- From: Ian Hickson <ian@hixie.ch>
- Date: Tue, 30 Oct 2007 23:42:32 +0000 (UTC)
On Tue, 16 Nov 2004, Ben Adida wrote: > > The REL attribute has been proposed as a means to extract *simple* > meaning from the HTML. For example: > > ==== > By viewing this web site, you are agreeing to its <a rel="terms-of-use" > href="terms-of-use.html">terms of use</a>. > ==== > > UAs could theoretically implement special UI for some of these typed > links, though certainly it wouldn't be required. Most importantly, an > automated crawler could glean structured properties from these typed > links in HTML. > > And, of course, in order to handle the scoping problem of these names, > one should allow fully qualified names in the REL. What scoping problem? > Thus, the HTML above could make use of well-defined properties like > those of Dublin Core, and look like this: > > ==== > By viewing this web site, you are agreeing to its <a > rel="http://http://purl.org/dc/elements/1.1/rights" > href="terms-of-use.html">terms of use</a>. > ==== > > The only thing that's necessary to enable this right now is to allow the > REL attribute to be a fully qualified URL. HTML 5 need do nothing more, > and no requirements need to be made of the UAs. It's an easy way to add > structure to HTML documents without having to define all the possible > structural terms. Actually URLs are already allowed in HTML5 -- anything that doesn't contain spaces can be used. http://www.whatwg.org/specs/web-apps/current-work/#rel Cheers, -- Ian Hickson U+1047E )\._.,--....,'``. fL http://ln.hixie.ch/ U+263A /, _.. \ _\ ;`._ ,. Things that are impossible just take longer. `._.-(,_..'--(,_..'`-.;.'
Received on Tuesday, 30 October 2007 16:42:32 UTC