Bidi reordering of soft hyphen

Richard Wordingham richard.wordingham at ntlworld.com
Wed Apr 2 15:46:51 CDT 2014


On Wed, 02 Apr 2014 03:17:35 -0700
Asmus Freytag <asmusf at ix.netcom.com> wrote:

> On 4/2/2014 12:36 AM, Richard Wordingham wrote:

> > But it is a *resolution* rule that converts the true hyphen or minus
> > sign to Bidi Class L; these apply before the scope reduces from
> > paragraph to line.

> When breaking a line at a soft hyphen, one is essentially modifying
> the text around the line break for display, because the SHY is not
> specific as to what should happen (as was the case with German old
> orthography, the changes go beyond simple substitution of a hyphen).
> 
> When you change the text, you have to fix up the resolution.

The argument was based on what happened to U+002D HYPHEN-MINUS.  The
change to the text then is to replace what is, in code order, 'CARROT
IS car<U+002D>rot...' by 'CARROT IS car<U+002D, U+2028 LINE
SEPARATOR>rot...'.  One can even argue that this replacement would
result from SHY under the rules of English typography.  Reapplying the
resolution rules, the left-to-right run now includes, even after
truncation, 'car<U+002D>'.

Richard. 



More information about the Unicode mailing list