A Practical Form of Semantic Web?
8 September 2007
This week I was impressed by a Technometria podcast with Rob Cook co-founder of Metaweb Technologies. We know of the pure semantic web ideas being promoted by Tim Berners-Lee and many others that require a priori ontologies in OWL to be produced before semantic web contents can be produced usually in RDF format. The continual refinement of the ontologies as new, unanticipated content is acquired is very messy and often impractical even though the goal of a machine-readable web is of immense benefit.
In a natural reaction to this pure semantic approach we have seen the growth of tagging folksonomies where all users are free to choose their preferred tags. this approach has worked reasonably well for small to medium scale data collections but rapidly becomes less useful as the number of users grows.
Thus it was refreshing to encounter the approach of the Metaweb team who have placed themselves between the pure semantic web and unconstrained folksonomies. To quote them:
Metaweb designs and builds enabling technology for application developers and publishers who are creating a new kind of Web – a Web that benefits from the power of people and machines. Metaweb’s technology powers commercial and open-source applications, with a focus on organizing and managing complex data structures. An example of the kind of applications made possible by Metaweb can be found at www.freebase.com.
Their freebase product mentioned above is well worth looking at if you either want to incorporate structured data search in your web application or as an individual want to perform a structured search over a large dataset such as the contents of Wikipedia. All content in freebase is covered only by the Creative Commons Attribution licence. For developers the RESTful API and the simple structured query language are very straightforward. I look forward to trying it out. Metaweb describe freebase as:
… home to a global knowledge base: a structured, searchable, writeable and editable database built by a community of contributors, and open to everyone. It could be described as a data commons.
At the moment freebase is in private alpha but accounts can be obtained on request from the freebase site within 24 hours.