pluto is dead. long live pluto!
The first version of pluto had a big WTF factor that threw most users off: everything was batch processed. pluto v1.0 bypassed Jupyter’s execution engine and sent the entire notebook to the back-end which then recomputed the whole thing and sent it back. The intuition behind this approach was that it was easier to bring code to data than data to code. But this was not how must users worked on notebooks. Most split their code into “data-load” and “analysis” blocks – load all the data first in the beginning and then iteratively code up the analysis blocks. Batching this lead to longer dev times. Hence, 2.0
- Jupyterlab interface – sets pluto up for further enhancements.
- Traditional dev cycle – no more batch processing.
- Notebooks are not automatically uploaded to github – users can do this by logging into the terminal.
- SSL – users can now work knowing that their comms are secure.
- Dark mode!