Roadmap 2023 Q1

[lihebi]

P1 Roadmap: the most important features missing

• various UI issues mentioned in Features without which coding in CodePod would be painful  #181
• (@lihebi, WIP) Local runtime Local runtime spawner #212
• (@lihebi , WIP) .ipynb & .py file import/export and sync with github Export to files and sync with GitHub #211
• Version control
• Auto-layout
• Mount drive
• Save outputs: potential issues: output too large (e.g. image) to send via GraphQL query.
• Chain-of-pods execution: and persist the arrows to database Persist arrows to the database #210
• minimap with ToC for jumping nodes
• (@li-xin-yi) Yjs sync for prosemirror rich text editor #213
• (@li-xin-yi) copy-n-paste a scope Copying and pasting cells  #114
• (@li-xin-yi) fine-grained access control Edit/View privileges of shared projects #177
• (@li-xin-yi) record user-project data such as lastViewPoint Record and restore the previous position and zoom level on repo entry. #157

P2 roadmap

• Better Local release for those who are not into cloud
• Add comment
• Latex support in rich text editor
• issue tracking
• CI/CD
• runtime for more languages

Backend roadmap: the improvements under-the-hood

• Yjs server routing & auth & highly-availability
• Separate server from runtime & storage server

[forrestbao]

1. What's the relationship between "Local runtime" in P1 and "Better Local release for those who are not into cloud" in P2?
2. Q1 is a very long time, 3 months. Can we set timeline for the tasks?

[lihebi]

1. What's the relationship between "Local runtime" in P1 and "Better Local release for those who are not into cloud" in P2?

• local runtime: Allow users to launch and connect to the local runtime kernels for code execution. This is especially useful when they want to use their GPU machines. Users still write code on https://codepod.io, so they can share with others and collaborate in real-time. They won't need to worry about data storage and backup as well.
• local release: The app is deployed entirely offline. The current local deployment requires docker and save code in a database. A desired local release would be a local app with a file-based backend for persistence. But this is not our priority.

2. Q1 is a very long time, 3 months. Can we set a timeline for the tasks?

Let's keep refining the timeline.

[li-xin-yi]

You can use projects to manage the progress of those features.

[forrestbao]

I'd wanna add one more: something like ToC for users to quick jump between different parts of the program. Paging up and down is a major reason that development in Jupyter is slow. We cannot let users pan, scroll, or zoom to find what they want.

[lihebi]

You can use projects to manage the progress of those features.

I did try "project" several times in the past, but I found it overkill / hard-to-use (confusing issue linking, no comment, no markdown, weak timestamp tracking, and the multi-columns view seems complicated). It's probably more suitable when we have a larger team.

[lihebi]

ToC for users to quick jump between different parts of the program

Agreed; added the ToC/minimap to the P1 roadmap.

[lihebi]

You can use projects to manage the progress of those features.

One more thing: I'm working on export-and-sync-to-github (#211). Once that is done, we'll move many of the project docs & progress management to a public/shared codepod repo.