Re: Comments on CSS 2.1 Specification

On Fri, 6 Aug 2004, MURATA Makoto (FAMILY Given) wrote:
> 
> In 4.4 "CSS style sheet representation", the user agent is required to 
> observe the charset of referring stylesheet or document (if any) for 
> determining the charset of CSS stylesheets.  (The fourth item in the 
> list)
> 
> I think that this should be removed for two reasons.  First, this is 
> ad-hoc.

It's also what every major UA implements, and is required in order to 
correctly render most of the east-asian Web.


> Second, some HTML documents are created from XML by XSLT on the client 
> side.  Although such HTML documents may be in UTF-16, they have nothing 
> to do with the encoding of referenced CSS stylesheets.

There are of course many cases where the heuristic in question can be 
shown to fail, but since this is the last-ditch measure, it shouldn't be a 
problem. It can be avoided by correctly labelling the stylesheet.

-- 
Ian Hickson               U+1047E                )\._.,--....,'``.    fL
http://ln.hixie.ch/       U+263A                /,   _.. \   _\  ;`._ ,.
Things that are impossible just take longer.   `._.-(,_..'--(,_..'`-.;.'