- From: Larry Masinter <masinter@adobe.com>
- Date: Fri, 5 Apr 2013 15:40:35 -0700
- To: Anne van Kesteren <annevk@annevk.nl>
- CC: "www-tag.w3.org" <www-tag@w3.org>
> > In fact, there are no current W3C or IETF specifications that "work against" > the "Web architecture" in this way. > > http://www.w3.org/TR/html5/syntax.html#determining-the-character- > encoding > treats BOM as more authoritative than HTTP Content-Type if you want to > get technical about this. IF you want to get technical, it's up to the media type (text/html) to define the meaning of the parameters. RFC 4288 lays out only minimal constraints on how the charset parameter is to be interpreted, it only requires that if there's a parameter The "text" media type is intended for sending material that is principally textual in form. A "charset" parameter MAY be used to indicate the charset of the body text for "text" subtypes, notably including the subtype "text/plain", which is a generic subtype for plain text defined in [RFC2046]. If defined, a text "charset" parameter MUST be used to specify a charset name defined in accordance to the procedures laid out in [RFC2978]. I think the main thing is to contain sniffing and media type guessing to only 'modern browsers' and not let it infect non-browsing contexts. The arguments for sniffing have been: - 'misconfigured web sites' - PLUS no way to reconfigure - PLUS existing browsers sniff - PLUS browsers don't want to stop sniffing because if they don't, users will switch to a browser that DOES sniff that is, all four of the conditions have to hold for sniffing to be a good idea to propagate. While that might apply to hyperlinks and transclusion FROM HTML documents, I don't think it applies to network APIs, applications other than browsers, (SIP? Instant messaging? hyperlinks from content other than HTML?). There are better ways to make the previous TAG finding on "authoritative metadata" be consistent with current browser behavior without tearing down the infrastructure for use by other applications. Larry -- http://larry.masinter.net
Received on Friday, 5 April 2013 22:41:08 UTC