Hello,
In the past years, several of us have experimented with using
text-based notebooks, as offered e.g. by Jupytext [1]. I find these
particularly suitable in teaching use cases, in particular for the
following reasons:
- it enables the use of text editor and other text tools (e.g. for
mass edits).
- it plays well with version control and collaborative forges.
- outputs are not stored, which saves space and makes notebooks more
stateless (e.g. students start each time with a fresh notebook with
no trace of former executions)
To improve the support for text-based notebooks by the Jupyter
ecosystem, it is desirable to not only have implementations, but a
standard for such notebooks. To this end, a group of people interested
in text notebooks for various use cases have taken the occasion of the
Jupyter community workshop on notebook formats [2] to prepare a
Jupyter Enhancement Proposal for such a standard [3].
Reviews and comments are most welcome there: [3]!
Once there will be such a standard, we can expect that implementations
of official parsers will follow, enabling native support of text-based
notebooks by tools such as nbconvert, nbgrader, jupyterlite, etc.
Enjoy your week-end,
Nicolas
[1]
https://jupytext.readthedocs.io/
[2]
https://blog.jupyter.org/jupyter-community-workshop-the-notebook-file-format-8133ed606118
[3]
https://github.com/jupyter/enhancement-proposals/pull/103/files
--
Nicolas M. Thiéry "Isil" <
Nicolas...@universite-paris-saclay.fr>
http://Nicolas.Thiery.name/