Freebase: I’m in…

Posted on May 19, 2007


I got a sign-in for the alpha release Freebase a few days ago and I’m pretty interested in what I’ve seen. Freebase is, to quote a couple of commentators: “possibly the most exciting thing to happen to the web for some time”. Or, as TechCrunch put it: “This is cool, unless it achieves consciousness and kills us all”…

type “museum”From the museum perspective this is particularly pertinent – anyone who has spent some time in the sector will be aware of the talk about the Semantic Web and what that means to us. I was invited to sit on one of the sessions of the UK Museums and the Semantic Web thinktank a while back and we talked a lot about the need for real examples and the issues around building this stuff from the bottom up rather than the top down.

Well, Freebase is aiming to fill a big hole in the Semantic Web world. As Tim O’Reilly says in his excellent article:

“[Freebase is] a wikipedia like system for building the semantic web. But unlike the W3C approach to the semantic web, which starts with controlled ontologies, Metaweb adopts a folksonomy approach, in which people can add new categories (much like tags), in a messy sprawl of potentially overlapping assertions. “

Or, again, as TechCrunch put it:

“Freebase looks to be what Google Base is not: open and useful”

So what does it actually do? Well, apart from being a “shared database of the world’s knowledge”…

Metabase (the company behind Freebase) has grabbed a bunch of existing databased content, including Wikipedia – as well as this, >22,000 films, >350,000 albums, >300,000 people are also apparently available. For starters, this is a formidable searchable resource (the startup hell of “great concept, no content” really isn’t an issue here)…

personal profileUsers can edit content, wiki-style, but – crucially – Freebase gives people the power to connect bits of content. And this is one of the things that makes it different. The concept of a “type” is central to the site – already a bunch of types (film, computer, astronaut…) – about 700 of them – are already defined. And users are encouragd to add more as they go about editing content. The data is structured – semantic web styley – but, as Tim O’Reilly says “in a messy sprawl of potentially overlapping assertions”. The power will come about as the community begins to hone these assertions into a more meaningful shape.

This is an interesting slant on Wikipedia, which has a pretty unstructured approach: Freebase hopes that this collective approach to creating connections is not only attractive to end users because it is much like the way the brain works, but will also lead to connections that can be tapped into in a very powerful, semantic way.

All the data on Freebase is made available under a Creative Commons license, and (here’s where it gets really exciting), there is an incredibly feature-rich and well documented API. This means that developers can begin to hook into this wealth of information and incorporate relevant content into their own web applications.

I’ll have a play and as soon as I’ve built something interesting, will let you know 🙂