we are currently looking into ways how we can establish a collaborative workflow with jupyter like notebooks.
The problem hereby is with the dynamic nature of these documents that are being generated. Ive seen some solutions / suggestions on this topic that would fit the Yjs approach, however i do feel that those implementations limit usability for input related content. However when we talk about dynamically rendered output (widgets like components, dynamic html elements etc) that’s where we struggling to find an approach that would fit within CRDT. Alternative to this approach we evaluate “co-browsing” as a solution.
I was wondering if someone would be interested in doing some high level consulting for us to establish the right approach in that regard and lay out the possibilities as well as limitations of current CRDT or OT technologies.