W3C home > Mailing lists > Public > public-xg-webid@w3.org > December 2011

RE: neither FCNS nor FOAFSSL can read a new foaf card (hosted in Azure). RDFa validators at W3C and RDFachecker say its fine...

From: Peter Williams <home_pw@msn.com>
Date: Sat, 31 Dec 2011 09:17:04 -0800
Message-ID: <SNT143-W64203CDC3CB6E6AB4E631992930@phx.gbl>
To: <mo.mcroberts@bbc.co.uk>, <kidehen@openlinksw.com>
CC: "public-xg-webid@w3.org" <public-xg-webid@w3.org>

Great, except it had NOTHING to do with the issue I was facing. It was peripherally related, though. it sounds like a bug (that had consequences, due to the 90% dominance of IE6).

 

lets et back to qualifying the spec, as a piece of engineering. Is it mature, is it ready, is it consistent, is it creditable, have the political agendas been removed, does it teach without leading?

 

Jurgen told me he expected 2 link to work, when a user enters them on some W3C hosted RDFa.py form, acting as a linked data client. I should enter a URI foo and URI foo#html. He expected triples to be returned from windows web server to the linked data client embedded in a server module, in either case. The only differnece between the 2 cases was presence of #html on the GET on the wire (IE6 like), as emmitted by the W3C server doing server-server communications over HTTP.

 

The case the triples MUST be returned in server-server communications with a #fragment on the wire of the GET is what I have learned is termed: "linked data nuances" for web server configuration. 

 

Henry's web server exhibits such linked data nuances. Tested with telnet over port 80, his server accepts #fragments on the wire on server-server communication, and it delivers triples. It does not reject the GET, with an 400 response, like Windows does. It may be possible to make Windows/IIS do what Henrys server does (though I have no evidence in practice its possible, only rumour)

 

I was unable to make windows native web server do what Henrys web server (probably apache) does. I thus could not exhibit "linked data nuances". I could not do the second case Jurgen mentioned. I thus failed to complete the "endpoint test".

 

Given the failure of the endpoint tests, I attributed the failure THEN of all 3 validation agent sites (FCNS, FOAFSSL, ODS) to perform (a) de-referencing (per linked data numances) and (b) positive QUERY=ASK of the RDFa stream (given in the spec) due to the failure of the window to perform all the linked data nuances.

 

The spec's RDFa example of webid profile is distinguished from the Turtle example of a webid profile; in regard to the use of relative name spaces. I the spec are using the distinct to EXPOSE and ENFORCE the MUST of the linked data nuance requirement, not tested for when using the specific Turtle example. The RDFa had relative naming, and thereby forced the issue of a linked data client PROPERLY and FULLY handles (per full power formal web architecture) full power de-referencing, which includes the obligation to properly and fully address the inner naming/addressing duality of URIs in 100% of the intended.

 

i though this was the better part of the spec, since it forced web engineering to test 100 compliance (and not the usual world of 80% fiddle and fudge web). This will show the class of the work, even if folks back off in reality in product, to 90% compliance, even for security interworking. its good to know that full rigor is possible.

 

 

 

 

 

 

 

  		 	   		  
Received on Saturday, 31 December 2011 17:17:32 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Saturday, 31 December 2011 17:17:33 GMT