[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]
Subject: Re: [office-metadata] Re: ODF and semantic web
Hi Mark, Great to have your feedback on this. Bruce knows the concerns that drove this and gave a good summary. In my heart, I have to admit that I had a sweet spot for taking the route you propose. but I also recognized the cogency of the concerns Bruce and others brings/brought up. We're still at a "first official proposal spec" phase, of course. I wonder, if you and others from RDFa taskforce would make your presence known in ODF-metadata, whether it still might be something to think about to make straight RDFa an option, perhaps in the next proposal go-round. It's modularity makes it pretty easy to bring across. The other point is we have mostly done demo documents using the out-of-content, more heavyweight model. I personally haven't authored any in-content documents (though Bruce will remember back at the very beginning of the process I was a strong proponent for an in-content option). It's always possible that when we get more prototype docs out there we'll find that pure RDFa has advantages we hadn't factored in. I guess what I'm saying is that if you and your colleagues would weigh in on this in group discussions, those of us already in the group would enthusiastically welcome your input. We're a incredibly open and enthusiastic group, and all of us are committed to making this a killer spec. On another topic, we've had a number of discussions--and I have authored a sample doc-- that use XForms to generate as output an RDF instance. This is using the out-of-content,heavyweight model. To me this is very exciting. There are some constraints and issues -- the way I did it was to generate a prototype RDF instance which I made into an XForms model, and then "fill in the bits". But because, in point of fact, a given RDF graph can be serialized in many different ways, this need to commit to a single prototype serialization a priori is a little clunky. It would be nice if XForms could address the underlying RDF graph, rather than one paricular serialization of it. But XForms can't so this, because its path language is XPath, and XPath addresses xml, of course. On the other hand, if the XForms spec would allow a form to use, let's say, SPARQL as its path language, then the form could address the underlying RDF model instead of just a particular serialization of it. Any chance XForms might eventually allow other path languages to address the model other than XPath for this reason? John John F. Madden, M.D., Ph.D. Associate Professor Department of Pathology (3712) Duke University Medical Center Durham, NC 27710 USA +1 919 681 6671 (voice) +1 888 681 6671 (fax) +1 919 597 0304 (mobile) +1 800 307 9538 (pager) On Oct 14, 2007, at 8:33 AM, Bruce D'Arcus wrote:
|
[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]