- From: Anne van Kesteren <fora@annevankesteren.nl>
- Date: Mon, 30 Aug 2004 16:30:17 +0200
It might be a good thing to describe META elements in a bit more detail and match them with current implementations. HTML 4.01 currently states that the 'http-equiv' attribute is supposed to be used by servers[1], while current practice is that browsers do something with this in *text/html documents*. If WHATWG can define this somehow it might be easier for new UAs to know what to implement instead of trying to figure out how other UAs are doing it. I believe that current UAs first look at the HTTP header. If no 'charset' parameter has been given the document is spidered for the META element. If it has been found the character encoding is taken and the document re rendered with the found character encoding. If no character encoding is found the UA uses some mechanism to find it or assumes iso-8859-1, the default for 'text/html'. (I can be wrong here, this is just an indication of something the specification should tell.) An emphased part should notice that this does not apply to XHTML documents. (Maybe an additional note that search engines don't really look at META elements anymore might be a good thing too. Especially if that note has an ID attribute to point to it :-).) [1]<http://www.w3.org/TR/1999/REC-html401-19991224/struct/global.html#h-7.4.4.2> -- Anne van Kesteren <http://annevankesteren.nl/>
Received on Monday, 30 August 2004 07:30:17 UTC