A few days ago, I got the wind knocked out of me with one of those rare “aha” moments. A headline caught my eye: “The State of Linked Data in 2010” by ReadWriteWeb. Being a curious sort, I read the blog post, and then spent the next few hours (days) exploring what Linked Data is. OMG.
Now I realize this may be old news to some, just as the WWW was when I got on board in 1993. But in 1993, I bought the first Mac built for online connection, and the good folks at Apple took days to successfully get me online as it was still such a novelty. So I’m thinking that Linked Data (also referred to as the Semantic Web) may still be an unknown idea to many of you as well.
Last May I was wowed by the concept behind Wolfram|Alpha, and I wrote on it here as it was launching. Since then, I haven’t heard that much about it (it’s a private effort), and I personally don’t believe it works that well.
Linked Data, an open-source movement, is more likely to be the real start of the answer of how inter-relationships between data can be seen and accessed on the Web.
Simple, one-dimensional searches are fine today, but what happens when your query is complex? Here’s an example I saw about football (I know, how unlike me!): you want to know which football players went to the University of Texas, Austin and played for the Dallas Cowboys as cornerback. What you’re really asking is that two databases – UT and Cowboys – talk to each other. But unless they’re connected in some way, they can’t talk. And that’s what Linked Data is all about – forming those connections through shared databases. (To see more – for those of you who are tech-inclined – visit this link.)
Today, it’s a little cumbersome to “code” the data for these linked databases, but when it becomes more automated, it will be quite powerful and I think more people and organizations will participate. Currently, the US government allows all federally collected data – data that we as taxpayers pay them to collect – to be shared across departments. It’s early days, but visit www.data.gov to see what the US government is doing. The Brits are also committed to this effort.
A company called Talis has a great example of how its platform is working for the BBC and I believe if you watch the short video, you’ll get excited about the potential of this new movement as an end-user.
What are the implications for other types of businesses? I can’t imagine they will share their proprietary information, but I do imagine that there will be “intranet-like” Linked Data resources where queries will merge public data with their private data and data from their vendors (such as agency media buys). Internal politics regarding who “owns” which data sources may be the biggest hurdle for full adoption.
I know that there will be more to write on all of this over time, but I would highly recommend a blog post by Scott Brinker. In his visual (below), you can see that he lays out seven business models going from indirect-to-direct revenue models and from ‘raw data’ delivery to ‘as an application’ delivery.
Are you as wowed as I am? What are your thoughts?