From: Peter Kirk (peterkirk@qaya.org)
Date: Sat Jun 11 2005 - 07:53:53 CDT
On 11/06/2005 02:17, John Hudson wrote:
> Kent Karlsson wrote:
>
>> That does not really follow. I think "inline" tags *between*
>> Arabic/Syriac/
>> Mongolian letters (possibly with combining marks) can be seen as acting
>> like
>> ZERO-WIDTH JOINER for the purpose of Arabic/Syriac/Mongolian shaping. ...
>
This cannot be correct. Consider the sequence of Arabic DAL, followed by
font markup, followed by Arabic HEH, as part of an Arabic script word.
As DAL never joins to the left, if there is no markup here this should
be rendered as isolated or final form DAL followed by isolated or
initial form HEH. And this same joining behaviour should be preserved if
parts of the word are to be rendered in different fonts, colours etc.
But this sequence with ZWJ should be rendered as isolated or final form
DAL followed by final or medial form HEH, which is certainly not what is
required. The requirement should be that, for shaping purposes, the
markup should be treated as completely transparent for the purpose of
shaping, in the same way that combining marks and "most format control
characters" are treated as transparent. In other words, they should be
treated as in class T (not class C like ZWJ) in Table 8-3 in the Unicode
standard p.199 (http://www.unicode.org/versions/Unicode4.0.0/ch08.pdf).
I can see John's point that this might cause implementation difficulties
where there is a change of font, but nevertheless this must be the
correct behaviour as it preserves the generally correct appearance of
the characters. It should be up to users will notice and correct for
mismatches e.g. when glyphs in different fonts and sizes do not join
correctly; it is not for Unicode to decide that because there may be a
mismatch completely different glyphs should be substituted.
>> ... Certain changes that the markup may result in, such as a size
>> change,
>> will
>> make the join more or less "misfit" graphically. But whoever wrote the
>> markup
>> asked for a size change, not a joining change. Ligature formation should
>> (always) be blocked over markup tags.
>
I agree. There seems to be a need to define markup as breaking
ligatures, much as ZWJ and <ZWJ, ZWNJ, ZWJ> do according to Figure 15-2
on p.391 (http://www.unicode.org/versions/Unicode4.0.0/ch15.pdf). But
this should be done in a way which is transparent to normal joining,
which is unlike the behaviour in any of the columns of this table: the
display for the last row should be as on the left column in the table,
but for the preceding row as in the right column. The alternative must
be to form the entire ligature as if in either the preceding or the
following font.
-- Peter Kirk peter@qaya.org (personal) peterkirk@qaya.org (work) http://www.qaya.org/ -- No virus found in this outgoing message. Checked by AVG Anti-Virus. Version: 7.0.323 / Virus Database: 267.6.8 - Release Date: 11/06/2005
This archive was generated by hypermail 2.1.5 : Sat Jun 11 2005 - 10:39:54 CDT