Esther Dyson's PC Forum 2003 with Jon Matonis

The following is from Esther Dyson's PC Forum 2003 gathering at The Fairmont Scottsdale Princess on March 23-25th, 2003. Photos are courtesy of Dan Bricklin.

As a counterpoint discussion to a single unified database such as Oracle, I was featured on a panel named "Beyond Relational: What and Where" along with Zvi Schreiber (CEO and Founder, Unicorn Solutions), Kevin Turner (President & CEO, Sam's Club/Walmart), and Maria Martinez (CEO, Embrace Technology). At the time, I was the CEO of Network Inference, a business intelligence software company in London.








from Esther Dyson's Monthly Report Release 1.0
Volume 21, No. 3 (March 23, 2003)
Showcase Panels, pp. 88-89

Onotologies: Beyond database

Ontologies truly make data come alive; just like living systems (or profitable companies), they produce more than they take in. If you feed an ontology a given amount of data, its reasoning powers can produce more information than you entered. This means not just crunching data to produce an invoice, but crunching models to make inferences. That enables users to do everything from data integration and re-engineering to complex modeling and design tasks - not only for physical products, but also for commercial relationships, workflows and perhaps one day of business processes. The models in themselves are interesting, but what makes them useful is that they can be used as tools both to implement processes in software, and to redesign what has been modeled.

In the old days, information was painstakingly entered into ontologies by domain experts, such as those who worked at Cycorp, the leading ontology research project that is now a for-profit company (though it still does a lot of research). Today, much of the information comes through Web services interfaces - the best way of handling the complexity and volumes of data that make ontologies both useful and in the long run necessary. Without data representation standards, the difficulties of seamlessly integrating the processing capabilities of traditional systems and the reasoning power of ontology-based systems would be almost insurmountable. Beyond that, the spread of ontologies will require not just agreement on technology, but also on domain expertise, as Jon Matonis of Network Inference notes below.

Jon Matonis, Network Inference: Ontology requires harmony

As it happens, one of the tools enLeague is using at Coca-Cola is Network Inference's ontology and logic layer engine, Cerebra. Based on FaCT, one of the leading academic description-logic engines, Cerebra has been refined to optimize for commercial capabilities such as speed, scalability and predictability. The company has a number of commercial customers in addition to enLeague, including Clinician Support Technology (medical), and Qinetiq, the commercial arm of the UK Ministry of Defense, specializing in defense, aviation, and energy projects. Network Inference is also working with IBM's T.J.Watson Research Lab.

Network Inference ceo Jon Matonis earlier worked at VeriSign, where he learned about the importance of trust and of bottom-up networks. He sees vertical domain experts ratifying objective ontologies and also creating a separate cottage industry for ongoing ontology development. "Pharmaceuticals, clinical healthcare, and financial services are early adopters for Network Inference because of the existence of explicit, mutually agreed-upon global ontologies," he says.