Use cases

[hammer]

As discussed at https://github.com/pystatgen/sgkit/discussions/954 we've decided to lead with use cases in the paper.

We discussed trying to implement these use cases as notebooks in the Examples section of our website. This would be helpful to ensure there's working code readers of our paper could reference, and it would a good place to try out rough drafts of use cases that can evolve over time before fixing a final draft forever in the paper. It also creates a separate step before the paper, is not quite the same as the writeup we'd like to produce for the paper, and also requires us to figure out a sample data set that is open and small enough to be run in CI/CD. It's up to the use case author whether they'd like to implement an Example first, or leave that for later.

Here are our suggested use cases, in no particular order:

• https://github.com/pystatgen/sgkit/issues/955
• https://github.com/pystatgen/sgkit/issues/602
• https://github.com/pystatgen/sgkit/issues/956
• https://github.com/pystatgen/sgkit/issues/688
• https://github.com/pystatgen/sgkit/issues/957
• https://github.com/pystatgen/sgkit/issues/958
• https://github.com/pystatgen/sgkit/issues/959

[hammer]

@tomwhite @jeromekelleher having thought more about the intermediate target of example notebooks, I think ensuring the data used in the example is public and small enough to run in CI/CD is a real killer. I'm curious how y'all are feeling after a week of thought?

[jeromekelleher]

Requiring these to be runnable in CI/CD would be a show-stopper for me all right - it pretty much rules out doing anything at an interesting scale.

[tomwhite]

Agreed - we don't need to add the CI restriction.