JupyterLab 4.0: a development environment for education and research

Most of the negative comments that I have read seem to be related to the very concept of a computational notebook interface (that owes most of its success to Mathematica going back to the end of the 1980s, and that is a great form of Literate Programming) with its pros and cons (including the out of order execution), rather than to the specific Jupyter lab implementation.

What seems to be very specific to Jupyter and more susceptible to criticism in my opinion is that Jupyter lab, now at version 4.0, seems to be still in a "young" phase of development, with a much stronger emphasis on adding features wrt fixing issues. As a consequence, the project has a very large number (>2k) of bugs open, some of which very old and some outstanding (e.g. the fact that printing notebooks from the browser is severely broken, and has been so almost forever, which is really weird given that we talk of literate programming). Very poor svg management is another notable issue, as the poor management of embedded images for markdown cells.

In general the ability to store files embedded in the notebooks to be used as datasets by the code or by markdown cells (e.g. for images) would be useful to enable self contained notebooks. The ability to associate a notebook with "requirements" metadata would also be interesting.

Another thing specific to Jupyter lab that seems worth mentioning is that other implementations of the Notebook interface have means to group cells (e.g. in notebook sections) and to associate specific context (scoping rules) to the sections.