Hacker News Reader

Comment by drauh | Hacker News Reader

Beaker Notebook (posted several times, but without much attention[0]) does something similar: http://beakernotebook.com/features

It supports these languages: Python, Python3, R, JavaScript, SQL, C++, Scala/Spark, Lua/Torch, Java, Julia, Groovy, Node, Ruby, HTML, and Clojure.

It has an experimental native version: https://github.com/twosigma/beaker-notebook/wiki/Electron-Be...

Talk at SciPy 2015: https://www.youtube.com/watch?v=iMPfLz6kKv8

[0] https://news.ycombinator.com/item?id=8364237

View on HN

Maybe because Jupyter has already more than 70 kernels and got adopted before other notebooks appeared (Beaker, Spark, Zeppelin)?

true but the languages are siloed, each notebook runs just one language. with beaker the languages can communicate with each other. there's no easier way to combine python and javascript for d3, for example: https://pub.beakernotebook.com/publications/7fdcaaa6-fb83-11...

there are lot more differences in the UI as well.

You can actually use R and Python in the same notebook. See https://blog.dominodatalab.com/lesser-known-ways-of-using-no...

The %Rpush and %Rpull magics are what you need.

Also, if you are using Spark, then Apache Toree[1] lets you use Python, R, Scala and SQL in the same notebook against Spark[2].

[1] https://toree.incubator.apache.org/documentation/user/how-it...

[2] https://github.com/ibm-et/spark-kernel/wiki/Language-Support...

Not everyone needs more than one language in the same notebook communicating between each other. But if required, then the cell magic system looks superior to me: have a look at %%fortran, %%cython, %%javascript, %%html, %%bash options. Also it is possible to switch kernels in the same notebook, but serialization of state between kernels is handled by user.

Cell magics just seem like one possible method of choosing which language to use, a menu such as what beaker uses seems alright too. I was setting up cling (the C++ kernel) today and I gave up at realising that the interpreter is spun up in a subprocess, rather than a kernel persisted across cells. Without cross cell support, mixing languages is a limited feature of jupyter.

what is superior about %%javascript?

here are some examples of JS working in beaker (in addition to the above python/JS demo):

https://pub.beakernotebook.com/publications/87a444cc-fb85-11...

https://pub.beakernotebook.com/publications/797f99f2-fccd-11...

https://pub.beakernotebook.com/publications/afa0f71c-4aaf-11...

Beaker is a cool idea. We were trying to use it instead of jupyter but it ended up being too buggy. Its been a few months, has it gotten better?

thanks. there are about 100 pull requests merged per month, so yea i think it's definitely getting better.

what bugs did you run across? i am very interested in your feedback.

I had a lot of fun hacking on Beaker Notebook at a hackathon a while back. It's a very powerful and cool environment.

thanks!

and Beaker Lab, a free hosted environment for running Beaker, was launched back in April: http://lab.BeakerNotebook.com

Small fix, on https://pub.beakernotebook.com "offing" should be "offering".

Does Electron really count as 'native'?

Definitely not in my book.

Yes.

Reply on news.ycombinator.com