- From: Ian Hickson <ian@hixie.ch>
- Date: Tue, 12 Dec 2006 00:56:17 +0000 (UTC)
On Tue, 12 Dec 2006, Karl Dubost wrote: > > > > > > <link rel="hcard" href="http://www.w3.org/2006/vcard/ns"/> > > > <link rel="hcard" href="http://www.w3.org/2006/03/hcard"/> > > > > I don't really understand how that would solve the problem; could you > > elaborate? > > ok. For microformats it is mandatory to have a profile URI in the head Mandatory but rarely done, so only mandatory de-jure. De-facto it's optional and doesn't do anything. > it helps specifically parsers. Parsers rarely actually pay any attention to profile="", both because most content omits it, and also because it's harder to do so. So while profile="" was originally intended to be used by parsers to reduce ambiguity, in practice it isn't used by them and doesn't help them. > It has also the benefits that an authoring tool can download > automatically XMDP profile for creating an help to edit microformats. Auto-generated UI is rarely optimal, though. > So a page containing microformats looks like that. > > <!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Strict//EN" > "http://www.w3.org/TR/xhtml1/DTD/xhtml1-strict.dtd"> > <html xmlns="http://www.w3.org/1999/xhtml"> > <head profile="http://gmpg.org/xfn/11"> > <title>Tantek's Thoughts</title> > <meta http-equiv="content-type" content="text/html; charset=utf-8" /> In HTML5, the above can now be written as: <!DOCTYPE HTML> <html> <head> <title>Tantek's Thoughts</title> ...which is far easier to write and understand. > Then in the page there are things like > > <ul class="xoxo facets"> > <li><a href="http://technorati.com/profile/tantek" > rel="me">Technorati</a></li> > </ul> > > rel="me" has a meaning because of the profile up there. With the new proposal, the above still works, but doesn't require the profile attribute. > With the new proposal > > * People can add this information even if they do not have access to the > template (head section). Most common use case form editing. > * People have it right under their eyes near the information they want to > describe (if they wish it). > * Parsers can still have the information to disambiguate when necessary. With the spec as written now, however, people still don't need access to the <head>. The disambiguation thing is nice in theory (which is why I wrote a detailed normative description for how to handle it about a year or two ago, in far more detail than HTML4 ever did), but in practice nobody uses it and it therefore it doesn't actually disambiguate anything. > > Unfortunately in both cases we don't really have any choice; for back > > compat, <link> and <meta> elements that aren't in the <head> must be > > moved to the <head> by the parser. > > Then for back compatibility you will have to keep the profile attribute. I don't really see why. Nobody uses it. What useful content would you be being compatible with? > See > 4. Using GRDDL with valid XHTML > http://www.w3.org/TR/grddl/#grddl-xhtml > > Parsers are not only browsers parsers. Removing profile="" makes GRDDL implementations easier and makes them more compatible with existing content. How is that not a boon? > Do you have an explanation for the why of > "<link> and <meta> elements that > aren't in the <head> must be moved > to the <head> by the parser." It's what browsers do... what do you mean? -- Ian Hickson U+1047E )\._.,--....,'``. fL http://ln.hixie.ch/ U+263A /, _.. \ _\ ;`._ ,. Things that are impossible just take longer. `._.-(,_..'--(,_..'`-.;.'
Received on Monday, 11 December 2006 16:56:17 UTC