On Sat, 4 Sep 2004, Paul Novitski wrote:
> I had one of those treasured Internet Explorer Moments lately working on a
> friend's website, when I plugged in the 8000-series character entities and
> still got straight quotes & apostrophes in IE6. By hacking desperately at
> my CSS for a while I finally figured out that if I used:
> font-family: Verdana, Arial, sans-serif;
> then I got straight quotes, while:
> font-family: Verdana, Arial, Xsans-serif;
> (or simply deleting sans-serif) got me the curlies I was looking for.
That sounds very strange, though not stranger than IE behavior I've seen.
But can you reproduce the phenomenon at will? I cannot.
If CSS is really involved that way, then the issue is on-topic in this
list. Otherwise it's clearly off-topic, since the methods of including
characters in an HTML document are not CSS.
On my computer, Verdana has curly glyphs for left and right quotation
marks (double and single). And naturally when Verdana is listed first in
the font-family value, those glyphs are used. In Arial it's different: at
normal (by IE behavior in "standards" mode) size and below, the glyphs are
vertical and hardly distinguishable from Ascii quotation mark and
apostrophe; but at higher sizes, they are curly. I cannot see why a third
value in the font-family list could have any impact when _both_ the first
two fonts are available and contain glyphs for the characters; but this
wouldn't be the first bug found in IE, would it?
I'm not a bit uneasy after realizing that setting font to Arial nullifies
the author's attempt at getting the quotation marks right, i.e. curly, by
using character references or by other means, instead of the common habit
of using just " and '. Setting font size larger than normal surely isn't a
feasible solution for copy text.
I have often noticed that IE displays _some_ (Ascii) apostrophes as
vertical (which is the correct way by Unicode definitions), some
apostrophes (maybe even in the same document) as curly, as right single
quotation marks (which is incorrect by Unicode defs but often what the
author wants). I have been unable to find any pattern in this, and I'd be
interested in hearing if someone has. Maybe the solution lies in the same
direction as the cause of your observations.
Jukka "Yucca" Korpela, http://www.cs.tut.fi/~jkorpela/
List wiki/FAQ -- http://css-discuss.incutio.com/
Supported by evolt.org -- http://www.evolt.org/help_support_evolt/