- From: Peter Williams <home_pw@msn.com>
- Date: Mon, 12 Dec 2011 15:40:29 -0800
- To: Kingsley Idehen <kidehen@openlinksw.com>
- CC: "public-xg-webid@w3.org" <public-xg-webid@w3.org>
It's what older triple store readers read. I've found that those humans who can read the XML show not only can triples be read but they have grasped the rdf'ness. Then I can easily test whether they have grasped rdfa' ness. With that done, one is where Ldif and object classes were a decade ago. If one stuffs data formats on properties where ldap had attribute syntaxes, you almost achieve COBOL80. Until webid showcases some benefit (and only Kingsley has ever shown something beyond tech demonstrators and poor grade identity management webapps) it's hard to rationalize why bother. I'm not being negative (just buying into tech hype). I can parse an XHTML stream for a script element of a certain mime type using xpath, and archive the same security semantics as webid validation today. What I cannot do is then benefit from the semantic web. But the webid doesn't actually say how to go to the next step, either. Sent from my iPhone On Dec 12, 2011, at 1:15 PM, "Kingsley Idehen" <kidehen@openlinksw.com> wrote: > On 12/12/11 1:09 PM, Henry Story wrote: >> On 12 Dec 2011, at 18:32, Pierre-Antoine Champin wrote: >> >>> On 12/12/2011 05:45 PM, Henry Story wrote: >>>> On 12 Dec 2011, at 15:24, Pierre-Antoine Champin wrote: >>>> >>>>> Hi all, >>>>> >>>>> having a look at the WebID spec, I spotted a few mistakes in the RDF/XML >>>>> example: >>>>> >>>>> * the rdfs namespace is not declared >>>>> * the closing tag for rdfs:label misses the leading '/' >>>>> * the datatypes xsd:hexBinary and xsd:integer should be expanded URIs, >>>>> not CURIEs >>>> very well spotted, Pierre. Thanks a lot. We have updated the spec here >>>> >>>> https://dvcs.w3.org/hg/WebID/rev/add4f836470d >>>> >>>> WE fixed this in today's meeting >>>> >>>> http://www.w3.org/2011/12/12-webid-minutes.html >>>> >>>> and we should have a new spec with all the latest changes out today. >>>> >>>> Does one really have to use full URLs for datatypes? That's really a bit lame... >>> well, unless I missed something the RDF/XML recommendation, I'm pretty >>> sure you have to... >>> >>> And I agree, it is a bit lame... >> That would be one point in favour of using Turtle as the other MUST format, (and drop rdf/xml) >> Though we would need quite a few other serious reasons to do something like that. > > RDF/XML is a relic that doesn't work at the front door. It has a place, but not the front door. We've made more than 70+ transformers cartridges/drivers and another 70+ cartridges/drivers. It is only when doing that sort of machine oriented work that XML, XSLT, RDF/XML, and GRDDL become useful. > > Again, RDF/XML has the horrible problem of obscuring pathways to the simplicity and potency of EAV/SPO triples. That's why it should never front anything that seeks to engage an attention challenged Human audience :-) > > Kingsley >> >> Henry >> >>> pa >>> >>>> Henry >>>> >>>> >>>>> best >>>>> >>>>> pa >>>>> >>>> Social Web Architect >>>> http://bblfish.net/ >>>> >> Social Web Architect >> http://bblfish.net/ >> >> >> > > > -- > > Regards, > > Kingsley Idehen > Founder& CEO > OpenLink Software > Company Web: http://www.openlinksw.com > Personal Weblog: http://www.openlinksw.com/blog/~kidehen > Twitter/Identi.ca handle: @kidehen > Google+ Profile: https://plus.google.com/112399767740508618350/about > LinkedIn Profile: http://www.linkedin.com/in/kidehen > > > > > >
Received on Monday, 12 December 2011 23:41:17 UTC