Re: s/URI/IRI/ in

On 2/16/2011 8:08 PM, Eric Prud'hommeaux wrote:
> The IRI lineage is a but funny. RDF mandates the %-encoding of any non-ASCII. It preceeded 2397 so it professed a compatibility instead with XML Namespace IRIs (which includes the small non-ASCII range [\xA0-\xEFFFD], as well as ' '). Somehow, the RDF world seemed to move as one from "<%E6%A4%8D%E7%89%A9#%E5%90%8D>" to"<植物#名>". SPARQL declares its IRIs to "correspond" to RDF URI References and provides a grammar which excludes ' ', but copies XML's Name production. At any rate, the world seems to have agreed to IRIs, so let's breath a sigh of relief and perpetuate this boon to I18N.

During the development of the 2004 recs the I18N WG advised RDF Core 
that we should align with the at the time unfinished IRI spec and there 
was (IIRC) no dissent, and there has been no effective dissent to that 
plan anywhere in the community. So I would date the move to "<植物#名>" 
as approx 2003, and we are still finishing off the plan of depending on 
IRI because the I18N guys were late to the party (but they had at least 
said they were coming, and did indeed turn up)


Received on Thursday, 17 February 2011 18:14:07 UTC