Metaweb launches semantic web database

By Vivek | March 9th, 2007 at 03:42 am ET         

Metaweb, the San Francisco, California based stealthy startup has finally given a glimpse of its first product - Freebase. From the early details, Freebase sounds like an ever expanding knowledgebase based on semantic web concepts. Any information fed into the Freebase gets analyzed and related to semantic categories created in the system. In some respects DabbleDB team has also been working in this direction like automatically mapping location data onto maps. However Freebase goes all they way by categorizing all the data at hand and making the interlinking possible based on semantics for faster and accurate data search.

Lets look at some possible cases. Search for Google at Freebase would bring up the company page with address, people, and other related information. All the information is categorized according to the content type. Address associated with address, company operational area with Industry, revenues with revenues,…….making it easy for the system to classify the details. Taking the case of “Operational Area” as “Internet” would bring up its own page which would have categories like “Protocols”. “Uses”, “Architecture”, “Marketing”. As you can clearly see this would be a folksonomy based categorization. As content keep coming in, Freebase keeps classifying and categorizing more and more data and in effect able to find more content of similar type, which in real world is not easy to achieve on web. Thinking again Freebase also feels like early days of Yahoo with the big difference ’caused by the use of semantics and all the people working towards associating the content with categories.

However Metaweb still has to deal with the real world of web that runs around 3 operations- Insert, Update, and Delete. Right now I am not sure how far in each of these changes affect the system. What are the rules around deleting an entry? How many links will be broken by such a change? How far down do the updates get reflected? I am pretty sure Metaweb team has answers for these basic transactions.

Right now Freebase is under private beta. You can always submit your ID for an invitation. If you get one, I would for sure be interested in listening to your views about the product.

New York Times and O’Reilly have couple of good writeups on Freebase.

Update: Chris at TouchstoneLive has few interesting points against Metaweb. His argument is that we as users might end up building another data silo at Metaweb. Instead the better idea is to go for the microformats approach and let search engines extract relevant information. Microformats as an idea has always appealed me but the lack of mainstream participation by websites will hurt the progress. Older sites already running in production mode have far lesser motivation to define and implement microformats in their environment. Of course this does mean that smart startups can identify such areas and leverage microformats and APIs to deliver high quality data. Again you need lots of interested parties to get this movement rolling just like Freebase needs.
Links:
Metaweb
Freebase