No, these aren’t the necessary precursors for a conspiracy theory about the moon landings, but three of the topics touched upon at the first VoCamp, which took place recently in Oxford. VoCamps are events where motivated individuals can come together and spend some dedicated time creating vocabularies/ontologies for describing data on the Web.

You may have heard of these vocabulary things before. Two popular examples that have been around for some time and are in widespread usage are FOAF (as in Friend of a Friend), for describing people and who they know, and SIOC, for describing the contents of ‘social media’ sites such as blogs and discussion fora. But why do we need more vocabularies, and why do we need VoCamps?

We need more vocabularies because people are increasingly motivated to share their data online, and need some way of describing the data itself in a structured fashion. If people use the same vocabularies when describing data of the same type, or at least some of the same terms, it makes sharing and integrating those data sets much easier. For example, imagine you and I both run online shops selling sports equipment, and we want to describe the stock we hold, if we use the same vocabulary to describe that stock data then anyone wanting to cross search our two shops will benefit by not having to map my data structure to yours — we’ll have saved them the job by converging on the same vocabulary from day one.

At this point in time there just aren’t enough vocabularies around to describe the wealth of data in the world. Left to their own devices people will simply create ad-hoc vocabularies which do little to aid data sharing. It’s for these reasons that we need VoCamps, where people can put day-to-day distractions to one side and concentrate on creating technically sound vocabularies in domains that interest them, according to some of the best practices in the field.

VoCampOxford2008 was the start of this process. I used the time to work with Ian and others on a vocabulary/ontology for describing Whisky. Leigh created his Space Flight vocabulary — not just a flippant bit of fun, but a crucial component in his desire to make NASA data more widely accessible and easily archived. Other groups at VoCampOxford2008 worked on a vocabularies for describing IRC discussions, evidence, discourse, participation, votes, journeys and scientific data. See this page for more information on the vocabs we created.

Now, while some people would no doubt argue that whisky and space flight constitute the two most important topics around, there’s still some way to go in creating the rich ecosystem of vocabularies required for a Web of data. That’s why the second VoCamp will take place in Galway, Ireland in late November. Anyone interested in getting their hands (metaphorically) dirty and creating some vocabs should register now before the event fills up — it’s free. Given the location I’ll have to spend a little time in Galway refining the Whisky ontology, but no doubt there’ll be plenty of scope for creating vocabularies in other areas. I may even attempt a vocabulary for describing conspiracy theories, but I imagine that no-one would be able to agree on the details!

The big news is that the GRDDL specifications just moved to Proposed Recommendation status at the W3C, which means it “…is a mature technical report that, after wide review for technical soundness and implementability, W3C has sent to the W3C Advisory Committee for final endorsement.“. To coincide with this I’ve prepared a laminator-ready GRDDL Quick Reference (PDF) card.

The purpose of GRDDL is to allow data contained in HTML and XML documents to be transparently interpreted as RDF, in other words making those documents first-class information on the Semantic Web. For most document authors there’s virtually no work necessary to put their material on this Web of Data. For example, if they’re using microformats then in effect the job is already done. (If the author wishes to create a proper chain of authority, they should use XHTML 1.0 and include a profile attribute in their document’s <head> element).

GRDDL is already supported by a fair number of programming libraries and tool implementations.

To get an immediate view of the kind of things that can be embedded in HTML, you might like to try the Tabulator Semantic Web data browser in Firefox – the 0.8 stable version of Tabulator is probably best right now. First you’ll need to make a quick change to Firefox’s settings (see the right-hand column on the Tabulator page). Then copy & paste Dan Connolly’s home page URI (http://www.w3.org/People/Connolly/) into the “URI” box at the top and click “Add to Outline”. You can now browse the information embedded in Dan’s page as well as any linked data (if you View Source on Dan’s page you’ll see the it contains a mix of microformat data as well as more general Embedded RDF).

Like most of the other W3C Semantic Web specifications, GRDDL is described in a suite of documents:

• GRDDL – main specification doc
• GRDDL Test Cases – demonstrate the expected behavior of a GRDDL-aware agent
• GRDDL Primer – introductory material
• GRDDL Use Cases – example applications of the technology

The core technique is for the GRDDL-aware agent (which may be part of a browser script or a server-side system) to transform the syntax of the source document into an RDF format, typically using XSLT. To anyone familiar with XML technologies this might sound simple enough, but there are several rather elegant mechanisms through which the required transformation(s) can be identified.

For HTML, the transformation(s) can be identified in an profile document, for XML the transformation(s) can be identified in an XML Namespace document. In each of these cases no extra work is required of the individual document author to enable interpretation of the document as RDF. If reasonable care is taken, the maintainer of the profile/namespace document can add the necessary material post hoc, bringing all documents using their profile/namespace document onto the Semantic Web in one fell swoop. The GRDDL-aware agent will see the profile attribute or namespace URI and “follow its nose” to the profile or namespace document, pick up and apply the transformation to produce RDF. This is a creative use of HTTP and the formats but is entirely in line with the letter and spirit of existing Web specifications. It also fits nicely with Tim Berners-Lee‘s Linked Data recommendations for maximising the findability and (re-)usability of data on the Web.

This blog is hosted by Talis, so it would be remiss of me not to mention the connections. Ian Davis (Talis CTO) and myself are both members of the W3C GRDDL Working Group, with Ian acting as an editor of the Primer. Quite a while before GRDDL’s specification process began, Ian came up with Embedded RDF (as found in Dan’s home page linked above) which enables arbitrary RDF data to be included in a standard HTML document. It’s a neat solution to this long-standing problem in the Semantic Web community, and (because it’s standards-based) it also happens to be compatible with microformats. With GRDDL, this data can be automatically extracted.

It’s still early days, but support for GRDDL is planned in the Talis Platform, following the commitment Ian expressed in this testimonial:

“Talis believes that GRDDL represents one of the most important steps along the road to the Semantic Web. It provides a very simple yet extraordinarily powerful mechanism to uplift documents into the web of data. Talis intends to fully support GRDDL in our Semantic Web Platform, allowing our customers to automatically extract searchable RDF metadata from their existing content with very little effort.”