In EVERSE, we consider representative software use cases (pilots) from the five EOSC Science Clusters, to be used in a feedback loop between EVERSE and the scientific communities. These use cases are:
- SSHOC: UDPipe, a multi-language textual analysis pipeline of tools (local or as a service)
- ESCAPE: software for ML-enabled data compression (Baler), reconstruction (ACTS) and data analysis software built for Open Science (xAODAnaHelpers)
- Life sciences RI: software for secure and federated workflow orchestration and movement of encrypted data (The Workflow Execution Service backend)
- PANOSC: software used to enable heterogeneous computing (ALPACA) and for scaling to high performance computing (PIConGPU)
- ENVRI Community: ENVRI-HUB knowledge base and Virtual Research Environment
While software quality, as well as its dimensions and indicators (and therefore RSQKit personas and tasks) are meant to be general and cross-field, there are cluster-specific aspects due to both technical and cultural/work environment differences. These differences mean that the personas above will be supplemented by other roles in some clusters, and that certain dimensions will be considered more relevant by some of the personas in certain clusters.
In the ESCAPE Cluster, the role of RSE is only emerging recently; software developers are historically physicists who know the physics challenges that the software must solve very well but may not have received formal training in software development. For this reason, EVERSE can help with the recognition of (and with retaining) professional figures that focus on software development, and building career paths that include software-centered roles.
Given that many scientists in ESCAPE and PANOSC work in large, often distributed collaborations, there are additional roles that are interested in software quality, namely: host laboratories, experiment coordinators, data/open data managers and central/distributed computing site managers (the sites are where the software runs).
In the life sciences (LIFE-RI), there is a large community of users who do not code but who are still interested in software quality -- and they are not necessarily principal investigators, although they could be considered within that category.
Different clusters prioritize quality dimensions based on their specific needs and contexts. While all dimensions remain relevant, emphasis varies across the science clusters depending on their domain requirements, technological constraints, and community practices.
Note: Visual representation of cluster-specific prioritization will be developed for future versions of this framework.
We use directly the EOSC Science Cluster definitions, with each cluster having sub-views based on the individual research infrastructures or other well-defined science communities within them. This ensures that domain-specific requirements are captured while maintaining alignment with the overall EVERSE quality framework.