RDFGoo - Web Searchable RDF

The RDFGoo project is an experiment in making RDF more accessible to tradional web search engines. It was inspired in part by Patrick Stickler's RDF Facet work. I'd originally intended to use a modified version of RDF Facet syntax to encode all combinations of predicate, subject, and object for each statement (P, S, PS, O, PO, SO, PSO), but was driven to seek a different solution by the word length limitation of my testing ground -- Google Desktop Search. I ended up using a hash (SHA1) of the combinations instead (I later disovered a similar solution has been suggested by the Swoogle folks). The approach requires embedding both the rdf content and the encoded form (since the encoded form is not reversible as it is with RDF Facets).

The idea is that folks will embed chunks of RDFGoo in their html pages and search engines will index it. None of the tradional techniques I've come across for embedding RDF in XML is suitable since they all place the rdf if non-indexed parts of the page (e.g. comments or scripts). My current thinking is to embed the RDF within a non-displayable div with a well-known id (so it is easily distinguished). The RDF is further wrapped in either a XMP tag, or marked as CDATA (if the page is XHTML) so that the RDF doesn't need to be encoded in any way. I chose to use the turtle syntax of RDF since it's commonly accepted as the most human readable (and best passes the view source test).

Here's an example of RDFGoo:

<div style="display:none" id="rdfgoo">
<xmp>
@prefix rdf: <http://www.w3.org/1999/02/22-rdf-syntax-ns#>.
@prefix foaf: <http://xmlns.com/foaf/0.1/>.

[ foaf:name "Geoff" ].

#Ob0ff576b72b51c15f62f12185ddff3f5d4d78429 P7e2d8376c0825ad025f24554ea71399bcff5d9a1 PO44050d489b0c159ad081ea5c4717b3b260b1fb47 
</xmp>
</div>

Tool and Other links

• A page to generate RDFGoo from your RDF
• A page to extract and display the RDFGoo in your page
• A simple query builder
• RDF Plugin for Google Desktop Search
• Details of the hash process