Home¶
Welcome to the Software and environments reproducibility working group homepage ! This working group is part of the French national reproducibility network.
On this website, you will find advices and best practices regarding the reproducibility of calculations (data production and management are excluded). We therefore cover software environments and software applications.
This is a collaborative effort based on our various experiences.
Reproducibility and open science¶
Reproducible research is essential to ensure the reliability and credibility of scientific results. By enabling independent verification of methods and findings, it strengthens transparency, reduces errors and biases, and improves overall research quality. Reproducibility also facilitates the reuse of data, code, and methodologies, accelerating scientific progress, fostering collaboration, and maximizing the impact of research across disciplines and society. Reproducing scientific results increases confidence in science, improves reliability and ultimately expands our collective knowledge, being a pillar of knowledge building.
Open science is a core pillar of reproducible research. By promoting open access to data, code, methods, and workflows, open science enables independent verification of results and ensures transparency throughout the research process. This openness allows reproducibility check, improves research quality, fosters collaboration, and accelerates scientific progress while increasing the societal impact and trustworthiness of scientific outcomes.
How to use this website¶
The website is designed as a guide, taking the form of a series of practical "cards" organised into three categories:
- 📖 definitions: the concepts and terms we wish to highlight and detail
- 🧭 overview: meta/summary-card focusing on a particular topic, providing an introduction to the more technical cards (e.g. about specific tools)
- ⚙️ technical cards: focus on a specific tool, pratice ...
You can easily identify each kind of card with the three symbols 📖, 🧭 and ⚙️.
About the working group¶
The main framework is the reproducibility of calculations (data production and management are excluded). This working group aims to produce guidelines and documentation of good practises around reproducibility in software developements.
By reading these cards, scientists may be able to publish reproducible results, providing all the artefacts needed by the reader of the publication to reproduce the results. Different levels of reproducibility or many time scale may be discussed here. Reproducibility levels may not be the same between statistical analysis and cybersecurity concerns! Similarly, it is often easier to reproduce a result some months later than some years!
So the main ideea is to guide the software developper towards the publication of reproducible results!
Another idea is to make links between good practises and technical documentation of a lot of tools by introducing difficulties and advicing technical solutions.
This working group drafts documentation sheets posted on this website (see section cards). These sheets deal with good practices for reproducibility, aimed at anyone writing code ranging from simple scripts to complex software. The sheets cover the points you need to know to make your computational output reproducible, from design to final distribution over time. They also present the tools that can facilitate this work, from the point of view of software development as well as distribution and execution.
In time, these sheets will also contain many references to existing publications and/or tutorials on the various subjects covered. The main idea is to help readers familiarise themselves with the tools available and find their way around a wide variety of concepts, the articulation and/or choice of which can be complex. The group meets on a monthly basis and contributes to the writing of the various fact sheets, while seeking external contributors who are specialists in a particular subject.
Contribute¶
If you want to contribute, you are welcome!
The group is open to anybody interested in computational reproducibility.
To join, subscribe to the mailing list and have a look to the contribute page.