bollwyvl
commented
2 years ago Not speaking officially, but I don't think any core jupyter project is going to bend over backwards to support a specific, proprietary product without a very compelling reason. It boils down to a poor use of volunteer time to support SaaS/proprietary products that can be rendered unusable at any time.
Thus far, github, and some of the CI providers, have received additional support, due to the criticality of those services to ship jupyter software, but I don't see that in this case.
With the free Wolfram Engine, it is now possible to create Jupyter notebooks that run the same kernel as Mathematica. I haven't fully explored this, but it seems VERY attractive. I love Mathematica's symbolic engine, but I can't justify the cost as I only use it once in a great while. The free Wolfram Engine + Jupyter setup potentially solves this.
The challenge now is to convert Mathematica notebooks to Jupyter notebooks. I've reviewed a few discussions on Mathematica stack here and here, but neither has a solution using the Wolfram Engine. I stumbled onto "nbconvert" and thought it might be possible to do this on the Jupyter end (instead of through Mathematica/Engine).
I suspect A LOT of people will be shifting to this free Wolfram Engine + Jupyter framework. Not having a tool readily available to make the conversion seems to be the only drawback.