INFO 202 Fall 08 Blog

I read a paper today that discussed a step covered in class today on document engineering.  We learned how the Berkeley Calendar Network team manually harvested and consolidated tables of terms in a huge excel spreadsheet.  The IEEE paper I read argues for semi-automating this process of deriving homonyms in IS-A relationships, for instance, and integrating the terms with the http://wordnet.princeton.edu ontology.

The buzzword-laden paper continues to argue for creation of a working Semantic Web by harnessing the large quantity of structured, “Deep Web” data. The, “Deep Web” (unindexed by conventional search engines) contains ~over 4 orders of magnitude of data than the, “Surface Web” and some data is structured in databases.  The Semantic web, they claim, has been hampered by difficulty in manually creating large OWL and RDF ontologies, and harvesting the richer potential of the Deep Web points to a possible solution:
Semantic Web + Deep Web-Ontology-aware browser.

Ironically, the paper, itself, is buried in the deep web:
http://ieeexplore.ieee.org/iel5/2/4623205/04623231.pdf?tp=&arnumber=4623231&isnumber=4623205

There’s an interesting article at Read/Write Web about the nascent market for semantic apps, and the dearth of apps using RDF. They speculate that perhaps using RDF, and imposing order, is too hard compared with discovering order (Tripit, Twine, others).

Is it too late to create structured data on the web? Are we stuck with making sense of the mess that’s there now?