- From: Eric Prud'hommeaux <eric@w3.org>
- Date: Wed, 30 Mar 2011 06:20:15 -0400
- To: Dan Brickley <danbri@danbri.org>
- Cc: RDF WG <public-rdf-wg@w3.org>
* Dan Brickley <danbri@danbri.org> [2011-03-30 11:59+0200] > As I understand / dimly remember, the RDFCore round of specs tried > their best to anticipate the IRI specs, but could only make normative > reference to the URI spec. > > http://www.w3.org/TR/rdf-concepts/ > > "Note: this section anticipates an RFC on Internationalized Resource > Identifiers. Implementations may issue warnings concerning the use of > RDF URI References that do not conform with [IRI draft] or its > successors." > > ...whereas http://www.w3.org/TR/2004/REC-rdf-syntax-grammar-20040210/ > does not mention IRIs. > > Meanwhile http://www.w3.org/TeamSubmission/turtle/ "Turtle uses IRIs > as term identifiers." > > For JSON my assumption has been that we would use IRI. Can this be confirmed? > > At the POI WG F2F we are looking at an example link to the page for > Amsterdam in the Korean Wikipedia. I hope these come through the list > OK. > > 1. the pretty link appears in Korean script (to me at least). > > { > "url": "http://ko.wikipedia.org/wiki/암스테르담" > } > > 2. if this is escaped so as to be a pre-IRI URI, we get instead an > ugly string, twice as many chars: > > { > "url": "http://ko.wikipedia.org/wiki/%EC%95%94%EC%8A%A4%ED%85%8C%EB%A5%B4%EB%8B%B4" > } > > I'm agnostic for now, on question of where-or-whether this stuff gets > canonicalised. But I would like to express a preference that verbose > URI escape sequences are not imposed on eg. Korean URLs like the one > given here. SPARQL's position is that RDF nodes look like <http://ko.wikipedia.org/wiki/암스테르담>, and if you want to GET them, you follow the %ification rules in RFC2397. That was a fairly bold and necessary step, IMO with a downside if folks have RDF data with nodes already in the %-y form. If so, your IRI assertions about <wiki/암스테르담> and their URI assertions about <wiki/%EC%95%94%EC…> will describe different resources. A still rarer problem occurs with when we try to GET their URIs and those URIs are for resources which actually include %XX in the (file) name. This is a pretty contrived screw case, and couldn't happen with modern web servers which %-decode the incoming URIs before resolving them against the filesystem or rewrite rules, etc. > cheers, > > Dan > -- -ericP
Received on Wednesday, 30 March 2011 10:20:50 UTC