W3C home > Mailing lists > Public > www-style@w3.org > July 2010

Re: [css3-fonts] unicode-range and unicode normalization

From: John Hudson <tiro@tiro.com>
Date: Mon, 12 Jul 2010 09:11:02 -0700
Message-ID: <4C3B3E96.7050904@tiro.com>
CC: www-style@w3.org, www-font <www-font@w3.org>
I presume that this is relevant to this discussion in some way?

Some comments on normalisation and fonts, from a font maker's perspective:

If things worked the way I would like, I could make fonts that only 
contained glyphs for decomposed characters, which would then be 
intelligently mapped, using Unicode canonical compositions, to text 
regardless of whether it were NFC or NFD. Why? Because my glyph sets 
could be much smaller and efficient, and I wouldn't need to support 
multiple ways of displaying the same typeform in a single font.

Unfortunately, no existing layout engines seem to support this approach. 
Most now offer mapping from decomposed character strings to precomposed 
characters prior to looking in the font cmap table for glyphs for the 
precomposed characters, but not the other way round: if a font does not 
contain a cmap entry for the precomposed character, a .notdef glyph is 
displayed or font substitution happens, despite the fact that the font 
supports the same character in decomposed form.

John Hudson
Received on Monday, 12 July 2010 16:11:42 UTC

This archive was generated by hypermail 2.4.0 : Friday, 25 March 2022 10:07:48 UTC