Srinivas Reddy’s Weblog

My Tech Rants

Semantic Web Insights

leave a comment »

Having recently read the book  ‘Semantic Web for the working Ontologist’ , it helped me better understand the problems of the current web and the new solutions to solving them. I would like to share my insights from it.  While most Software Developers may not think of themselves as Ontologists, models and modeling is at the core of a lot of what we do.  While the problem domain of most software systems are more constrained, it still needs a good understanding of “what exists in a system”

So, what’s the problem with the current web? We have a lot of information/data, but very little knowledge. The current web is dumb. You only need to go to Wikipedia to get information on any topic. We would like to use the information to make meaningful associations and take reasonable contextual actions. To take a simple example, if you would like to visit a place there are different kinds of information you will be interested in – historic information, points of interest, accommodation, current events, driving directions, …   In the current web, you will need to visit multiple web sites, manually aggregate/associate different information to arrive at the information you really need. So how do we make the web smarter? This is the problem semantic web technologies try to solve using new modeling constructs and technology standards.

The fundamental concept of the web is ‘Anybody can say Anything about Any topic (AAA)’.  Given a ‘open world’ where information follows the ‘bazaar model’ with everyone trying to enforce their view of the world, defining a model to describe information in a highly flexible manner is important. If the model has to enable the web to be ‘smart’, it should allow ‘inferencing’ between the different modelling artifacts. To relate this to the earlier travel example, it should be possible based on the travel destination to make inferences on possible accommodation. Its defining such models and inferencing that are supported by Semantic web related standards such as RDF & OWL. RDF provides a graph based Data structure using the Subject/Predicate/Object paradigm to enable the modeling flexibility needed. A recent Java one session on this topic is also a good introduction on this topic.

The current mainstream approach to the information problem is to create mashups (be it yahoo pipes, popfly, google mashup editor,…) based on the service interfaces exposed by relational data models. To make the distributed web data smarter, semantics needs to be added to the data models at model layer (ie enhance the metadata) to enable meaningful inferences by the machine. For mashups to takeoff at a large scale such a semantic model is needed and its this distributed data integration problem that semantic web technologies are trying to solve. Some of the interesting applications to watch that are adapting such technologies are Freebase, DBPedia, Opencalais and Semantic-mediawiki to name a few.

While quest for knowledge is as old as human history, Semantic web technologies may well make this easier for the future generations. Given the exponential growth of web data (what will all the latest mobile gadgets,…) is set to continue the need for semantically richer models is greater than ever. While it will take some time for domain experts to understand the new modeling constructs and Software developers to leverage them, there may well be light at the end of the information overload tunnel

Written by srinivasreddy

September 7, 2008 at 7:48 pm

Posted in 1

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: