Semantic Web: Snake Oil or Balm for What Ails Us?

This blog is in response to Seth Grimes' recent post "Semantic Web Snake Oil."

Our Business Intelligence industry is held back by a chronic lack of techniques to unify information. I saw in semantic Web technology a means to address that. I don't know if the Semantic Web will ever happen. I watch it very closely and report when I find something that looks like it might work. It's sort of a lonely outpost. I do know that ontologies appear to be a superior way for representing information, sharing it, reasoning from it and avoiding lots of duplication of effort. I call this making the data smarter so the applications can be dumber.I'm a big-idea guy, but I'm not a dreamer. My clients appreciate that I can think out of the box, but they also value that I'm practical. I'm not Tim Berners-Lee; I don't have an army of people who are willing to work nights for free to help me realize my vision. I have to focus on what's solvable within a business timeframe, not a geological one. I don't really care if some Semantic Web researchers or standards organizations or academic conferences tag their Web sites. Who does? I do care that Seth associated the Semantic Web with the term "snake oil" because it has now made my life a little more difficult. I don't think Seth proved his case, either.

My definition of the Semantic Web is turning the Web into a knowledgebase that can be understood by people or machines. But that isn't my primary interest. For the time being, I'm just trying to find a replacement for our inadequate metadata techniques that amount to dead data in a drawer, neatly arranged. The reason we keep getting into this mess is that IT organizations focus on systems, knowledge workers focus on data. Metadata as we know it is designed to facilitate (or at least document) what's what from a systems perspective. It is almost totally useless for DOING work.

There are alternatives to the Semantic Web. Sony Computer Science Laboratory is positioning its "emergent semantics" as a self-organizing alternative to the W3C's Semantic Web that does not require any tagging at all. There is some interesting development in discovery of semantics through machine learning. If you really want to put down the Semantic Web, no one does it better than Clay Shirkey: "The people working on the Semantic Web greatly overestimate the value of deductive reasoning (a persistent theme in Artificial Intelligence projects generally.) The great popularizer of this error was Arthur Conan Doyle, whose Sherlock Holmes stories have done more damage to people's understanding of human intelligence than anyone other than Rene Descartes. Doyle has convinced generations of readers that what seriously smart people do when they think is to arrive at inevitable conclusions by linking antecedent facts. As Holmes famously put it 'when you have eliminated the impossible, whatever remains, however improbable, must be the truth.' This sentiment is attractive precisely because it describes a world simpler than our own. In the real world, we are usually operating with partial, inconclusive or context-sensitive information. When we have to make a decision based on this information, we guess, extrapolate, intuit, we do what we did last time, we do what we think our friends would do or what Jesus or Joan Jett would have done, we do all of those things and more, but we almost never use actual deductive logic."

Shirkey didn't really put down the Semantic Web, but it was still entertaining. He used a classic rhetorical technique of arguing something beside the point. Of course, people don't apply formal deductive logic to most of their decisions (see my blog about the gut), but the Semantic Web doesn't require them to. It draws those conclusions itself through graph queries, bringing together the various concepts. To me that that has great merit.

I know that Seth is a proponent of text analytics, and he seemed to imply it was a more viable approach (for what, I'm not sure). I do know that if you're analyzing text or you're classifying an ontology of RDF tags, ultimately you are creating a representation in some sort of binary structure that can be searched. Whether you use tags or discover them, it doesn't matter. Underneath the covers is some sort of knowledge graph, just a different way to get there.Our Business Intelligence industry is held back by a chronic lack of techniques to unify information. I saw in semantic Web technology a means to address that. I don't know if the Semantic Web will ever happen... I do know that ontologies appear to be a superior way for representing information, sharing it, reasoning from it and avoiding lots of duplication of effort. I call this making the data smarter so the applications can be dumber.