Steven Pemberton wrote: > > On Wed, 14 Feb 2007 15:51:13 +0100, Elias Torres <elias@torrez.us> wrote: > >> Steven Pemberton wrote: >>> >>> On Tue, 13 Feb 2007 14:57:10 +0100, Ben Adida <ben@adida.net> wrote: >>> >>>> Yes, we've seen this, too, which is why all of the examples no longer >>>> use LINK and META outside of the HEAD. >>>> >>>> Basically, we're working on having all of RDFa implementable using only >>>> extra attributes, since such extra attributes are supposed to be >>>> ignored >>>> by browser, as per the HTML spec, and in fact other tools (the Dojo >>>> toolkit) use custom attributes already and the browsers don't mind. >>> >>> Ha! I'd missed that. Good idea. >>> >>> How about: >>> >>> <p> >>> <span role="meta" property="title">XHTML<sup>tm</sup> Basic</span> >>> ... >>> </p> >> >> Why do we need role="meta"? Is it for reification purposes, etc? >> >>> >>> and >>> >>> <p> >>> <span role="link" rel="index" href="p-index.html"/> >>> ... >>> </p> >> >> Why not just use the fact that @href is there? >> >> It's snowing outside, so forgive me if I'm missing something. :) > > The standard rule for @rel and @property is that you look for the > closest @about above in the tree, or otherwise use the document as the > about. > > For meta and link is was different; the about was the parent element > (unless there was an explicit about on the link or meta). Right. It is used to create blank nodes, but could we do something a bit more obvious instead, like @about="[_:whatever]"? Instead of going around the way and defining an encoding to something we defined in a draft specification. Anyways, we are discussing your suggestion, but Ben might already have another since he mentioned it. -Elias > > Steven > >Received on Wednesday, 14 February 2007 15:22:47 GMT
This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 8 January 2008 14:15:03 GMT