RDFWeb  
overview | feedback

RDFWeb: a linked Information System

RDFWeb is an experimental linked information system. This site provides links to documentation and early prototypes. The RDFWeb general overview provides a basic explanation of the system's goals and capabilities.

Prototypes

We have two implementations in progress: one using Perl and DBM, the other using Java and SQL. Both provide tools to aggregate, store, query and present views of the RDFWeb data graph.

Perl implementation (danbri)

Note that I am rewriting this, and the current demo is malfunctioning (although it looks nice and some of the built-in toys (path-finder, rdfviz, rss) still work.

Java (libby)

Demo status note: the perl demo is part broken currently; gets confused on dan b's data due to bad handling of person identifiers, and mention of dan b in another rdf file.

For Developers

Disclaimers: the Perl demo service is rebuilt 20 mins past every hour by traversing from all known RDFWeb files. The system may be fragile, and currently does not have a large dataset (ie. don't make closed-world assumptions - if it doesn't know about some connections they may nevertheless exist...). Currently, the version you see here is the version I'm hacking on.

Background reading

Excerpt from the AmSci article:

Some of the most interesting large graphs are those in which we are the vertices. These social graphs are associated with the phrase "six degrees of separation," popularized by a 1990 play of that title and a later film, both written by John Guare. The idea is that the acquaintanceship graph connecting the entire human population has a diameter of six or less. Guare attributes this notion to Guglielmo Marconi, who supposedly said that wireless telegraphy would so contract the world that any two people could be linked by a chain of 5.83 intermediaries.

Acknowledgements and Context

Eventually this stuff will be expressed via the rdfweb system instead of in prose. meanwhile...

The rdfweb testbed is based on lots of work done elsewhere.

RDF Parsers: We use the W3C Perllib RDF/XML tools and the W3C SiRPAC Java RDF parser.

RDF Storage/Query: The Perl RDF store is currently built using a DBM-based triplestore developed by Jasper Tredgold and Phil Cross (both of ILRT) in the DESIRE project. Libby's Java version uses PostgreSQL as an RDF store instead.

Visualisation: the RDF graph visualisation (select 'RDF' next to the URL for people who have an RDF file describing them) uses the RDFViz service, which in turn uses the GraphViz tools from AT&T/

Graphics: Jan Chipchase (formerly of ILRT, now at GomaGoma) was kind enough to draft some designs for a possible RDF logo last year; these are seeing the first light of day on the RDFWeb site.

A bit more context: Dan and Jan originally collaborated on the first version of the Social Science Research Grapevine, which provided a (centralised rather than distributed) user profiles and connections system for likeminded social science researchers. Jan's P-Lab online photo archive applies similar ideas to image collections.

Logic: Jan Grant contribution some code for exploring the RDFWeb data dumps in logic programming environments.

Feedback

Comments/question to danbri or rdfweb-dev@egroups.com

Valid XHTML 1.0!

dc:contributor Powered By RDF