Monthly Archives for May 2007

Sparqling 5 years of semantic web evolution

The Semantic Web School has monitored press releases about semantic technologies and related stuff over the past 5 years. A collection of about 1.200 links and summaries on articles are the result. These articles were tagged and those tags are mainly linked to wikipedia – so this might be an interesting data collection for some web applications…

Inspired by the work of of LinkingOpenData (W3C SWEO) and dbpedia we´re happy to announce that now there is a Sparql endpoint for that data which is open for the public.  Ideas for some mashups are:

  • Showing on a timeline which topics came and went
  • Publishing a list of press releases linked with wikipedia articles
  • Extracting names of companies which are mentioned in the articles and calculating a tag cloud for them

Any other ideas?

Yahoo Researcher Declares Semantic Web Dead – and reborn again…

When Mor Naaman from Yahoo said in a special track on Web 3.0 at WWW2007 that the “Semantic Web” is dead, he obviously tried to attract attention. Nevertheless, in my opinion he is absolutely right – there is no chance to “teach” people to annotate web content in a more sophisticated way than “social tagging” (and I´m pretty sure that also in the future it will always be a small community which will tag their content).

But in one point Mor Naaman missed the point: The “Semantic Web” was always there, under-cover more or less. Living in a tin with a lousy HTML-lid. And inside the tin there has always been enough semantics. There is no need to re-invent the data models, the namespaces, the ontologies (at least for most of the basic “things”) as Naaman proposes in his talk (slide 13). How easily all the existing semantics can be released and mapped against the “Semantic Web” (and suddenly it was born again 😉 ) is demonstrated by projects like [1] or [2].

RDF-Data is the Next Intel Inside…

…is the ultimative design pattern for the Semantic Web. In search for high quality data W3C provides a platform called “Linking Open Data on the Semantic Web“. Some basic datasets like geographic or encyclopedic information are already available. The more meshups from that datasets people will build the more data will be available, and so on…

Around two years ago virtually NO data for the semantic web was available, therefore no applications…

The RDF-Web has just started…

Yet another search interface for wikipedia?

It has always been worth using exalead from time to time to “exalead” (somehow that verb isn´t that popular as “to google”) one´s favourite search phrases (you usually type in when you start your computer in the morning ;-). That´s because exalead finds sometimes really “new” websites for you when you´re a regular google-user. So, until now exalead was rather interesting as an alternative like alcohol-free beer – give it a chance!

Now there is a reason to go to exalead: Try out exalead´s search over wikipedia and you will have a similiar experience as you have with dbpedia.

You´ll be able to refine your search phrase, therefore exalead offers even typed tagclouds (different colors mean different types of associations, like people or places). A nice Web 2.0-like GUI and high performance convinces finally.

One last question: What if dbpedia and exalead would combine their different approaches (dbpedia tripled wikipedia, exalead still remains in the area of automatic text-extraction) to put even more semantics into wikipedia?

Semantic Search for the WWW2007

When I tried to figure out which tendencies, hot topics and trends at the upcoming WWW2007 will be discussed, I first looked up the official website of the conference. I thought to myself – why isn´t there a semantic search implemented for the WWW2007? But the weblog revealed: There is one!

After doing some research with this really well implemented tool, I´m really impressed: Not only by the possibilities the search engine offers ( also searching the archive is there), but also by the relative importance the semantic web has gotten by 2007.

Meaning-based search engine

The recently published Top 100 Alternative Search Engines are mainly covering bad copies of Google. The reason is, that (1) no semantics/meaning of the content is being extracted by most of the “alternatives” and (2) they don´t focus on special domains. Therefore Google won´t be topped in the next few years… But, if you take a look at Cognition Search you will find a really interesting new way to search for information (or is it at this stage even knowledge?): First you can select a special domain, you want to search in, second you can select resources and then a really helpful way to formulate your search phrase will amaze you…