Funnily enough, I've been thinking lately that it might be nice for journals to publish annotated Jupyter notebooks or similar instead of conventional prose-based methods sections.