- From: François Yergeau <francois@yergeau.com>
- Date: Fri, 10 Oct 2003 23:35:21 -0400
- To: www-i18n-comments@w3.org
3.2 says "Where practical, declare the page's character encoding by setting the charset parameter in the HTTP Content-Type header." I think this should be just the opposite: servers should shut up and let (X)HTML documents speak for themselves. Several years' experience has shown that servers get it wrong much of the time when they bother to set charset. Since this setting has precedence, this really breaks the whole thing. Please recommend that servers *not* set charset, unless they have a good reason to do it and actually do it right. See also the recent thread on the TAG list on this (starting at [1]) as well as section 4.9.6 of the Web Architecture WD [2]: "In the case of XML, since it is self-describing, it is good practice to omit the charset parameter". [1] http://lists.w3.org/Archives/Public/www-tag/2003Sep/0042.html [2] http://www.w3.org/TR/2003/WD-webarch-20031001/#xml-media-types =================================================================== In 4.1, there's an ed. note that says "Describe the evils of using <font> to cheat on the charset and represent other scripts." You may want to take a look at my old (1996) rant on this at http://babel.alis.com/web_ml/html/fontface.html. -- François
Received on Friday, 10 October 2003 23:41:05 UTC