Hi, On 10/02/17 05:45, Serge Stinckwich wrote:
On Fri, Feb 10, 2017 at 12:48 AM, Offray Vladimir Luna Cárdenas <offray.luna@mutabit.com> wrote:
Excellent! I will create the proposal to have it ready for tomorrow. Great, tell me when this ready.
It is. See at: https://github.com/pharo-project/pharo-project-proposals/pull/34
Jupyter support could be nice, but I would think is some kind of ZeroMQ support really. That would enable multiple kernels inside Pharo with the advantages of live coding, powerful inspectors and debuggers, not present in the Jupyter notebook interface. Hydrogen[1] and nteract[2][3] are taking this route of enabling Jupyter kernels inside desktop apps with results far more interesting that what is possible on the web interface. I think there is a ZeroMQ implementation in Pharo ? In order to built a Pharo kernel we have two solutions: built something in Python that run Pharo from command-line or implement a kernel in Pharo. The 1st solution will allows you to have a working solution very fast, but I will prefer the other solution ;-)
Yes I would prefer the second one via ZeroMQ and Pharo. In that way you could use the same technology that Jupyter and nteract are using to enable several programming language kernels talking with the notebook. I think that the notebook we can create inside Pharo is far superior that the existing ones in terms of UI, inspectors, visuals and live coding.
[1] https://github.com/nteract/hydrogen [2] https://nteract.io/ [3] https://github.com/nteract/nteract Yes I know about these tools. I think we need to have first a basic support for Pharo and we can have something more live after that. and maybe use Grafoscopio as an interface ;-)
Regards,
Yes, using Grafoscopio as and interface for literate computing and reproducible research, native to Pharo and talking with several external languages would be a dream come true :-). Cheers, Offray