- From: Danny Ayers <danny.ayers@gmail.com>
- Date: Wed, 31 Aug 2011 18:30:04 +0200
- To: Dan Brickley <danbri@danbri.org>
- Cc: Manu Sporny <msporny@digitalbazaar.com>, Linked JSON <public-linked-json@w3.org>
On 31 August 2011 18:16, Dan Brickley <danbri@danbri.org> wrote: > On 31 August 2011 18:10, Danny Ayers <danny.ayers@gmail.com> wrote: > >> [21:12] <manu> Not dealing with normalization is a mistake that RDF >> serializations have been making for years... >> >> Right, RDF/XML has so many different possible serializations that it's >> impractical to use standard XML processing tools (XPath, XSLT, XQuery >> etc) on it. Which kind of defeats the object of having it in XML in >> the first place. > >> There have been a few 'normalized RDF/XML' proposal, and a few people >> have used them locally - but they've never really caught on, probably >> because they do demand a normalization step. > > Or because similar graphs normalise differently? Picking the 'top' > node, and deciding when to use refs vs inline sub-elements, isn't > trivial. well, there is the <rdf:Description rdf:about="#subject> <property rdf:resource="#object" /> style as well > Really, a lot of very smart XML experts have looked deeply at RDF over > the years. If it was so easy, ... such an elementary mistake, ... > wouldn't someone have got this right, and proposed a better, > normalised RDF-in-XML? > > The fact that such a mechanism hasn't emerged in 14 years, suggests to > me that the goal might not be coherent or achievable. That is a distinct possibility. In effect we have a whole parallel graph-oriented stack alongside the tree-oriented XML stack. But what are the implications for JSON-LD - maps seem the most common idiom in JSON, do they demand yet another stack?! Cheers, Danny. -- http://dannyayers.com
Received on Wednesday, 31 August 2011 16:30:41 UTC