Getting Man & Machine to Communicate

"Semantics is hot," declares Seth Grimes, "but only in a geeky sort of way." And then Grimes offers 11 different ways semantics is impacting -- search. In fact Semantic Search takes two geeky subjects and, when conjoined, attempts to become a precision tool for finding what the user really wants.

But what does the user really want? And here is the crux of the problem. The machine needs to discern the true question the user is asking when searching -- but the challenge is the machine and man do not share the same language. Man wants to ask a question -- well, naturally -- meaning highly unstructure. And machines need people to be structured. NLP, Natural Language Processing, is where the machine finally understands human-speak. (We're a long way off still -- but heading in the right direction.)

It helps if we know Boolean -- but few do. Really smart people I know may know that putting a + sign between words means AND, but few know that the - (minus) sign means NOT. Nor do they think to put phrases in "". Really smart people I know forget that really smart people don't know Boolean -- which leaves many really smart -- and not so smart people --having a dismal time when they are searching for things. (To prove my point -- this article on 10 Google Search Tricks has been in the top 10 in popularity for the last 10 days.)

Which circles back to why semantics is hot. Generating semantic metadata means machines can locate content about this -- when a person asks for that. It presents content to us by clustering content under categories -- so we may browse through and find what are looking for. Semantic Metadata is a great first step to getting man and machine to communicate.