Carter Francis

This should be partially covered by #868. I would like to implement logging similar to hyperspy as well before the 1.0.0 release.

That sounds good to me. I'm interested in how you handle the different I/O backends and the benefits of them. It seems like there is a lot of potential optimization...

Both of those work for me as well!

> Should we maybe go for Tuesday and potentially a follow-up on Thursday? I have a feeling that we'll have quite some things to discuss and try, so we should...

Works for me!

I think that this is a really good idea. That being said we should probably have some type of standard going forward. Changing the commit history can be a major...

> I have started to use a zarr backend in this [branch](https://github.com/ericpre/hyperspy/commits/zarr_backend), which seems to address the issue you mentioned above, pretty much in the way you described them. This...

@din14970 Thanks for the suggestion. I've considered it a couple of times but never very earnestly. The part that I have had trouble recreating without a for loop is iterating...

@din14970 From my experience not rechunking the data is a huge advantage. When using memmap your original chunk is defined by which images are next to each other in the...

It seems like there are a couple of different best case scenarios. There is the compressed data vs uncompressed. Let me see if I have this correct... For uncompressed data:...