The yearly World Wide Web conferences are highlights for my research: every time again, the world’s most fascinating people meet to discuss novel ideas. This year’s edition moved to Seoul, and I happily represented Ghent University for the third time, together with my colleagues. In addition to hosting the WS‑REST2014 workshop, I presented Linked Data Fragments at LDOW2014. The combination of these workshops represents for me what is important to move the Web forward: flexible data and API access for automated clients.

19 April 2014

The WWW conference series have so far been very rewarding for me, and Seoul was no exception to this. I had fruitful conversations with many of the attendees, and picked up several ideas for the months to come. Content-wise, I get the most inspiration out of the workshops that accompany the conference. The discussions that happen there are in my opinion the most interesting, even though there is unquestionably great research in the main tracks as well.

The drawback of many workshops is that you can only pick two of your favourite ones to attend. Fortunately, LDOW2014 and WS-REST2014 were scheduled on different days this year, so I could attend both.

WS-REST2014: REST and APIs

The WS-REST workshop, or “International Workshop on Web APIs and RESTful Design” as we rechristened it this year, has special meaning to me. Three years ago, at the WWW conference in Hyderabad, I arrived by accident at the 2011 edition of the workshop. I had nothing planned for the day and I was curious to see how the state of giving things pretty URLs was advancing. (I didn’t actually know anything about REST back then.) My PhD research was just starting and honestly, I was out of inspiration at the time. A presentation by Thomas Steiner on the hypermedia constraint caught my attention—and the rest is history. Tom and I have been working on many fun projects and papers ever since. At the next edition of WS-REST, I presented our work on RESTdesc, followed by Distributed Affordance in 2013.

You can imagine my surprise when Erik Wilde asked me to take over WS-REST organizational duties from him, Cesare and Rosa. I didn’t have to think for long… ;-) And so I organized WS-REST 2014 together with Tom and Carlos. It was a though challenge, as the 2013 edition of the workshop saw a declining number of people. However, by widening the scope to include everything about APIs, and focusing on new challenges such as client design, we gathered an exciting line-up and crowd.

My personal highlights of the day include the keynote by Sam Goto, a very honest story about how he (and Google) learned about hypermedia APIs and found ways to describe them in Schema.org. Then came a wonderfully authentic Mike Amundsen performance, who managed to capture the essence of hypermedia clients with a phrase from the Pixar dog Dug: “I have just met you and I love you”. Truer words have seldom been spoken. I was also very excited to see Pete Gamache’s presentation on a very pragmatic generic hypermedia client. His approach for capturing hypermedia in a Ruby DSL is particularly elegant. I’d be interested to see how that grows, and I think there would be a good match with Hydra.

The audience seemed to have a great time, and so did we! A broad smile on the face of Cesare, the original workshop’s co-founder, seemed to indicate his blessing for another edition organized by us :-) Here are challenges for next year.

LDOW2014: Linked Data on the Web

I must admit that I’ve been looking forward to LDOW2014 for a very long time. Since I started working on Linked Data Fragments, I’ve been imagining how I could present it in front of the LDOW crowd. My slides are online, and you can check out the live demo too. The response has been very positive, with some interesting remarks and questions by the LDOW audience.

The Linked Data Fragments story, together with Stefan Scheglmann’s presentation on inverse link traversal, opened up the discussion of extending Linked Data with other hypermedia controls besides the links used for dereferencing. Given my background in REST APIs, this is something I strongly believe in: with more relevant controls hypermedia at their disposition, clients can engage in more intelligent interactions on the Web of Data.

While work on Linked Data querying has shown the potential of links, extending the control set to forms or URI templates allows support for more scenarios and lets us support existing scenarios more efficiently. In particular, finding subjects that match certain conditions is hard with Linked Data querying, because you browse by subject. In contrast, basic Linked Data Fragments expose triples of a dataset in all possible ways through hypermedia forms, which allows to answer questions that are hard with dereferencing, such as “give me all artists” or “who was born in Seoul?”

The next 25 years

The Web’s 25^th birthday was of course a theme at the WWW2014 conference. Tim Berners-Lee cut the Web’s birthday cake, and there were several ongoing discussions on which directions the future Web could go. I personally believe that scalable and flexible access to data & APIs are crucial for the Web to move forward. This is why I’ll be continuing research on semantic hypermedia, and if all goes well, you’ll see more about that at WWW2015 :-D