As knowledge organization systems increase in complexity, we have increasing numbers of dimensions or features or ways to define the terms. Semantic networks help with bringing all those concepts together and tagging them so that you can look at them in lots of different ways.

The concepts are still the nodes. The nodes are represented by articles, paragraphs. What we are trying to do is to connect that information to our information; connect that kind of conceptual treatment to the document. That is the way you can surface them, search or display them, or browse, or we can offer a collection.

There are a lot of systems used. People talk about semantic networks but I haven’t seen anybody fully achieve them. Everybody wants to have a semantic web, a semantic leverage.To me, the work based on Tim Berners-Lee’s ideas has become so over-engineered that it will never be successful. But the concepts at the base of it all are really intriguing.

I just think that we will have to achieve them another way.

So, when we talk about ontologies, and when we talk about semantic networks, we are really talking about an ideal that has not yet been achieved. I am watching.

I think we ought to be ready to embrace ontologies and semantic webs when it is feasible. Right now it is feasible in the lab. It is not feasible for a serious collection. If you have 600,000 or 1.2 million documents, you can’t do it. Not now.

There are some prototypes. Jim Doe has done a perfect 5,000 documents and when he adds the 5,001st document, it crashes. So, if you have a collection of more than 5,000 documents , I suggest you best wait a bit and get yourself something that could become a semantic network, when somebody smart figures out the way to implement it. How you are going to do that is to tag your data well, get an ontology or get a thesaurus, or you could just have a taxonomy to start, but as long as you use those building blocks, you will be well positioned to move forward.

So, you can tag your data now, you can build your thesaurus or taxonomy now, but if someone tries to sell you a semantic network, I can sell you that bridge that goes across the Hudson River in New York. If you can buy the Brooklyn Bridge, you can buy a semantic network.

Marjorie M.K. Hlava
President, Access Innovations