Two complimentary strands of semantic web
Strand 1: “semantic” aspect, powerful knowledge representation, database quality data, centralized workflows for knowledge management. E.g. oracle triple store. It has enterprise uses case, e.g. bio-medial study.
Strand 2: “web” aspect. Publish knowledge rather than text on the Web. Rooted in the original vision of the Semantic Web. while the promise look great, the real world use cases are fairly poorly understood.
Challenge: “Can strand 2 semantic web applications overcome the data chaos of the emergence semantic web”
Semantic Wiki lives in both strands. It inherits the web 2.0 nature from wiki and is quite easy to be adopted, and in the mean time, it has pretty good support to encode structured data using RDF.
On promising potential is that semantic wiki may enable ontology convergence. Note that without convergence, semantic data may be in chaos and thus less useful. In halo experiment, ontology convergence has been observed in collaborative annotation contributed by college students.
Several findings learned from
* user interface matters, (sure, semantic web developers should pay more attention to UI for better adoption)
* gardening matters (wikibots works, so does semantic wiki bots)
* user created ontology are not always well-designed (that’s why administrators are needed to clean up, but how to deal with such imperfectness and will that cause data chaos? )
* natural language is necessary to augment bare RDF(S) semantics (the “situation calculus” problem indeed is a good justification, as we cannot encode all in semantic web way, some free text may help fix the empty space in the absence of semantic wiki. )
Digital Aristotle for scientific knowledge – Halo project (2006): to build a question answer system that allows domain experts to build robust system for answering challenging and complex questions.
* the two strands of semantic web should meet each other.
* semantic wiki is one the applications that can bridge both strands
* halo demonstrated that by addressing hard AI problem using semantic wiki.
By Li Ding, Greetings from ISWC 2008