Execution Flow Branching

[SeriousM]

Is your feature request related to a problem? Please describe.
When I explore new ways how to use a nuget package I prepare all required dependencies eg. "#r "nuget:..." and objects to work with.
After that I start to play with it. At certain checkpoints I want to change the execution flow to see alternative ways to use it but don't want to create a new notebook.
The same is true for working with sql data and different ways to visualize them without throwing already written code away.

Describe the solution you'd like
I want to add one/more vertical slices after a "checkpoint step" to have multiple execution alternatives.
Kind of LogicApp designer but easier to use.

Describe alternatives you've considered
I could achieve the same with multiple notebooks and include files to share certain pre-steps, but this would allow only a single branch checkpoint.

[jonsequitur]

Would this interaction model need a different UI so people understand the branching? Notebooks are very linear.

Another question this raises for me is whether #!import can be used in this way. Each branch could be a separate notebook with predecessors indicated by #!import. For example, common steps could be a.ipynb while branches could be a.1.ipynb, etc,. via #!import a.ipynb.

[SeriousM]

Notebooks are very linear.

I think notebooks really try to invite the user to play with data and supporting branching would also allow play with different flows.
Looking forward for other input!