Return to ENVRI Community Home
IV actions model the processing information objects in the system. Every action is associated with at least one object. Actions cause state changes in the objects that participate in them.
The figure shows a collection of action types specified in the information viewpoint.
Add additional data according to a predefined schema (metadata schema). This partially overlaps with data annotations.
Perform Annotation of an information object. Adding structured or unstructured information to describe a data object.
There are two basic types of annotation: free text annotation and semantic annotation. free text annotation refers to adding short explanations or opinions to a text or drawing (dictionary definition). Semantic annotation refers to linking data to structured conceptual model (ontology)
Semantic annotation of persistent data with concepts of predefined local or global conceptual models.
In practices, this can be done by adding tags or a pointer to concepts within a conceptual model to the data. If the concepts are terms e.g., in an SKOS/RDF thesaurus, and published as linked data, then data annotation would mean to enter the URL of the term describing the meaning of the data.
There is no exact borderline between metadata and semantic annotation.
Semantic annotation of metadata with concepts of predefined local or global conceptual models.
This can be done by adding pointers from concepts within a conceptual model to the metadata. For instance, if concepts are terms of a SKOS thesaurus, identified by URIs and published as linked data, then annotation amounts to associating metadata with the terms' URIs.
Obtain a unique identifier and associate it to the data.
Establish a local or global model of interrelated concepts.
This may involve the following issues:
Replicate data to an additional data storage so it may be used to restore the original after a data loss event. Long-term preservation is a special type of backup.
Actions to verify the quality of data.
For example it may involve:
Quality checks can be carried out at different points in the chain of data lifecycle.
Quality checks can be supported by software tools for those processes which can be automated (e.g. statistic tolerance checks).
Describe the accessibility of a service or processes, which is available for reuse, the interfaces, the description of behavior and/or implemented algorithms
Execute a sequence of metadata / data request --> interpret result --> do a new request
Usually this sequence helps to deepen the knowledge about the data. Classically this sequence can:
It can be supported by special software that helps to carry out that sequence of data request and interpretation of results.
Review the data to be published, which will not likely be changed again.
The action triggers the change of the data state to be "finally reviewed". In practices, an annotation for such a state change should be recorded for provenance purposes. Usually, this is coupled with archiving and versioning actions.
To add a short explanation or opinion to an information object.
Execute transformation rules for values (mapping from one unit to another unit) or translation rules for concepts (translating the meaning from one conceptual model to another conceptual model, e.g. translating code lists).
Measure parameter(s) or observe an event. The performance of a measurement or observation produces measurement results.
Process data for the purposes of:
Data processes should be recorded as provenance instances.
Make data public accessible.
For example, this can be done by:
Make the registered metadata available to the public.
Send a request to a data store to retrieve required data.
In practice, there are two types of data query:
step 1: query/search metadata;
step 2: access data
For example, when using OGC services, it usually first invokes a web feature service to obtain feature descriptions, then a web map service can be invoked to obtain map images.
Requests can be directly sent to a service or distributed by a broker.
Send a request to metadata resources to retrieve metadata of interests.
Enter the metadata into a metadata catalogue.
Retrieve the reference to the specific set of objects that correspond to a set of annotation terms.
Linking of data objects whit structured information from predefined local or global conceptual models.
Specify the mapping rules of data and/or concepts.
These rules should be explicitly expressed using a language that can be processed by software.
A minimal set of mapping rules should include the following data:
specify design of investigation, including sampling design:
Specify the details of the method of observations/measurements.
For example, it may include the specification of a measurement device type and its settings, measurement/observation intervals.
Archive or preserve data in persistent manner to ensure continued accessibility and usability.
Automatically generate and store metadata about the actions and the data state changes as provenance instances.