It’s time to get real with Linked Data. The World Wide Web Consortium’s Linked Data Platform Working Group, convened almost a year ago, is on the case, with expectations by June to publish a last call working draft of the specification, and to have a final recommendation, the last stage of the W3C’s standards process, by early next year.
“The Linked Data Platform is expanding on the concept [originally] put forward by Tim Berners-Lee on his web site, to turn it into a specification,” says Arnaud J. Le Hors, co-chair of the working group and IBM’s Linked Data Standards Lead. He will address the work at this session during next month’s SemTechBiz conference in San Francisco.
Why the need to formalize Linked Data? While there is a fairly significant list of W3C standards around the Semantic Web, the more loosely-defined Linked Data has led to an environment where interoperability suffers. That’s because people are left to solve the same problems, such as those around publishing and retrieving data, over and over again, and they take different paths to get there, Le Hors says. The guides that are out there are just that, guides, with people free to use or ignore them, if they can even find them – which in itself isn’t easy to do for those who aren’t well-informed members of the community, he says.
The Linked Data Platform extends the model to provide the industry with a formal definition for read-write access to Linked Data; it mandates publishing data in a standard format, RDF, and using a standard protocol, HTTP, “which is completely symmetrical with the way the web works today, with HTML and HTTP,” Le Hors says.
Providing a formal definition will help the enterprise use Linked Data more effectively and more broadly, as it will be easier to develop tools and make it easier to create apps. “Today an issue with Linked Data is there aren’t too many tools available, so it’s difficult for very practical enterprise companies to use it,” he says. And, importantly, no Linked Data that an enterprise or other organization already has been leveraging need be left behind: Says Le Hors, “There are certain features at the technical level in the spec that are attuned to trying to make it possible for people to use their existing Linked Data with the Platform specification.”
Ready to help with the embrace of Linked Data, and key to its broad adoption, are several open source implementations, like Apache Marmotta. “That will make it very easy for people to use the specification when they can download an open source implementation,” he says. The Linked Data Platform workgroup counts among its members reps from Oracle, EMC, and Fujitsu, as well as IBM, pointing to a coming wave of commercial tools that support the spec, too.
IBM has several products that depend on it, and is coordinating the changes it’s seeing in the spec with the product groups that are implementing it, such as IBM Rational and Tivoli. “In the Linked Data work we do, from the IBM point of view, it is related to the Open Services Lifecycle Collaboration initiative whose supporting organizations include Big Blue and many other vendors, as well as end users like GM that are looking to use Linked Data and the Linked Data Platform as the foundation layer for the work of standardizing the way that software lifecycle tools can share data.
Linked Data today, Le Hors thinks, for most, is more about accessing data in a distributed way, about performing application and data integration, that is completely compatible with web architecture, less about built-in, full-tilt semantic web capabilities. “People haven’t quite appreciated that yet and the specification we are developing lets you do that,” he says.
So for the present the work is focused to publishing data in a standard format with a standard protocol vs. “talking about all that good [semantic layer] stuff” that, he says, tends to scare people away, at least for now. But, hopefully, not forever. “We believe using Linked Data today leaves the door open for future development that takes advantage of the semantic web capability built into Linked Data,” he says. “When the basic layer is completely understood and people start using it, you can bring up the semantic web stuff that builds on it – that now that you have data in that format, RDF, and that model, now you can do all this good stuff with inferencing and reasoners on top of it, and it’s a much easier learning curve.”
To register for SemTechBiz and learn more about it, go here.