On Wed, 28 Jan 2004, brucie wrote:
in post: <news:bv************@ID-196529.news.uni-berlin.de>
"Damian" <no********@adress.getting.too.much.spam.org> said:
I was reading some old groups.google.com posts about css class names not
being allowed underscores
corrected in the errata but not part of the spec
Substantively changed in the errata, or purported to be (errata aren't
supposed to do that, but they didn't seem to care); as you say, the
change isn't officially part of the spec, but see draft CSS2.1.
http://www.w3.org/Style/css2-updates...12-errata.html
I think the hon Usenaut might also have been interested in why it was
so. Originally the idea was that names in HTML would be SGML name
tokens, and they're defined to start with a letter and then use
letters, digits and hyphen-minus. Nothing else.
Since CSS class and id names are referencing name attributes in an
SGML-based markup such as HTML, they were originally happy with a set
of rules for CSS that matched the rules for HTML...
However, both HTML and CSS subsequently got much more lenient about
what characters could be used in names. But most of the extra
characters have to be escaped if you want to use them in CSS.
For a latin-based language it's much easier to stick with the simplest
subset, if the choice is yours. I think W3C were worried that authors
writing Japanese or whatever would complain that they weren't allowed
to write their name attributes in their own language, and would then
accuse the W3C of being too English-centric.