We have received hundreds of good feedbacks from early adopters of the Jupyter extension using native notebook support. The clean UX, seamless integration with VS Code ecosystem and improved performance are mostly loved. However we are also seeing bugs and UX issues users are struggling with. Here I tried to compile the feedbacks into issues so we can track where we are with them.

Bugs & UX improvements

Core

• Problems with ctrl+z (cause the cell to run endlessly undo (Ctrl+Z) affects notebook cell metadata run state #106362)
• The control enter to run appears an increasing number of times the control enter to run appears an increasing number of times in markdown cell #113243
• Vim Coexistence of native notebook and Vim VSCodeVim/Vim#6104
  • jk to move between cells
  • the shortcut to clear the output is not working with vim bindings either
  • compatibility with vim extension when pressing escape
• Cell/output collapse Clickable area to expand should go across entire cell #104114, Bottom border of collapsed cell looks messy #104115,
  • When dealing with cells containing very long output is quite clunky, you can right click on the cell and collapse output, but then it's fully collapsed and you don't see anything
  • Collapsing a very long cell when you're looking at its end, the document scroll position should probably jump to the code cell that produced the collapsed output
  • Show the first line when a cell is collapsed
• Tooltip for every button

Jupyter

• 🐍 Kernel
  • 🐍kernel crashes (multiple reports of crashes)
  • 🐍could get kernel to connect
  • 🐍two kernel selection options at the bottom When using native notebooks, if kernel selector is shown, interpreter selector should be hidden vscode-jupyter#3979
  • 🐍the constant failing to start up a kernel
  • 🐍severe performance issues. appears to degrade over time, requiring regular restarts of the kernel Jupyter notebooks after a while become very slow vscode-jupyter#1622
• 🐍The old cell output stays when running empty cell
• 🐍Matplotlib plots (contour plots) very slow Jupyter notebooks after a while become very slow vscode-jupyter#1622
• 🐍The linter often gets stuck
• 🐍Renderer
  • Plots with all content outside the axes (i.e., tick labels, etc.) shaded such that they are difficult to read. Notebooks - HTML elements (widgets) in output overflow & get displayed under the editor #108947

Feature Requests

Core

• Markdown
  • 🏃 Latex Explore customizing the markdown renderer for notebooks #106701
  • Mapping of all Markdown capabilities into a menu
  • Markdown features (underline, bold, compared to Azure Data Studio)
• Outline Support outline with notebooks #98173
  • Heading navigation (like google colab)
• 🏃 Live share support API gaps to support Liveshare in Notebook #102503
• 🏃 Cell multi selection: selecting multiple cells to copy or cut would also be nice Support multi cell selection in notebook #114156
• Cell metadata editing

Jupyter

• 🐍Execution
  • 🐍Run all cells above or below a selected cell add run above and run below to native notebooks vscode-jupyter#4240
• 🐍 🏃 Full jupyter keyboard shortcut Explore notebook keymap support #106679
• Variable explorer Implement Variable Explorer in VS Code sidebar vscode-jupyter#250
• 🐍Renderer (❓ how many of them are covered by nteract/jupyter-ext?)
  • Date viewer
  • built in save picture as
  • option to save the plot in svg format
  • image zoom
  • full widget support
  • support full interactive plotting suite from pyviz
  • bokeh
  • holoviews
  • Interactive matplotlib plots
• 🐍Jupyter lab extensions (There's also jupyterlab extensions which would be great to include in vscode for plotly)
  • mathjax3-extension
  • geojson-extension
  • drawio
• 🐍Kernel
  • 🐍easily open a console connected to same kernel as notebook
• 🐍Language feaures Language Servers for Native Notebooks vscode-jupyter#1193
  • Go to definition and peek
  • Intellisense
• 🐍Resource monitoring tools for CPU/GPU/memory
• 🐍Set git not track outputs Settings for controlling if outputs should be saved to disk vscode-jupyter#4670
• 🐍Performance
  • optimizations for plotting
  • performance can be sluggish with large plots
• Customizable CSS to modify appearance of rendered markdown cells
• 🐍 ❓ Debug line by line Run-by-line (Backlog) vscode-jupyter#1160 & Native Notebook / Webview parity checklist vscode-jupyter#1231
• MISC
  • More Julia support