Cloud Solutions for Open Software
Authors/Creators
- 1. University of Bologna
Description
This work at the Univeristy of Bologna (UNIBO) is intended as a systematic literature review and a benchmark on the main cloud-based services with environment and system specifications for code reproducibility, mostly cloud notebooks, up to March 2023. We first screened the available tools for supported languages, costs, collaboration, set-up modes, version control allowance, PID (Persistent Identifier) attribution, licensing, accessibility, privacy, metadata, if Open Source, and long-term preservation strategy. We conducted this evaluation to identify a tool compliant with the FAIR principles (particularly those declined for Research Software from FAIR4RS) while facilitating code reproducibility. A tailored focus was on Code Ocean, a computational workbench used to associate code with scientific publications. The study highlighted that the landscape of tools facilitating computational reproducibility is varied, and it is still challenging to find a single solution presenting all the features making it both usable and adherent to the FAIR principles to ensure seamless sharing of code that is reproducible.
Background information:
In the era of data-driven research, software has become an essential tool for data generation and analysis. However, it cannot just be considered instrumental: initiatives such as the Software Sustainability Institute advocate recognizing software as a research output and clearly stating how its management can benefit research, especially if shared and referenced in scientific publications.
The emphasis on improving research reproducibility, which led to the conceptualization of the FAIR principles for research data in 2016, is also becoming relevant for research software (e.g., FAIR4RS – FAIR Principles for Research Software).
Development strategies are already consolidated in the community, yet the discussion on methods, tools, and guidelines concerning the computational reproducibility of published results is still open. The main barriers to overcome are connected not only to software’s sharing and preservation strategies (lack of source code, also along with raw data and protocols) but also to the need to consider its specific features, e.g., executability, composite nature, and continuous evolution and versioning (by recording relevant information about versions, parameters, and runtime environments).
Various tools, such as workflow management systems, digital protocols systems, packaging, and containerization platforms, can facilitate code sharing and reduce barriers to computational reproducibility. Some publishers endorse adopting some of these tools to improve peer reviewing of research code.
Files
OpenSoftware_02032023_Caldoni+Coppini.pdf
Files
(2.8 MB)
| Name | Size | Download all |
|---|---|---|
|
md5:963a5844ed82b2337b0254fdfd7e44e1
|
2.8 MB | Preview Download |
Additional details
Dates
- Available
-
2023-03-02