- From: Graham Klyne <GK@NineByNine.org>
- Date: Wed, 12 Feb 2003 19:09:46 +0000
- To: Jeremy Carroll <jjc@hplb.hpl.hp.com>
- Cc: www-archive <www-archive@w3.org>
Jeremy, Some questions were being asked about <rdf-wrapper> in IRC. I don't know if I got the intent right. But there was a suggestion or two that might avoid the need to use <rdf-wrapper>. Transcript follows. [[ <gk> Eric, as I recall, <rdf-wrapper> stuff is to make a well-formed document so that canonical form can be described by simple reference to XML C14N specification. It feels a bit kludgy, but I couldn't see a neater way to do it. * danbri wishes http://www.idleworm.com/nws/2002/11/iraq2.shtml were in SVG <danbri> gk, is rdf-wrapper in the Syntax spec anywhere? *** grove_ has joined #rdfig * DanCon returns from call, wonders where we are on the calendar agenda; adjourned? <ericP> gk, i'd recommend saying that each child of the propertElt with parseType="Literal" is a exc-c14n document <danbri> adjourned, i believe. <DanCon> thx <gk> Danbri, not as I recall. It's part of the mapping from abstract graph to the domain of interpretation. <danbri> it has angle brackets in it, so seems syntaxy... <ericP> if i find such a literal as a result of a query, will it have <rdf-wrapper> around it? <danbri> depends on the query api <gk> Eric, I'd be happy to suggest that to Jeremy, but is that enough? What about the space between the children? I don't know XML subtleties well enough to know. <ericP> so that i need to XPath my real data out? <danbri> presumably we won't see <rdf-wrapper> in ordinary rdf serializations? <ericP> gk, inter-children space, hmm, thinking... <gk> danbri, absolutely; not in RDF/XML serialization. <ericP> what if rdf-wrapper were made more abstract, ie, some element, so that folks would never try to put it in or take out out with XPath. <ericP> ? <amy> bye all *** amy has quit IRC ("Leaving") <gk> As I understand it, it's just a device to help the formal specification, and (maybe) the description of equivalend RDF graphs under XML canonicalization. The presence of XML literals in the RDF somewhat middies that issue, I think. <ericP> assuming my current interp is correct, that it is a temporary construct to make sure the canonicalization starts one down from the root element *** DanCon changes topic to 'RDF & Semantic Web hack'n'chat http://rdfig.xmlhack.com/' <ericP> algorithm: <gk> Eric, I think there's a problem of exposition if folks are tempted to put in <rdf-wrapper> in their XPath. <ericP> get byt string from API or parser <ericP> stick inside <rdf-wrapper> <ericP> allpy exc-c14n <ericP> remove from <rdf-wrapper> <ericP> stick in graph <ericP> done *** arnarl has quit IRC ("Client Exiting") <gk> Eric, I'm very sympathetic. Don't know if that's enough. I believe the intent of what you describe matches the WG intent. <ericP> i'll think about this and see if i can come up with anything <zoyd> other than RDF, what else will/does make the semantic web happen ? <ericP> right now, my code (serving annotea) does something like that, except it doesn't need to stick it inside <rdf-wrapper> as that fucntionality is available in my canonicalizer <ericP> other than inter-child spaces, is there an affect of the <rdf-wrapper> <ericP> ? <ericP> (other than to confuse my poor canonicalizer's understanding of which namespaces are currently in play) *** Ol has left #rdfig <gk> Eric, you just reminded me: <rdf-wrapper> is also used to bring the appropriate language tag into play. I suppose that might be done on each child element. But now I feel the complexification of this approach may start to bite. ]] #g ------------------- Graham Klyne <GK@NineByNine.org>
Received on Thursday, 13 February 2003 06:04:23 UTC