March | 2009 |The Semantic Puzzle

Linked Data is not owl:sameAs Semantic Web

March 30, 2009 by Andreas Blumauer, Category: Linked Data & Open Data, Search Engines

While some people work heavily on the extension of the semantic web infrastructure, like Talis Connected Commons or OpenLink´s Amazon EC2 Instantiation others have started to bring the semantic web closer to the developers and therefore to a much broader audience: They offer search facilities or Linked Data Navigators like OpenLink´s Entity Finder or DERI´s VisiNav.

Those kind of applications should not be confused with “semantic web” end-user-applications like Google´s Wonderwheel or INTSPEI´s Cloudlet: To add some semantics to existing user-interfaces can be helpful and obviously users are ready for such experiments, but of course this is NOT the innovation which the semantic web will bring but it is a very important step to be taken in parallel with the linked data initiative.

Let´s take a look at Cloudlet: This tool is an easy-to-use free Firefox extension that adds context-sensitive tag clouds to the most popular search engines and helps people more efficiently navigate through their search results. The previous version of Search Cloudlet worked with Google and Yahoo; the new version also works with Twitter. It adds Tag Clouds, Author Clouds, Recipient Clouds and Hashtag Clouds to Twitter search, Twitter user profiles and home pages. See some reviews on this popular tool.

Cloudlet is a child of the Web. INTSPEI has learned all lessons from Web 2.0 especially how to promote ideas using the blogosphere and how to identify market trends as early as possible, and it generates some added value for the users which is obvious. Sure, it doesn´t make use of linked data yet, but as a typical representative of the fast growing “semantic search evolution” it reminds me on Chris Welty´s famous insight: “In the Semantic Web, it is not the Semantic which is new, it is the Web which is new.”

Web 1.0 was the WWW without tons of network effects. Web 2.0 changed that a lot.

Linked Data is not the Semantic Web, it´s the basement for it. From a software developer´s and an IT archictect´s perspective it might seem as those two concepts were the same. But this community represents a very small percentage of all web-users.

So where is the User´s Web in the Linked Data architecture? If you´re looking at TimBL´s Linked Data principles one can clearly see that this is a “Web” for developers.

But things evolve. And some Web companies will jump on the bandwagon and will, for instance, improve their tagclouds, their semantic search, their recommender systems (Twine?) or their similarity search a lot by making use of linked data.

Like semantic search becomes mainstream (or call it “semantic search 2.0″) right now, then (in about three years, I guess) linked data will become part of a lot of mainstream applications. Linked data will generate tons of new network effects, maybe even new business models, it won´t be avant-garde anymore. It will be part of the Semantic Web.

the next google

March 25, 2009 by Thomas Thurner, Category: Search Engines, Software Development, Tools & Software

: Image via Wikipedia

Maybe you have noticed it already; today in the morning something new appeared at Google’s search engine interface: A bunch of corresponding search-suggestions based on your search query. Google spoke about this enhancement:

Starting today, we’re deploying a new technology that can better understand associations and concepts related to your search, and one of its first applications lets us offer you even more useful related searches (the terms found at the bottom, and sometimes at the top, of the search results page).

I tried it. So, if you type in “time travel” you also get search proposals like “theory of relativity time travel” or “wormhole time travel”. Google annouced, that the service is available in various languages. The direct test with German is a little disillusioning: Searching for “zeit reise” (which is the same concept as above, in german) leads to alternative searches like “reisen 50er jahren” (travel 50ies) and “reisen im mittelalter” (travel in the medieval).

Even if this semantic-like extension of the basis search function still needs some tuning, the point is getting clearer: Also Google is doing developments to get more meaningful results into their search algorithms. And parts of the semantic methodology are finding their way into mainstream services like search engines – as we have seen with Wolfram Alpha some days ago. So keep your eyes open – maybe next morning you’ll find another piece of the semantic puzzle embedded into one of your favorite web-apps.

Open now: LOD Triplification Challenge 2009

March 23, 2009 by Andreas Blumauer, Category: Conferences & Events, Linked Data & Open Data

The yearly organized Linking Open Data Triplification Challenge (as part of this year´s I-Semantics conference, 2 – 4 September 2009, Graz/Austria) awards prizes to the most promising triplifications of existing Web applications, Websites and data sets.

The challenge (Patron: Sir Tim Berners-Lee) is open to anyone interested in applying Semantic Web and Linked Data technologies. We envision submissions such as following:

Applications of Linked Data tools and techniques such as for example Triplify, Virtuoso or D2RQ on custom Web applications and data sets exposing a large quantity and variety of content.
Implementations of exporters and mappers from existing content repository formats (such as mbox mailing list archives, Bib Te X, XML-Schemes etc.) into RDF and Linked Data.
Adoptions / configurations of Triplify for standard Web applications, such as for example Wikis, Weblogs, Webshops, Forums, Web-Gallery, ERP/CRM systems and Web-calendar software. You can find popular Web applications for example at Source Forge.
Portings of the Triplify script into other Web application programming languages such as Python, Ruby, Perl, ASP. The Triplify script is very small (<300 lines of code) however, the port should be as compatible as possible with the current reference implementation but integrate well with the environment given by the programming language.
Applications showcasing the benefits of Linked Data to end-users such as for information syndication, specialized search, browsing or augmentation of content.

Submissions should consist of a two page description in JUCS format of the application, accompanied by (a link to) the software source code and a link to an online demo. The descriptions should be submitted electronically via email to Michael Hausenblas with the subject Triplification Challenge Submission by May 30th, 2009.

	if not otherwise specified