Ben Adida
Thu, 15 Jun 2006 06:30:45 -0700
My 2 cents while glancing at this.Ian, thank you very much for pointing out a real technical issue that we should absolutely address. This is one of the first issues we'll take up in July (when I start leading the telecons again.)
Anne, your point is well taken: there might be a way here to hit two birds with one stone: moving away from xmlns declarations to address Ian's point, and towards something that is more text/html friendly at the same time (but still XHTML compatible, of course), which might make RDFa adoption workable for the HTML5 folks, if they so choose.
Ian, I see on your blog that you said that using '.' and '-' were arbitrary decisions. So, I'm wondering, since I now clearly understand how QNames are not supposed to apply to attribute *values*, why didn't you choose the ':' separator? There would be no overlap with QNames from a machine parser standpoint, since QNames don't apply to attribute values, but at least, from a "human parser" perspective, they would immediately ring a bell.
(I also see that you're trying to find a way to make copy&paste work with eRDF so that schema declarations aren't purely in the HEAD, excellent!)
-Ben On Jun 15, 2006, at 6:13 AM, Anne van Kesteren wrote:
On Wed, 14 Jun 2006 20:38:23 +0200, Ian Davis <[EMAIL PROTECTED]> wrote:Hi Ian,Good point. NewsML 2 has also taken your approach #2. What syntax is eRDF using for this?eRDF uses a convention that I first saw described in a W3C workshop report from 1996 [1]. This was later adopted for the Dublin Core[2].Basically a schema prefix is declared in the head of the document like this:<link rel="schema.dc" href="http://purl.org/dc/elements/1.1/"; />That's interesting. It also seems a better solution than RDFa as it integrates with text/html documents rather nicely.-- Anne van Kesteren <http://annevankesteren.nl/> <http://www.opera.com/>