We’re excited to announce the publication of our recent paper titled “A Terminology for Scientific Workflow Systems” to the journal “Future Generation Computer Systems”.
This work represents a joint effort by workflow developers and practitioners from the Workflows Community Initiative, who came together to address a recurring challenge: with hundreds of existing workflow management systems (WMSs), how can researchers compare them meaningfully and choose the right one for their needs?
In the paper, a flexible and descriptive terminology to characterize WMSs across five key axes is introduced:
- Workflow characteristics
- Composition
- Orchestration
- Data management
- Metadata capture
Each axis defines terms to describe high-level features, like how workflows are structured (e.g. task- vs data-driven), how they are composed (e.g. schema, API, GUI), how tasks are orchestrated, how data is moved and stored, and how execution metadata is captured and used.
As a demonstration, 23 widely used WMSs are classified, including (but not limited to) AiiDA, Nextflow, Snakemake, and Pegasus, using this shared vocabulary. We hope the terminology developed in this work becomes a shared reference for developers and users alike, and we invite the broader community to adopt and contribute to its future evolution.
Happy computing!
The AiiDA team