T1: Scientific context

[RenskeW]

Explanation of the choices which were made in the design of the workflow and parameter values.

• SC1 Workflow design: Annotations on the design of the workflow and its components. Purpose of the workflow, why steps were included or excluded, the meaning of particular input parameters, etc.
• SC2 Entity annotations: The meaning of individual input and output data entities. Why were they chosen? How are the results interpreted?
• SC3 Workflow execution annotations: Annotations about a set of parameters in a particular workflow run. Allows to distinguish between the ROs of multiple workflow runs.

[RenskeW]

What is represented in CWLProv RO Bundle in RDF:

Scientific context is not represented in RDF. However:

• SC1 - Workflow design can be conveyed through label, doc, and intent fields in the Workflow and CommandLineTool documents and are present in packed.cwl.
• SC2 - Entity annotations, if supplied, are in primary-job.json, but there are no guidelines to do this in a structured way in the CWL standards v1.2.
• SC3 - Workflow execution annotations, if supplied, are in primary-job.json, but there are no guidelines to do this in a structured way in the CWL standards v1.2.

[RenskeW]

Results of analysis of RO-Crates converted by runcrate from CWLProv RO Bundles:

• SC1 - Workflow design annotations, if supplied, are present in packed.cwl.
• Scenario 1: Of all workflow components which can be annotated with label, doc, and intent fields, only the doc field of CommandLineTool documents is included in ro-crate-metadata.json.
• SC2, SC3: Given that primary-job.json is not included in the RO-Crate, input entity and workflow execution annotations, if supplied, are not present in the RO-Crate, and also not propagated to ro-crate-metadata.json.

[RenskeW]

Suggested enhancement 1:

Include doc, label, and intent fields of all workflow components for which these are allowed in the CWL Standards v1.2 (see also Table 1 of the CWLProv analysis):

[simleo]

https://github.com/ResearchObject/runcrate/issues/10 added conversion of doc, label and intent, so SC1 is now fully represented in the RO-Crate. Also, with https://github.com/ResearchObject/runcrate/pull/25 we now add the input object document to the RO-Crate, so we can say that SC2 and SC3 are represented, albeit unstructured.