Guide for Extensions
2 Best practices for producing valid the annotations
Below we describe some of the best practices that testbed owners should follow in order to create semantic annotation. Please note that the APIs provided in the section 1.6 already use/satisfy already use/satisfy the best practices. The best practices are:
- Do not use blank nodes. Create individuals for concepts used.
- Have meaningful URIs for Individuals.
- Respect the cardinality provided within the ontology
- Do not push unnecessary/fake/test data
- Remove unused concepts. They might cause validation errors. Further, you are not obliged to provide certain tuples that have predicates such as http://www.w3.org/2002/07/owl#equivalentProperty, http://www.w3.org/2002/07/owl#inverseOf
- Make sure that annotations for observations are not repeated. Meaning you do not replicate information. This will unnecessarily use the network resources.
- ssn:SensingDevice individuals should have identical URIs in observation graph and resource graph
- In observation realm, uuse both ssn:madeObservation and ssn:observedBy. This also applies to all other properties that have inverse relationships.
- Do not abuse iot-lite:Metadata by providing just any information about the sensing device.
- Get the annotations validated first before pushing data to FIESTA-IoT repository by using certification portal or validation services.