A few years ago, Terry Jones sold his Barcelona apartment so that he could single-mindedly pursue a rather radical idea. What if a database worked like Wikipedia—it was not only readable by everybody, but also writeable by everybody?
The problem with databases today is that they are silos of data. Even the databases on the Web are separated from one another and must speak to each other through highly-defined APIs. While most developers equate the flowering of machine-to-machine APIs with the open Web, Jones sees them as controlling. So Jones created a new kind of database called FluidDB (and a company, FluidInfo) with a much more flexible database language which can essentially create database mashups and allow developers to pull data together in unanticipated ways. FluidDb is a shared database for applications. “Imagine if applications with related data were no longer forced to store it in unrelated places,” says Jones.
In his demo at TC Disrupt, he showed an app he threw together called Tickery which is based on FluidDB. It allows you to put in any two Twitter account names and compare the overlapping set of people they follow. But Jones also pulled in data from other sources besides Twitter, such as TunkRank, which sliced the results by influence, and an app he built called Wemet.at which lets people indicate which of the people they follow on Twitter they’ve actually met in real life. He pulled in data from other sources as well, including CrunchBase.
Once in FluidDB, which acts as one giant repository, the data becomes more malleable. Anyone can add to it or query it in any way, even in ways unanticipated by the designers of the database.
On stage, Jones used Tickery to winnow down the overlapping people several investors are following until he ended up with one profile: his own. That was his way of announcing a new $800,000 round of seed funding from betaworks, Founder Collective, RRE, Roger Ehrenberg, and Joshua Schachter. Previously, he raised $540,000 from friends, family, and Esther Dyson.
“The future of data is a writeable future” says Jones, “where any person or application can come along and add information without asking for permission, where data is social, where data that is related lives in the same place.”
Q&A
Chi-hua Chien, Kleiner Perkins: How do you get people to put data i?
Jones: You can make your data more valuable by storing it in a context where it is more useful. Also a flexible permssion system in FluidDB, so you don’t have an edit war like in a wiki.
Who else has the guts to ask a question?
Yossi Vardi: I am concerned that it has some terrifying privacy consequences. I mean it is good, but we will have to kill you.
Jones: the best answer to that question if it was a question is that FluidDb has been written to be extremely simple. There is not a lot that can go wrong. Not a lot in FluidDB, tags
Don Dodge, Google: Security and privacy are fundamental issues for anyone building applications. The hard part is making the developers feel comfortable with it
Jones: I think that is an evolutionary process. The fact that I’ve met John Borthwick, I don’t really care if that is public
Vardi: Is Ron Conway here? Two years ago, he said nothing exists outside Silicon Valley. I want to hear his view. Actually, he told me , he said he was blind, now invested in 20 companies in NYC.
Jones: FluidInfo is in NY because of the vision of John Borthwick and Andrew Weissman
Bijan Sabet, Spark Capital: It is nice to see an entrepreneur who has committed so much of his life to a product. We like to see that. I think the privacy thing is an important issue.
Jones: We will opensource FluidDB so that people will be able to open the code and develop confidence that it is robust. It is very easy to use FluidDB in an incremental way as long as you have meta data
We took Esther Dyson’s Flickr photos and calculated the white balance, show me the photos that are overexposed.
FluidDB’s presentation is at 35:30.