Version 4.0 of Freebase (or “the client” as it is called in Metaweb-lingo) just got deployed. Big congrats to the many involved!

I think this is a very important milestone: not only the new site sports a completely redesigned (and fresher, if you ask me) style and a special logo, most importantly, it outlines Metaweb’s renovated focus to go after communities of data enthusiasts, rather than trying to entice anybody in being a data contributor.

There are several good reasons for this change of direction:

1. reducing data agoraphobia – it might start by feeling mighty powerful to be able to access and edit anything in a database that contains 5 million different topics, but it turns into scary pretty fast. How do you know you’re not doing something bad? That you’re not damaging somebody else’s application that depends on this topic having this type set even if it’s totally empty and feels out of place? The best description I heard of this feeling is the ’stadium effect’: editing Freebase sometimes feels like walking in a stadium full of people to do your thing. Unfortunately, not exactly ideal to entice volunteer contributions.
2. one size does not fit all – one of the results of solutions that want to be general enough to apply to a vast variety of problems is that they tend to look so flexible they tend to feel boring and built out of compromises. Freebase is only one way of interacting with the underlying data that Metaweb hosts (which is why it’s called ‘the client’), but it’s clearly the most visible and for that reason the most important. But one size cannot fit all comfortably, and bases are the first step (of more to come and currently in progress) toward giving the communities of data enthusiasts the tools to personalize the way the data they care about is presented, viewed, edited, controlled and ultimately used.
3. increasing relational density – both for data and for the social network of the contributors that care about it, the value is not necessarily in the quantity of nodes in the graph, but how well connected these nodes are to one another. I strongly believe that relational density is a much more meaningful indicator of data quality (and value!) than others, like quantity, completeness or coherence. It used to be an heretic position in the data community, but times are changing and it’s all very intriguing for me since I’ve been researching in this space (how to mine and emerge insights and information from highly dense relational networks) for many years. To me, bases represent an important tool to increase relational density both for the data and for the social network of contributors.

Very exciting times.

But stay tuned because other announcements are in the pipe and coming real soon!

Also, if you happen to be in San Francisco next Saturday (Nov 09, 2008), come to the Freebase HackDay, it’s free and you’ll get to see a lot of shiny new data toys.