Common Workflow Language

The Common Workflow Language (CWL) is a standard for describing computational data-analysis workflows.^[1] Development of CWL is focused particularly on serving the data-intensive sciences, such as bioinformatics,^[2] medical imaging, astronomy, physics, and chemistry.

Common Workflow Language

The Common Workflow Language standards
CWL Logo
Abbreviation	CWL
Status	Published
Year started	10 July 2014 (2014-07-10)
Latest version	1.2 7 August 2020 (2020-08-07)
Related standards	BioCompute Object
License	Apache 2.0
Website	commonwl.org

Standard

A key goal of the CWL is to allow the creation of a workflow that is portable and thus may be run reproducibly in different computational environments.^[3]

The CWL originated from discussions in 2014 between Peter Amstutz, John Chilton, Nebojša Tijanić, and Michael R. Crusoe (at that time their respective affiliations were: Galaxy, Arvados, Seven Bridges, and Michigan State University) at the Open Bioinformatics Foundation BOSC 2014 codefest.

CWL is supported by multiple analysis runners and platforms^[4] such as Apache Airflow (via CWL-Airflow ^[5]), Arvados, Rabix,^[6] Cromwell workflow engine, Toil, REANA - Reusable Analyses and CWLEXEC for IBM Spectrum LSF, and was identified in 2017 as one of the future trends for bioinformatics pipeline development.^[2] Several additional analysis environments are currently implementing support for CWL including Pegasus^[7] and Galaxy.^[8]

Availability

The CWL Project^[9] is a multi-stakeholder working group consisting of both organizations and individuals. A member project of Software Freedom Conservancy, it publishes the CWL standards freely available via its GitHub repository under a permissive Apache License 2.0.