You must believe in spring - Bill Evans

May 2009 Archives

Did someone get lost?

| No TrackBacks
Today is a lovely day in the Netherlands and I was walking in Utrecht when I stumbled on this.

Thumbnail image for 30052009.jpg Thumbnail image for 300520091.jpg

Ok, they say that Holland is the Venice of the north, but this is a bit too much. What is a Venetian gondola doing in here?!? And what are those people thinking? Here a brief list of what they could have been thinking during their lovely trip:

  • ...look at that! I didn't know there were so many bikes in Venice...
  • ...it's not that warm here... strange, even in June I need to wear a jacket sometimes...
  • ...I read that in Italy the law it's very strict but here it's very easy to get some marijuana...
  • ...the sailor said it was just 10 minutes but I think it's taking a bit too long to go to San Marco's square...
  • ...wow, the Venetian dialect is very different than Italian... I speak Italian a bit but here I don't understand anything...
  • ...I didn't like the typical dish "nieuwe haring" they serve here around. Strange, they told me the Italian kitchen was one of the best...
tot ziens,

Hadoop in semantic web

| No TrackBacks
It has been few months that I'm busy writing my master thesis. I chose the problem of reasoning in semantic web over a massive amount of data.

The choice mainly lies on the fact that very soon the amount of data will grow to the point it will be impossible to reason over it using a single machine. Semantic web is increasing its popularity and very soon it can reach a level that the computational and storage power of single machines are not enough. Think about the traditional web: storing and managing all the information present on the web is simply science fiction.

This problem can be solved trying to parallelize the computation and this is what I am investigating on for my master thesis. More in specific what I am doing is to use the Hadoop framework to implement a RDFS/OWL reasoning. The task is not trivial because distributed algorithms give their best when we can partition the data and unfortunately this is not the case of semantic web where the data is heavily connected to each other. After I finished it I want to publish the results in a dedicated page on this website hoping someone will find it interesting. Let's see how it goes.