I have been a fan of ontology based models for some time now. The danger here is that the ontology and AI paradigm has been around for sometime and with that "perceived" expert opinions on how to apply ontologies to the blogosphere and semantic web can and will be flawed (even by the most respected technical gurus).
I believe that a new model and approach will emerge. One that will reflect the current environment and future needs not pandering to old academic papers and theories.
One area that I see clearly are new tools that will automate the ontology "seed" process. Having built nameservers in the past for keyword search and navigation I can tell you the taxonomy issue is HUGE.
RSS and new syndication technologies combined with search capabilities (not paid search) then overlaid ontop of an always on network will have a profound effect on user experience.
"What you saw in first-gen solutions [was] Ask Jeeves—fragile ontologies that required tons and tons of human labor to keep them updated. The next gen has succeeded because they have a more automated way of keeping the taxonomy up-to-date and being able to give a consistently good-quality solution to the end customer."
This important addition to already developed ontology paradigm will change the nature and definition of "search eninge optimization"
http://furrier.typepad.com/john_furrier/2004/12/search_engine_o.html
Comments