Loraine Lawson spoke with Vickie Farrell, Neoview Product Management, of Hewlett-Packard's Business Information Optimization about the business potential of semantic technology.
Lawson: Does HP have a product that uses or does semantic integration or offers aspects of that?
Farrell: You probably know that the W3C (World Wide Web Consortium) has had working groups adopting or reviewing adopting standards for a number of years. HP has been on many of those committees. For example, HP was well represented on the working group that approved the adoption of the RDF and OWL standards in 2004.
Probably the most significant thing that HP has done is provided Jena. It's an open source framework for building semantic Web applications. It incorporates RDF and owl APIs; it also includes a rules-based inference engine, it includes in-memory and persistent storage for the data. It includes the SPARQL query engine, and it's by far the most popularly used framework for developing applications. There have been over 200,000 downloads.
Today there's a rules interchange format standard that is being developed and reviewed on a W3C working group. So, HP is represented there. They're in the final stage of producing a member submission on a SPARQL update language, which has a lot of backing. They expect that'll probably be going through the standards process in the next year or so. That's probably the most significant thing that HP has done.
Our consultants who focus on information management for a number of years have been working with clients to get started, to develop a taxonomy that will lead to the ability to make use of semantic technologies over time.
Lawson: The thing that I'm really struggling with is to wrap my head around what semantic integration means. As I understand it, semantic integration is a different concept maybe based on some of the same fundamentals as the semantic Web. Can you define semantic integration?
Farrell: When I wrote in 2005 the article, The Need for Active Metadata Integration: The Hard-Boiled Truth, that you read, I typically shied away from the use of semantic or semantic technology because, at that time, I think it scared people. What I was calling "active metadata management," is probably what you're calling today - and what is a good term - "semantic integration."
The concept of semantic Web, you've probably seen references to the 2001 Scientific American Article written by Tim Berners-Lee, Ora Lassila and Jim Hendler. Basically what they defined was this concept of a semantic Web, where information is given well-defined meanings so that computers can relate information in a meaningful way, interpret it, and act on it appropriately.
Humans can do that because they understand context, relationships and they can infer implicit knowledge from explicit fact. Computers have not really been able to do that. Semantic technology is helping computers start to be able to do those kinds of things.
This is why I think so many people are confused about this. In the examples that you see, you book a hotel in San Francisco and a flight on the Web and it knows that you like Greek restaurants. It tells you recommendations for Greek restaurants in the neighborhood of your hotel. You kind of scratch you head and say, "How is that going to be relevant to my business?"
The reality is that this technology could be used behind a firewall for real business applications today. It can be used within your organization to integrate data. This could include data coming from outside. That's the leap I think you don't often see in what is written out there.