Skip to content

What is Linked Open Data (LOD)

January 19, 2010

“Linked Open Data” has been stirring a lot of discussion on the Web. My research in the past few years is very related to the LOD in many aspects. Here I tried to describe the main ideas of the Linked Open Data in simple words. I hope that junior students, businessmen or just tech curious people can benefit from this post somehow.

What’s in it for you?

For website owners, product manufacturers, online stores, or in general someone interested in publishing data: would you like pieces of your website to appear in other websites? For example, would you like your wine description to show in as many wine lists as possible? Do you have a product that is much better than the competitors, and would like to make it easy for people to compare it in some nice interface? Linked data will make it easier for other websites (retailers, reviews, search engines) to get to your data, to understand it correctly and display it more effectively. Would you like to allow and track DIRECT access to “objects” in your website? No more wondering about what part of the page the user is looking at, Linked Data goes beyond pages and promotes objects to first class citizens on the Web.

(Update: If you want a very nice, completely non-technical explanation, see the post by Michael Hausenblas using a metaphor to describe Linked Data.)

What is it?

So… in a few words, what is this Linked Data thing? What does it bring to the table?

  • unambiguous reference: Linked Data defines a way to refer to an object such that a computer program can identify it without ambiguity. For example, you can say: “I’m talking about Michael Jordan as Berkeley CS defines him, not the Chicago Bulls Michael Jordan.”
  • description: Linked Data also defines a standard way that one can describe (and read a description) for an object. For example, you may want to assess “Who is this Michael Jordan at Berkeley guy? Tell me more about him? What’s the difference between the two Michael Jordans? ”
  • browsing: More importantly, by using the mechanisms of unambiguous reference and description in conjunction, the Linked Data meme facilitates browsing to other objects. So one can say: “Oh… he’s a Machine Learning professor? Let me look at his co-authors.”
  • querying: But on a giant Web of Linked Data, browsing is just a manual way to explore information. To automate access to all this interconnected data, Linked Data also facilitates querying. For example, “Get me all Michael Jordan’s papers co-authored with Stanford professors.”

    How does it work?

    Now if you are tech oriented, you are probably interested in how can Linked Data accomplish all those things.

  • unambiguous reference: Every datum in the Linked Data Web is described by a globally unique identifier. The standard that defines how to create such identifiers is the Uniform Resource Identifier (URI). You may be already familiar with one type of URI: the URL that gives you unambiguous pointers to documents on the Web. The idea here is that those URIs now identify real world entities (e.g. a product in your store).

  • description: To retrieve a description for a URI, all you need to do is to make an HTTP connection to that URI. Much like a Resource Oriented Architecture, every URI in the Linked Data can be “dereferenced” into an object/resource description. But with Linked Data the descriptions are represented in the Resource Description Framework (RDF), so you know exactly how to process any resource description on the Web in a unified way.
  • browsing: By providing descriptions containing links to other dereferenceable URIs, the Linked Data meme provides a way for anyone to navigate from one description to the next.
  • querying: Since Linked Data advocates the use of RDF, there is a standard way to parse the data and associated query languages (such as SPARQL) to formulate and process queries on that data.

    So, that’s it?

    The 4 rules of Linked Data recommend best practices for you to represent and share your data.
    Following those suggestions will make it easier people to read and present your data, link to it, etc.

    But don’t get me wrong. That is not going to solve all problems in the world. As with any tech problem, solving one creates ten. There are still a bunch of challenges to be solved there.

    For example, the unambiguous reference is one-directional. A given URI represents only one concept, and therefore allows people to specify exactly what concept they refer to. But one concept can be described (e.g. by different people) by many URIs. Therefore, disambiguation (aka deduplication or reference reconciliation) is a scientific problem of interest for the Linked Data community.

    A few problems that I am interested in include:

  • facilitating LOD querying for regular Web users,
  • using Linked Data for enhancing exploratory search,
  • real-time extension of LOD with contextual dynamic relationships,
  • representing opinions, observations and comments as Linked Open Data to facilitate collective analysis,
    amongst other topics.

    Advertisements
  • Leave a Comment

    Leave a Reply

    Fill in your details below or click an icon to log in:

    WordPress.com Logo

    You are commenting using your WordPress.com account. Log Out / Change )

    Twitter picture

    You are commenting using your Twitter account. Log Out / Change )

    Facebook photo

    You are commenting using your Facebook account. Log Out / Change )

    Google+ photo

    You are commenting using your Google+ account. Log Out / Change )

    Connecting to %s

    %d bloggers like this: