- From: Ivan Herman <ivan@w3.org>
- Date: Mon, 03 Dec 2007 13:19:20 +0100
- To: public-rdf-in-xhtml-tf@w3.org
- Message-ID: <4753F448.8060001@w3.org>
By the way, this issue was on my mind for a while (not a really very very important one, but nevertheless) The current conformance clause gives an exact list of which triples are to be generated and it does not allow to generate extra triples (in the same graph, that is). And, in general, I agree. The issue I have, however, is what to do if I want to generate, say, a warning to the output. Eg, if I want to generate a warning if the @profile attribute is not present (if we end up saying that @profile is a SHOULD, then processing should probably not stop in the absence thereof, but some sort of a warning would be good). As a simple approach I was wondering whether annotations (ie, generation of rdf:label and/or rdf:comment) should not be exempt from the conformance rule I described. Ie, whether it would not be acceptable for an RDFa processor to add rdf comments into the result where such warning text could be added... Just a thought... Ivan Ivan Herman wrote: > Same here. Well, I use an XML parser, ie, the input must be valid XML. > Other than that, I just apply the processing rules. > > In the future, when we have made a decision on that in the group, I > intend to examine and do 'something' with the @profile value. What the > 'something' is will depend on what the spec will exactly say. > > Ivan > > Benjamin Nowack wrote: >> At the moment, ARC ignores the doctype and simply applies the >> RDFa processing rules to the passed document's node tree. I can >> make the RDFa doctype mandatory if you think that my current >> approach encourages the creation of non-conforming RDFa. >> >> Benji >> >> >> On 03.12.2007 10:44:48, Simone Onofri wrote: >>> I hope using xmlns and DTD can cover all questions about it. >> And so, can be >>> fine if Validator also supports validation for xmlns is >> the best. But I hope >>> to give the question also to who have implemented >> RDFa extractors to have also >>> the point of view also another point of >> view. I've added Fabien, Ivan, Bengee, >>> Dave (any others)? >> The @profile is the solution we've used for GRDDL and it >>> works fine. >> Cheers, >> >> Simone >> >>> Le 30 nov. 2007 à 21:54, Sean B. Palmer a >>> écrit : >>>> On Nov 26, 2007 2:43 AM, Karl Dubost <karl@w3.org> wrote: >>>> >>> would the following be a solution for you? >>>>> <html >>> xmlns="http://www.w3.org/1999/xhtml" >>> xmlns:dc="http://purl.org/dc/elements/1.1/" >>>>> version="xhtml11 rdfa >>> svg"> >>>> In what specification would the interpretation of the @version >>> values >>>> be given? Would they be extensible by users other than the W3C? >>> I'm >>>> not sure they'd need to be extensible, admittedly. >>> There are >>> troubles with different type of mechanisms. >>> >>> * known values >>> - Dominant >>> players may impose its values >>> - Strong Communities will impose a set of >>> values on small communities >>> - Sometimes the known values are not known to >>> you, how do you find >>> the doc >>> * URI system >>> - burdensome for authors >>> without an authoring tool >>> - Weakness because of Cache Implementations >>> (Single Point of Failure) >>> >>>> It's been suggested to me that you meant for >>> @version to be a hook for >>>> namespace GRDDL to dispatch off of; is that >>> something that you thought >>>> about? >>> An identifier more than a namespace. >>> A flag which says: "Hey watch >>> out, here there might be RDFa" >>> >>>> This >>> *would* solve the RDFa discovery problem for me, but I'm not sure >>>> how well >>> it would work as a discovery mechanism in general, especially >>>> given the >>> extensibility question and so on. From what Mark and Shane >>>> have said, it >>> sounds like they're only considering @profile at the >>>> moment. >>> It >>> doesn't solve the extensibility question indeed. >>> >>>> See also >>> http://www.w3.org/2001/tag/issues.html#standardizedFieldValues-51 >>> >>> >>> At a >>> personal level, I'm for URIs, though I would prefer a mechanism >>> ala CSS, >>> where I can declare all my namespaces in *one specific file* >>> on my site, and >>> be able to link this file from all my documents. >>> >>> GRDDL suggests the use of >>> profile. >>> http://www.w3.org/TR/grddl/#grddl-xhtml >>> >>> <head >>> profile="http://www.w3.org/2003/g/data-view"> >>> >>> but there are two issues for >>> me, >>> >>> * the file which is delivered at >>> >>> http://www.w3.org/2003/g/data-view >>> is a document I have to read, there's >>> no predefined format that I >>> could automatically grabbed. >>> * You have to be >>> able to edit head, which is impossible in many >>> scenarios. Being able to >>> point to another file locally would be cool. >>> ala CSS link rel="stylesheet" >>> | style element | style attribute. >>> gives a great >>> flexibility. >>> >>> >>> >>> >>> >>> >>> >>> >>> >>> -- >>> Karl Dubost - W3C >>> >>> http://www.w3.org/QA/ >>> Be Strict To Be Cool >>> >>> >>> >>> >>> >>> >>> >> >> > -- Ivan Herman, W3C Semantic Web Activity Lead Home: http://www.w3.org/People/Ivan/ PGP Key: http://www.ivan-herman.net/pgpkey.html FOAF: http://www.ivan-herman.net/foaf.rdf
Received on Monday, 3 December 2007 12:19:26 UTC