequences, I would be very leery of
using any of them to solve the Biblical Hebrew issue. One possibility
is to use a combining mark instead. That is, something with (α) no
visible glyph, (β) combining class = 0, and (γ) general category = Mn.
Unlike the Cfs, this would *not* break a combining sequence. There
would be two possibilities.
a. define a new character with these characteristics.
b. use a variation selection character.
Now, we decided that VS characters would not apply to any but base
characters, but one of the primary reasons for that was so that they
wouldn't disturb canonical order. So easing this restriction in this
case might be reasonable, since that is exactly the point!
Of course, such a change would need to be sanctioned by the UTC, and
it might take a while before fonts supported it, but it may be a way
out, one that doesn't require waiting for the assignment of a new
character. So this is in the spirit of Ken's original proposal.
Mark
__________________________________
http://www.macchiato.com
► “Eppur si muove” ◄
----- Original Message -----
From: "Kenneth Whistler" <kenw@sybase.com>
To: <rick@unicode.org>
Cc: <unicode@unicode.org>; <kenw@sybase.com>
Sent: Thursday, June 26, 2003 17:48
Subject: Re: Biblical Hebrew
> Rick wrote:
>
> > > I now like better the suggestions of RLM or WJ for this.
> >
> > I'll have to disagree with Ken. I'm not so sure about either of
these. I
> > don't think anyone has, in the past, considered what conforming or
> > non-conforming behavior would be for a RLM or WJ between two
combining
> > marks. This needs a bunch more study to determine what on earth it
would
> > break in existin implementations.
>
> Point taken.
>
> >
> > On the other hand, ZWJ between two combining marks has at least
been
> > discussed, and in the case of Indic anyway, it has known,
documented
> > effects.
>
> This, however, has the same problem. The specification of the use
> of ZWJ and ZWNJ in Indic scripts is not *between* two combining
> marks, but following a combining mark (halant), preceding another
> base character, usually a consonant. So we don't really know what
> the implications of trying to put it between two combining marks
> would be -- there aren't any specifications for doing so (yet).
>
> --Ken
>
>
>
This archive was generated by hypermail 2.1.5 : Thu Jun 26 2003 - 22:21:24 EDT