High Semantics in a Down Economy – Jeff Pollock

It’s tough times out there for little companies swimming in big markets, and unlike some pundits, I don’t pretend that the Semantic Web has already “made it” as a quantifiable software market. In the Semantic Web for Dummies book, I’ve tried to offer a lot of practical advice about which uses of the Semantic Web are making people’s lives easier today. After almost 10 years of working on data semantics for business software, I can testify that your next “greatest thing since sliced bread” software invention better have a dead-simple proposition or else your ideas will be simply dead after a few years. In this market, at this time, the little companies better stay focused on simple, practical, easy-to-understand solutions to problems that lots of people have.
That’s why “high semantics” are a bad idea right now.  High semantics are software implementations that demand strict conformance to rigid and inflexible data models or ontologies, forcing developers to be a bit too neat. Deviating from these high semantics in a particular software system might produce wildly inaccurate results from regular application queries. Semantic Web purists who imagine a world of only RDF/OWL or those who believe that Ontologies will right all the wrongs of relational databases, UML and XML will surely be disappointed. The folks who are getting things right today are busy getting on with solving real problems, perhaps even using the Semantic Web to do it!
Low semantics are a bit scruffy, a tad messy, and definitely not pristine. Using RDF/OWL opportunistically means finding those sweet-spot areas that a graph data model, a wee-bit of inference, or a spot of URI/Web-based data can make a regular software application sing.
Yes, even in the down economy there’s still money for semantics. Just look at $100 billion-dollar-plus enterprise software market, Service-Oriented Architecture (SOA) and Business Intelligence (BI) alone account for close to $20B annually, that’s real money that businesses pay to operate and optimize their business processes. Now, with the rise of the Cloud Computing meme, and the increasing buzz around Master Data Management (MDM), the opportunity for better data semantics in business couldn’t be better. But heed the advice in Semantic Web for Dummies, if you try to buy or sell semantics for semantic’s sake, you’ll get nowhere. Instead, look for places where a little bit of scruffy, down-low semantics can help an already mainstream software solution become better.