September 15, 2011 Leave a comment
We talk about modeling a lot at Thetus. In fact, if you ask us what we do, we’ll tell you we are a modeling company that tackles complex problems. To us, models are a way of expressing the concepts, properties and relationships that describe a complex system. We use many concepts and the standards from the Semantic Web, and the types of models we build are ontologies to describe what things are, how they are related, and where they fit into the overall model.
The word ontology elicits a bit of eye rolling in some circles due to an abundance of enthusiasm for the potential of ontologies in the early years of the Semantic Web. Ontologies were heralded as the magic answer machines. Nevertheless, the power of describing concepts and relationships and following inferred connections through models to answer questions is real, but the model by itself is not an application. We’ve always taken a pragmatic approach toward how we use these models and how we build software to get value from them. Savanna, our analysis environment, uses models for everything from driving the user interface to link charting and modeling the analytical process. Models are simply rich building blocks for characterizing.
This modeling capability is especially useful for sociocultural analysts because the information becomes process-centric. Analysts can work through a problem and then be able trace the process that helped them arrive at the understanding they’ve gained. At Thetus, we’re interested in functional ontology that accommodates the constant change of real-world understanding even when it contradicts our models. Definitions aren’t simply one-sided–they’re multi-faceted and dynamic.
In other words, we define, but we don’t restrict.
For example, imagine that a sociocultural analyst is working to understand how regional alliances are formed and how they can be influenced. The analyst needs to employ a number of interconnected models to really characterize the environment to create the basis for analysis and, ultimately, to make an assessment. Analysts need to describe concepts and relationships that span social, economic, political and environmental factors. These factors intersect with temporal and geospatial vectors to form a rich model of the region. Some relationships and concepts will have concrete and observable manifestations, and others will have more subtle and inferred connections and associations.
The truth is, we’re unintentionally using models in our everyday lives every time we make a decision. For example, if we hold a dinner party, we must consider what kinds of guests we’re cooking for—some people are vegetarians and other people like meat–some people have food allergies and some people are just picky eaters. We’re also thinking about external environmental influences like weather (serving soups in the winter or BBQ and fresh tomato salad in the summer), or the availability of ingredients (corn on the cob might be harder to find in January). All these aspects contribute to the final dish, and it’s models that help us make the decision.
Life is full of complexity. Even dinner choices involve many factors and defy easy definition. We embrace this open-endedness because we know that the answers aren’t always set in stone.