If you structure your code from the start so that it is easy for you and others to use, you save effort later, for example when publishing or after a review. The goal is code that is open and FAIR: free to use and at the same time findable, accessible, interoperable and reusable. The following points will help you achieve this.

Before you start

• If possible, choose an open programming language such as Julia, Python or R.

In the code repository

• Manage your code with version control, for example on GitHub or GitLab. This makes changes traceable and makes working together easier.
• Make your code understandable:
  • consistent file and folder names (see tips)
  • good readability and documentation
  • a README file
• Where possible, read and save data in open standard formats, e.g. CSV instead of .dta or .xlsx.
• Record which packages and program versions your code requires, for example in a renv.lock or requirements.txt file. For complex requirements, a code notebook or a container may be an option.
• Choose an open licence and store it as a LICENSE file.
• Add machine-readable metadata in a CITATION.cff or codemeta.json file, and list all authors with their ORCID.

When publishing

• Create a release, i.e. a fixed, versioned state of your code.
• Archive the release publicly in a repository such as Zenodo, which assigns a DOI. GitHub alone is not permanent. Find out more under Publish Open Code Yourself.
• Link your code to the associated data and publication, for example via their DOIs in the README file.