Re: Unicode 3.1: IDS and ZW(N)J

From: John Jenkins (
Date: Thu Jan 25 2001 - 12:50:11 EST

On Thursday, January 25, 2001, at 08:55 AM, John Cowan wrote:

> Certainly. But what if you want to display the content of Figure 10.8
> (p. 270) in such a font? In that case, you can use ZWNJ to break up the
> ligature. The first example could be written as <U+2FF1,
> U+4E95,U+200C,U+86D9>, which cannot come out as a ligature.

No, in that case you turn off ligatures. If I'm doing something like
Figure 10.8, I'm *not* doing it in plain text, and if I have intelligent
fonts that can handle ligatures, then I'd better have intelligent
applications than can turn them off.

Naturally, you can also insert the ZWNJ anywhere in the sequence so that
the rendering engine won't recognize it as a valid ligature sequence in
the font and won't show it as a ligature. Granted, it won't be a valid
IDS anymore, either. The question in my mind is when do you envision
that you would want, in plain text, to have a sequence which will be
parsed as a valid IDS but must not be displayed as a single glyph? To
my mind, that really is analogous to trying to force the display of e +
acute as something other than e-acute but have it still parse as e-acute.

My bottom line remains where it has always been: Ligation control does
not belong in plain text.

This archive was generated by hypermail 2.1.2 : Tue Jul 10 2001 - 17:21:18 EDT