I've begun using the terms Weak Semantics and Strong Semantics to help people understand the difference and the application state of the art. The terms are obviously derived from Weak and Strong AI discussions. The terms have been used before but not, i think, in quite this way.
Weak Semantics are Flickr tags ... simple taxonomic annotation
Strong Semantics are fully constituted ontologically complete triple stores
WEAK----------------------------------------------------------STRONG
Applications in the world are somewhere along a continuum between weak and strong. Most of the tagging applications and a lot of the social networking apps we are seeing and working on today fall on the weaker side of the middle of the scale.
Research efforts have typically deal with stronger semantics which ae harder to implement and manage. In particular they are hard to get performance out of.
I've noticed that more and more of our efforts fall on the left side of the scale. They are easier to understand by our partners, easier to implement, and quicker to show results. Butthe time will come for efforts on the right side of the scale. There has to be a committment to solving very hard problems with semantics to get there though and they take a long time.
Bill
1 comment:
Bill, could not agree more.
Two years ago I picked up from fellow museum researchers similar distinction: soft and hard semantics. The former where tags & co., the latter where ontologies, etc.
But trying to push that in scientific publications was somehow tough. Eventually, there is not too much (or none at all) mentioned about this naming scheme in my thesis.
Post a Comment