This search engine uses wordnet to find synonyms of search terms, and hypernyms of marked-up keywords. So a search for skyscraper finds one picture and a search for structure finds the skyscraper and a load of other pictures too.

A browseable hierachy of pictures organised by the hierarchy of the wordnet terms depicted in them. Start browsing at Artifact or look at the RSS backend.

As with the ruby rdf parser, this runs XSLT over your RDF/XML to make ntriples then parses those. It spits out mysql insert statements ready for querying with my rdf sql query stuff.

I felt it was time to learn a bit more about SVG and do some concrete work with it. I made a fairly simple SVG document that renders picdiary feeds. It parses the RSS and builds the page on the client-side using Jim Ley’s lovely javascript rdf parser.

A simple algorithm for translating rdfdb-style queries into SQL to query a triple store with a very simple schema. A perl and a PHP implementation is provided.

A mail sent to www-rdf-interest after attempting to embed XMP metadata in JPEGs.

(It turns out my code doesn’t do the jpeg embedding quite right. It looks like the JPEG-XMP package will do it correctly.)

Continuing the theme of harvesting information from MT entries into RDF summaries, I wrote another MT plugin (src) that uses HTML::LinkExtor to extract <a> links.

The RDF for this entry shows that we have xlink:href triples for each link in the…

A mail to the bots list about a web-to-IRC gateway.

Subject: [Bots] gatebot — making bots URL-addressable
From: Matt Biddulph
Date: Mon Aug 26 14:41:05 2002