- From: Satoshi Nakamura <snakamura@infoteria.co.jp>
- Date: Thu, 06 Sep 2001 19:43:06 -0400
- To: Graham Klyne <Graham.Klyne@Baltimore.com>
- Cc: Jeremy Carroll <jjc@hplb.hpl.hp.com>, w3c-rdfcore-wg@w3.org
At 06 Sep 2001 17:25:37 +0100 Graham Klyne wrote: > I therefore suggest the above syntax is appropriate, and the current RDF > specification is about right on this > point. > (http://lists.w3.org/Archives/Public/www-archive/2001Jun/att-0021/00-part#54) I think it is nearly impossible to determine which element is the boundaries for 'generic' RDF/XML parsers without rdf:RDF. We can even write RDF/XML without rdf-ns declaration. For example: <x:foo xmlns:x="..." xmlns:y="..."> <y:bar>baz<y:bar> </x:foo> This XML is legal RDF/XML document. But there is no hint to determine whether it is RDF/XML (generate 1 triple) or some other kind of XML document (generate no triple). I think that only the parser can do is to treat whole document as RDF/XML. In addition, parsers based on event-based XML API cannot start processing before they find rdf:RDF element which may not be present. If there is rdf:RDF element, a parser should treat only its contents as RDF/XML, and if not, parser should treat whole document as RDF/XML. So a parser need to create a tree model (such as DOM) or something to hold contents which may be RDF/XML and may not be. I think it is useless to hold 1MB XML contents to get few embeded RDF/XML contents. Thanks. --- Satoshi Nakamura <snakamura@infoteria.co.jp> Infoteria Corporation
Received on Thursday, 6 September 2001 19:45:44 UTC