- From: <Patrick.Stickler@nokia.com>
- Date: Thu, 11 Sep 2003 12:06:19 +0300
- To: <Patrick.Stickler@nokia.com>, <bwm@hplb.hpl.hp.com>
- Cc: <gk@ninebynine.org>, <danbri@w3.org>, <w3c-rdfcore-wg@w3.org>, <Patrick.Stickler@nokia.com>
> -----Original Message----- > From: Stickler Patrick (NMP/Tampere) > Sent: 11 September, 2003 11:53 > To: 'ext Brian McBride' > Cc: gk@ninebynine.org; danbri@w3.org; w3c-rdfcore-wg@w3.org; Stickler > Patrick (NMP/Tampere) > Subject: RE: I18N Issue alternative: collapsing plain and xml literals > > > > > Here's a twist on the two approaches ... > > > > > > Eh? > > > > Yes. It needs some changes to the MT though, as PatH pointed out. > > Changes, yes, but not new stuff, that I can see. More > of a merging, fusion of existing stuff. My stab at defining the merged plain+XML literal would be: Lexical Space: Same as now defined for rdf:XMLLiteral. I.e. canonical XML, with comments, etc. etc. Value Space: Same as now defined for rdf:XMLLiteral. I.e. XML "things" corresponding to sequences of characters and (optionally) markup which correlate to Infosets and are compared for equality based on their lexical forms. Graph representation: Lexical form + optional lang tag. I.e. "LLL" or "LLL"@xx RDF/XML expression: Two options: As plain string, in attribute value or element content, (without rdf:datatype) converted to escaped, canonical XML when mapped to graph. As XML fragment, in element content with parseType="Literal", (without rdf:datatype) canonical XML when mapped to the graph. For either, a lang tag is derived from any non-null xml:lang scope. Patrick
Received on Thursday, 11 September 2003 05:06:26 UTC