I made an argument and backed it up with a series of technical points. Feel free to disregard my opinion piece in the end there.
How "i'm coming across" is at best irrelevant, at worst a silly attempt to avoid the issue. Either stay on topic and address the points I made or don't post at all.
All of life is 100% about that. Wish you all the best! Your points were paper cuts.
None of these things are broken and I use Juypter on my server. This is an non-issue for me and most people's use case. It is a document for sharing code and results and not for debugging. This is why Nature (#1 Science Journal using it and well just about http://www.nature.com/news/interactive-notebooks-sharing-the...)
> + Can't easily use it for remote debugging (which is really the most WANTED usecase for something like this) since the concept of remote kernels is alien to it. One has to hack around the internals to expose a plain ZMQ kernel and it's still not working without issues.
+ Jupyter can't connect to remote kernels it hasn't spawned itself, gotta use stupid tricks like tunnels to fake them as local.
+ IPython forces integration with single-threaded event loops and takes the "we know better, threads are complicated and bad" approach to exposing something simple that can be used by others to integrate into any sort of app, multi-threaded or not.
How "i'm coming across" is at best irrelevant, at worst a silly attempt to avoid the issue. Either stay on topic and address the points I made or don't post at all.
TY