Very cool. I think this is, ultimately, what Jupyter hopes to move towards.

Literally their first example is embedding D3...

http://sharing.beakernotebook.com/gist/anonymous/e21582541d7...

In case anyone's curious, this is what it looks like using just javascript: https://tonicdev.com/tonic/d3-example-from-beaker (a JS port of http://sharing.beakernotebook.com/gist/anonymous/e21582541d7... ).

I'm happy that I found this and pissed that I didn't find it earlier.

Looking forward to testing it. The fact that it works with Clojure is bad ass by itself.

I wonder if using JSON as a data bridge works well in practice. If it does, I imagine that a polyglot jupyter notebook type must be on the horizon.

Part of the reason we built Beaker was that the IPython devs explicitly stated they didn't believe in multiple languages within one notebook. It's possible that this opinion has changed with the Jupyter migration, but I haven't heard anything to that effect.

Why did they object to multiple languages within a notebook? What problem did they think it would create?

How have I not heard of this before.