Tasks in COG

Tasks

Below are the important tasks we plan to carry out under this initialtive.

Content provision

Domain resources such as description of oceanographic facilities, concepts and measurement data are the blood and flesh on top of the semantic web infrustructure. We particularly encourage oceanographic domain experts to contribute their expertise on these resources. a wiki has been made available to facilitate this task of collaberatively acquiring knowledge in its original format.

Ontology management

Ontology is a conceptulisation of the domain, which can be formally defined using the W3C OWL specification, to provide pre-defined controled concepts and relationships for knowledge representation of the domain resourcs. We will provide a Collaberative Ocean ontology to underpin the rest of the semantic web activities such as semantic annotation and intelligent semantic processing of the annotated oceanogrphic resources.

Semantic annotation

Semantic annotation is like tagging resources using control vocabulary (ontology) in order to enrich the resources and improve their interoprobability. The RDF data generated, together with the ontology, forms part of the semantic web.

RDF storage

Technologies on semantic web repository play a more and more important role while the amount of the semantics data increases. It aims to provide a scaleable RDF storage mechanism, often using DBMS as a backend, coupled with remote management, semantics processing capacities (e.g. SPARQL query and Semantic Web inference).

Semantics processing

Typical processing includes semantic web query and reasoning

semantic web query - SPARQL becomes the W3C standard of the querying lauguage on the semantic web. Its syntax is similar to SQL but is optimised for the semantic web with the potencial to integrate inference engines for intelligent semantic web reasoning.

reasoning - This is all about generating new semantic web statements by applying domain rules over existing semantics, to allow machine to "think" by reasoning. Lots of inference engines have been made available to facilitate this promising task.