It’s just data

GRDDL

DanBri pointed me to a specification named GRDDL.  It looks like exactly what I was looking for last August.  Any XML file is one namespace declaration, one attribute definition, and one XSLT file away from RDF.

Oddly, it defines a completely different mechanism for XHTML.  I'm not sure what I think about that.

Definately something to watch, and probably something to include in a best practices guide to Atom.


I'm not sure how settled GRDDL is, and perhaps it could be more consistent across formats. But something to watch, absolutely. Best practices? Certainly, as long as you can get people to agree on the same schema and mapping, otherwise it might end up being pretty incompatible practices. (I brought this up on-list: [link] )

On the topic of RDF actually in XHTML, there's another proposal doing the rounds (a lot of people seem to like it, despite qnames in attributes) : [link]

Posted by Danny at

In the IBM Emerging Technologies Toolkit version we just released, there is a demo technology called "Semantic SDO" in which we demonstrate that it is entirely possible to use the RDF model without requiring instance data to ever be in the RDF XML syntax.  All you need is a data model, a mapping between your XML Schema and the data model, and a processor intelligent enough to understand  'em and you can get the benefits of RDF without the need to convert everything to RDF.

Posted by James Snell at

IBM Emerging Technologies Toolkit: Semantic SDO

Semantically-Adapted Service Data Objects (Semantic SDO) is a prototype extension to the base Service Data Objects implementation provided in the...... [more]

Trackback from Raw

at

James - cool, thanks for the pointer.

It's a neat approach - rather than mapping to the RDF model via syntax (XML => RDF/XML, maybe with XSLT) it specifies the mapping in OWL, with the XML entities identified in an XPath-like pathFromRoot property. The bonus is that the same technique can map between different XML syntaxes that share the same model. Are you thinking what I'm thinking?

Semantic SDO demo/docs: [link]

Posted by Danny at

Danny: absolutely, that was one of the motivations for the design.  A single instance document can be mapped to multiple models depending on what semantics you wish to apply to the data at any given time, without incurring the cost of transforming the instance data each time you want to look at it differently.

Posted by James Snell at

Sam Ruby: GRDDL

A single instance document can be mapped to multiple models depending on what semantics you wish to apply to the data at any given time, without incurring the cost of transforming the instance data each time you want to look at it differently....

Excerpt from del.icio.us/pkeane at

Add your comment