Jerome Kelleher

Nice, that sounds like a good idea. Do you want to push forward with that, or merge this much and iterate? (I haven't reviewed again, will need to take another...

Yeah, your way is better.

Also looks useful: https://www.dhiwise.com/post/the-ultimate-guide-to-github-reusable-workflows-maximize-efficiency-and-collaboration

Also potentially useful for doing repetitive things across multiple repos: https://github.com/samlader/gmux

There's a general issue here that we've been too lax with checking mutation parents. We'll probably just have to figure out how to detect whether mutation parents are set properly...

Your probably right. We could check for cheaper things like if we have two mutations on a branch, the parent must be set, but that's a fairly limited check. So,...

That all sounds very sensible to me @jni. I don't really have a sense of whether I think the zarr2 package is worth it or not, but I'll shut up...

> A useful exercise that we could use some help with is trying to define the "public" Zarr API in v2 Everything that's documented in the [API reference](https://zarr.readthedocs.io/en/stable/api.html)?

How about we make the internal functions first, and decide about the external API later? The internal function would be based on numpy arrays as input and output, taking an...

> Should we be worried about the discrepancy between 0.01339734 and 0.01351895 here, given that the method is trying to make these the same? This seems like really close match...