Kernel issues

[ghost]

Hello, When I attempted to install the C++ kernel, I found that it was incompatible with windows. Could you please both a) recomend a kernel that works on all platforms or b) bundle kernels for the common languages in (maybe a command - line option?) like Python, R, C++, Java, etc?

[takluyver]

It looks like it is meant to be possible to use Cling on Windows: https://root.cern.ch/build-prerequisites#windows

But I doubt many people have tried that with the cling kernel. We don't maintain it, so if you think it needs some changes, see https://github.com/root-project/cling .

As far as I know, that's the only C++ kernel anyone has tried to make. There aren't many C++ interpreters out there to base a kernel on.

[ghost]

I understand that there aren't many C++ interpreters to base a kernel on, but shouldn't the top 5 languages be packaged with jupyter? (perhaps as an install option, such as --install-kernels)

[takluyver]

That's a much bigger job than it sounds, and I don't see us tackling it - sorry. We generally recommend installing with conda, which makes it relatively easy to install quite a lot of kernels.

From a Jupyter perspective, I doubt that C++ would make the top 5 languages, in any case.

[ghost]

Weird... At least for the C++ kernel you recommended, conda doesn't install it, and there is no binary for windows. C++ doesn't make the top 5 for data science, but for general use, it's on the top 3.

[ghost]

That's not the case with popular kernels like r - R kernel works perfectly

[gnestor]

@ndrewxie Have a look at https://github.com/twosigma/beakerx. This is a Python/conda package that includes kernels for C++, Java, several JVM runtimes like Groovy and Scala, Kotlin, and SQL...in addition to notebooks extensions for plotting, tables, forms, etc. It's currently in alpha but should be installable with:

conda create -y -n beakerx python=3.5 jupyter
source activate beakerx
conda install -c beakerx beakerx

[gnestor]

Now you should be able to just do pip install beakerx 👍 👍