Best practice metadata and attribution

CWL Questions
1

Hi,

while trying to implement the recommended practices for proper attribution to software developers (e.g. 2.16. Best Practices — Common Workflow Language User Guide 0.1 documentation, 2.18. Metadata and Authorship — Common Workflow Language User Guide 0.1 documentation), I got a bit stuck.

Even for a simple “single-action” (one input, one output) CommandLineTool, there may be multiple options whom or what to list for citation, license, author, etc..

When wrapping a published tool, one would probably add (reproduce) the info for citation, license, code repo, etc. as much as possible from the original developers and publication. What if conda or docker dependencies (not developed by the original developers) are referenced?

Or are fields like author (also) meant for the person writing the CWL document?
That would probably be best suited for Workflow documents, that make use of a number of tools in a specific conjunction?

Thanks for suggestions.

2

I can relate to this! Currently we use the citation section for the authors of the CWL file CommandLineTool or workflow since the authors of the tool itself are usually not affiliated with the CWL file.
(But it could indeed possibly lead to incorrect citations when someone used the CWL version)

Next to this we try to put as much info for (actual) tool citation as we can in the SoftwareRequirement section.

For example:

hints:
  SoftwareRequirement:
    packages:
      sylph:
        version: ["0.8.1"]
        specs: ["identifiers.org/RRID:SCR_026478","https://anaconda.org/bioconda/sylph","https://doi.org/10.1038/s41587-024-02412-y"]
  DockerRequirement:
    dockerPull: quay.io/biocontainers/sylph:0.8.1--ha6fb395_0
1 Like
3

@mrc answered this here: extend Kallisto tools by Brilator · Pull Request #158 · common-workflow-library/bio-cwl-tools · GitHub

The citation, author, etc. refer to the CWL document, not the tool.

2 Likes