Tom Nicholas

@applio we talked about this at SciPy this year - are you or Colin (whose github handle I don't know) planning to finish this off? The other thing we discussed...

> Given the new manifests coming, I'm not sure if this is a good time to do this. I actually agree. Changing the file format would only help people who...

> You still need to kerchunk-scan the original files at some point - how would you write these references? If you use VirtualiZarr then Kerchunk gets called to do the...

I had the exact same question as @julioasotodv. It's interesting to see in the readme that HDF5 is described as > Apparently now discouraged for TF/Keras. Seems like a great...

Thank you for the detailed answer! > HDF5 was getting deprecated in TF/Keras at the time of writing this lib, which means building upon it didn't look like a great...

Is it just the codecs support left to do here now? I'm wondering if the references generation can be punted on by e.g. warning / raising if you try to...

I think the reason I originally put the sort call in there was because in the `itertools.groupby` docs https://docs.python.org/3/library/itertools.html#itertools.groupby it says > Generally, the iterable needs to already be sorted...

Sorry for taking so long to reply to this 😓 > AFAICT, the sorting before groupby makes sure that all Datasets with same variables are grouped together regardless of the...

I think this would also be useful for VirtualiZarr.

Good point, this wasn't really an intentional breaking of that convention on my part. Happy with either / both of your suggestions. I think I originally put parent in the...