jupyter-lisp-notebook

[snunez1]

My common-lisp-jupyter (CLJ) installation on MS windows is broken, again.

I've had some useful feedback from users of lisp-stat that the on-ramp is just too steep. Statistics users come because they want to try a natively compiled statistical computing environment, only to learn they need to become lisp engineers too (emacs, slime, ASDF, common-lisp-jupyter, etc).

So I'm thinking a way to build a better onramp is to provide a docker image in the style of the jupyter-docker-stacks, e.g. community stack for jupyter-lisp-notebook, based on jupyter/minimal-notebook.

I know about CLJ and repo2docker and the provided docker image and want to ask if there was some good reason not to 'go with the flow' and base it on Jupyter Docker Stacks? Having something related and similar to a project like that means we could borrow the documentation and the usage patterns will be familiar to people crossing over from R, Python and Julia. Eventually we might be able to get them to host a lisp notebook.

The reason for this thread is to debate the pros & cons, identify what would have to be done and if there are any major blockers or impediments. I've been a long time user of docker images, but not so much experienced in building them (nor jupyter kernels).

@yitzchak, any thoughts?

[yitzchak]

Take a look at https://github.com/yitzchak/archlinux-cl. It has various CL implementations and quicklisp already set up. You'd just need to add JupyterLab.

[yitzchak]

Also, the dockerfile here is a bit out of date. I'll update it shortly. You could probably base on off of that.

[snunez1]

How would I do the reverse? Add the lisp kernel to an existing jupyter-docker-stacks minimal notebook?

[yitzchak]

The kernel installations are here https://github.com/yitzchak/common-lisp-jupyter/blob/master/Dockerfile#L74-L98