W3C home > Mailing lists > Public > www-html@w3.org > June 2001

Re: wbr revisited

From: Karl Ove Hufthammer <huftis@bigfoot.com>
Date: Sun, 3 Jun 2001 00:28:23 +0200
Message-ID: <003001c0ebb3$76084260$7e448ed5@huftis>
To: "Einar Westermann" <einar.westermann@trygdeetaten.no>, <www-html@w3.org>
----- Original Message -----
From: "Einar Westermann" <einar.westermann@trygdeetaten.no>
To: <www-html@w3.org>
Sent: Saturday, June 02, 2001 9:23 PM
Subject: wbr revisited

> Is the
> <!ENTITY zwnj CDATA "&#8204;"--=zero width non-joiner-->
> meant to function like the (Netscape-specific) wbr (possible line break
> without hyphenation)?

No. The only practical example I can think of with the English alphabet, is
ligatures. In some fonts (quite a few, actually), the the top of the 'f'
letter "crashes" with the dot in 'i', e.g. in the word 'fish'. In professional
typesetting, the letters 'fi' are therefore replaced with a single 'fi'
"letter", where the top of the f *is* the dot of the 'i'. You can see some
examples at <URL: http://www.will-harris.com/ligatures.htm >.

The browser is free to do glyph substitution, so that the character sequence
'fi' is rendered as a single 'fi' ligature (no browser actually does this). If
you put a 'zero width non-joiner' between the to letter, you can ensure that
the browser doesn't "join" the two characters to a ligature. The ZWNJ is more
useful in non-Western languages.

For information about line breaking and line breaking characters, see <URL:
http://www.unicode.org/unicode/reports/tr14/ >.

Karl Ove Hufthammer
Received on Saturday, 2 June 2001 18:29:37 UTC

This archive was generated by hypermail 2.3.1 : Wednesday, 7 January 2015 15:05:57 UTC